Turning a simulation idea into a working application means assembling assets, connecting physics and rendering, and checking that the scene behaves as intended. Developers are combining frontier AI models with NVIDIA Omniverse libraries to help carry out that work — building applications for exploring scenarios, investigating failures and improving designs.

Developers direct AI agents through natural-language instructions, review results and guide changes. Omniverse libraries provide GPU-accelerated physics, rendering and sensor simulation capabilities.

Explore the projects below to see frontier AI models such as GPT-6 Astra at work, and check back for new examples from NVIDIA teams and developers across the ecosystem.

Build a Humanoid Simulator for a Warehouse Environment

Explore a gamified, physics-based control of a humanoid robot in first- and third-person view.

Before automating warehouse tasks, developers need an interactive simulation environment to explore task behavior and evaluate how the work gets done. Frank DeLise, Omniverse product manager at NVIDIA, used Astra to turn a SimReady warehouse and humanoid robot into an interactive simulator with first- and third-person views.

DeLise directed Astra to connect NVIDIA Omniverse libraries for physics (ovphysx), scene updates (ovstage), rendering (ovrtx) and the user interface (ovui). He also used Astra with SimReady (simready-foundation) to create the physical scene in simulation. Astra then generated animation and application code to bring those capabilities together.

Learn how to prepare and validate SimReady robot assets with frontier AI models and NVIDIA Omniverse libraries.

Connect an Autonomous-Driving Testing Workflow

Changing a scene, sensor or driving model can significantly affect autonomous-vehicle simulations. Doyub Kim, a manager on the simulation technology team at NVIDIA, asked Astra to build Zero to Alpamayo — a reusable simulation environment based on San Francisco’s Market Street.

Kim directed Astra to map out the workflow, then connect asset creation, traffic, Omniverse RTX sensor simulation and Alpamayo driving in stages, checking each integration. The resulting prototype was a testing ground for comparing models and tracing how scene or sensor changes affect downstream driving behavior. A separate Cosmos3-Nano experiment varied weather and lighting in recorded simulation videos, allowing Kim to compare the driving model’s responses to the same scenario under different conditions.

Use Sensor Differences to Create and Improve Digital Twins

Comparing recorded and simulated camera and lidar outputs gives developers key performance indicator-based feedback for refining a scene.

To test robots and autonomous vehicles, developers need to know how closely simulated sensors match real ones. Ashley Reid, who works on RTX sensor validation at NVIDIA, directed Astra and Claude Fable 5 agents to compare ovrtx camera and raw LiDAR outputs with recorded data. The agents created two digital twins from scratch and improved two existing ones.

Over about three days, Reid guided an iterative workflow in which agents measured differences, created or modified OpenUSD scenes, and checked the results. Changes addressed missing objects, geometry and materials, with acceptance depending on camera and LiDAR metrics. This gives developers a way to use measured discrepancies to guide scene creation and improvement.

Start by rendering an OpenUSD scene with the ovrtx minimal Python example, then define a sensor measure to compare with recorded data.

Robo Olympics: Test Robot Skills With Simulation

Teaching robots new movements means checking whether those actions work under physical constraints. Tae Kim, who leads NVIDIA Omniverse engineering and product, used sports videos and natural-language instructions to guide Astra in building Robo Olympics, an experimental project that tests simulated Unitree G1 humanoids performing sports movements.

Under Kim’s direction, Astra built controllers and refined them through physics trials. Newton Physics Engine simulated behavior, the open source NVIDIA Warp framework accelerated calculations and ovrtx rendered scenes and virtual-camera images. In one experiment, the robot cleared a single hurdle in 64 of 100 simulation trials. The trials gave Kim feedback for improving the robot’s timing and control.

Watch Kim explore building simulations with Astra and NVIDIA Omniverse libraries.

Test Robotic Disassembly With Computer-Aided Design and Simulation

Before a robot can disassemble a product, developers need to know whether its tools can reach and remove the parts. Jens Jebens, a senior product manager for OpenUSD at NVIDIA, directed Astra to model a car suspension in PTC Onshape and configure it in NVIDIA Isaac Sim.

With Astra, Jebens explored computer-aided design and tooling revisions for robots informed by simulation. The agent measured the available space and designed a wrench the robot could use to reach the suspension’s bolts. Jebens reported successful removal of a suspension component in simulation. This connects design and tooling decisions to disassembly results, offering a starting point for robot policy training.

Explore the Onshape importer guide.

Bring the International Space Station Into the Browser

Turning 3D models into an application requires connecting assets, live data and an interface. Nic Johns, an engineering director at NVIDIA, prompted Astra to assemble NASA assets into an OpenUSD International Space Station model with telemetry. Johns built the application with a single prompt, then used a follow-up prompt to shift the scene to Earth’s daytime side so the planet was visible.

The workflow used Blender for asset preparation and Omniverse libraries for rendering (ovrtx), scene runtime (ovstage) and streaming (ovstream). The application brings 3D models and operational data into the browser, with Johns guiding its development through prompts and corrections.

Try the Omniverse Real-Time Viewer skill.

Turn Captured Rooms Into Testing Environments

A digitally reconstructed room needs editable objects and accurate physical behavior before developers can test interactions. Chirag Majithia, from the Isaac engineering applications team at NVIDIA, directed Astra to turn stereo camera captures into an editable OpenUSD studio.

The workflow combined PyCuSFM, FoundationStereo and nvblox for reconstruction, with user review guiding object selection and placement. Astra assembled generated and Blender-authored assets and used USD Content Agents to configure how objects move and interact in simulation. Isaac Sim tests guided collision and contact revisions for doors and drawers. The studio connects captured geometry to interaction testing, making gaps and object behavior easier to inspect. 

Explore USD Content Agents.

Have a simulation idea? Explore NVIDIA Omniverse libraries to start building with an AI agent.

Gears of War: E-Day leads the charge on GeForce NOW this week, bringing Marcus Fenix and Dom Santiago’s first fight against the Locust Horde to the cloud with GeForce RTX-powered performance.

A new way to join the action is also coming: Fire TV users will soon be able to purchase GeForce NOW memberships directly through Amazon, creating a more seamless path from device setup to gameplay. Keep an eye out for availability in the coming weeks.

There’s more in store this GFN Thursday, with the cloud serving up 10 games, including seven new releases.

Emergence Day Is Here

Gears of War: E-Day is now streaming on GeForce NOW following its global release on Tuesday, Oct. 6. Set 14 years before the original Gears of War, the prequel follows Marcus Fenix and Dom Santiago as they confront the terrifying arrival of the Locust Horde on Emergence Day.

The series’ brutal combat, desperate survival and larger-than-life action return with a new origin story for the legendary partnership. Ultimate members can take on the Horde with GeForce RTX 5080-class performance in the cloud, plus technologies including NVIDIA DLSS and NVIDIA Reflex.

From campaign missions to cooperative and competitive modes, GeForce NOW keeps the fight ready across low-power laptops, Macs, mobile devices, handhelds, Firefox and other browsers, and more — no massive downloads or expensive gaming rig required.

New on the Cloud Menu

Members can look forward to the following games joining the GeForce NOW library this week:

• Gears of War: E-Day (New release on Steam and Xbox, Oct. 6), available on Game Pass
• STAR WARS: Galactic Racer (New release on Steam, Oct. 6)
• Clive Barker’s Hellraiser: Revival (New release on Steam, Oct. 8)
• Forever Ago (New release on Steam, Oct. 8)
• Order of the Sinking Star (New release on Steam, Oct. 8)
• Silver Pines (New release on Steam, Oct. 8)
• Permafrost (New release on Steam, Oct. 9)
• Graveyard Keeper 2 (Steam)
• Solasta II (Steam)
• Tyr (Steam)

The Xbox version of Dune: Awakening didn’t join GeForce NOW in September as previously shared. Stay tuned to GFN Thursday for updates on game availability.

Dates listed above reflect when games are released on their respective stores. GeForce NOW availability may vary, as games are onboarded after they are released and added throughout the week. Keep an eye on GeForce NOW channels and GFN Thursdays for availability updates on announced titles.

Gamers who want to test out high-performance GeForce NOW cloud gaming can start with a day pass to try premium cloud gaming before committing to a membership. Better yet, the cost of the day pass can be applied toward a first monthly membership — making it easy to try GeForce RTX-powered cloud gaming with the latest releases, then level up for more.

What’s on the play list this weekend? Let us know on X or in the comments below.

At a Microsoft event in San Francisco on Wednesday, NVIDIA founder and CEO Jensen Huang and Microsoft CEO Satya Nadella outlined how NVIDIA and Microsoft are co-engineering hardware and software for AI agents to run on Windows PCs.

NVIDIA was founded because of Windows, Huang said. Now AI agents are coming to Windows.

“If you look at the entire journey of our company, Windows was at the core of it,” Huang said, tracing back NVIDIA’s support for Windows and Microsoft over several decades. “If not for Windows there would be no GeForce.”  

Huang connected that history — and his long-term vision — to the agentic era Wednesday during a fireside chat with Nadella hosted by Sriram Krishnan, a former senior White House policy advisor on AI at the Windows AI and Surface event, held at Dogpatch Studios in San Francisco.

“The thing that I really give Jensen all the credit for, quite frankly, is that consistency of vision of what this can be—not today, not tomorrow, but as a secular trend,” Nadella said. “And that’s what brings us to this moment.” 

“The personal computer is the ultimate tool, it’s my ultimate tool, and for a whole generation of people, it’s our ultimate tool,” Huang said. “What happens in the era of agents, when the agent is on your computer? It is now your personal assistant.” 

The event opened with Nadella before Pavan Davuluri, executive vice president of Windows and Devices at Microsoft, walked through the Windows updates designed for the agentic era.

Turning Windows into a secure platform for agents, Microsoft announced general availability of Microsoft Execution Containers (MXC), the OS-level infrastructure that lets agents run safely and persistently in the background, under operating system control.

“We are building these primitives directly into Windows, so agents can be secured, observed and governed,” Davuluri explained. “With MXC, Microsoft Security and Agent 365, we’re unlocking the full potential of agents on Windows.”

Nadella returned to that theme in his talk with Huang. “We needed to make the desktop the most secure place for agents to execute,” Nadella said. 

“What Satya just said is going to be the foundation of the next generation of IT as we know it,” Huang said. “Just as Windows and DirectX revolutionized how applications were built, MXC is going to revolutionize how agents are built and deployed.” 

A New Beginning for Windows PCs: Preorder RTX Spark Laptops Today

Among the announcements Huang and Nadella outlined Wednesday, RTX Spark puts the full NVIDIA AI stack into Windows laptops and compact desktops, with laptop preorders open today and available Friday, Oct. 16. Compact desktops will be available for sale in November.

Windows is bringing local AI closer to everyday work, while new hardware gives developers room to run increasingly capable models right on their PC. Introducing Surface Laptop Ultra, Davuluri tied that vision to NVIDIA RTX Spark.

“We built Surface Laptop Ultra around NVIDIA RTX Spark,” Davuluri said. “With up to 128 gigs of unified memory and up to a petaflop of AI compute, you can run models on this laptop that simply don’t fit on a traditional machine.” 

Systems are coming from Acer, ASUS, Dell, HP, Lenovo, Microsoft, MSI and Gigabyte with designs ranging from slim laptops to compact desktops built for always-on agents. 

RTX Spark combines an NVIDIA Blackwell RTX GPU with up to 6,144 cores, and an up to 20-core NVIDIA Grace CPU connected at 600 GB/s. 

One petaflop of FP4 AI performance and up to 128GB unified memory makes RTX Spark effective for local AI. It can run models such as Qwen 3.8 Flash Next, a 125B model with 51B n-gram that matches the intelligence of many cloud models, unmetered and without sending data to the cloud.

RTX Spark runs the full NVIDIA CUDA platform — the same software stack that runs across NVIDIA hardware — and is built for every way people use a PC:

• Developers can move models, tools and workflows without rewriting, running the same NVIDIA AI stack from RTX Spark to DGX Station.
• Creators get 5th-generation Tensor Cores with NVFP4 support, hardware-accelerated AV1, and 4:2:2 video encode and decode, DLSS and RTX ray tracing across the full production pipeline.
• Gamers can run AAA games at 1440p over 100 frames per second with DLSS 5, Reflex and G-SYNC. 

The compact desktop configuration puts the same RTX Spark superchip in a small chassis designed for 24/7 operation — a dedicated local AI system that keeps agents running continuously. Preorder RTX Spark today.

NVIDIA DGX Station for Windows: Frontier AI Compute on the Enterprise Desktop

The event previewed NVIDIA DGX Station for Windows today — the first deskside AI supercomputer to bring GB300 Grace Blackwell-class AI infrastructure directly into the Windows ecosystem.

“This unlocks the power to run frontier-class models locally,” Davuluri said. “Capabilities that once required renting a cluster, now in a deskside supercomputer.” 

Until now, DGX Station ran on Linux — which meant enterprise developers maintained two separate environments: Linux for heavy AI workloads and Windows for the productivity tools, applications and workflows. 

The vast majority of Fortune 500 companies are standardized on Windows, and that gap has cost developers time and resources — developers either moved to Linux to access AI compute, or stayed in Windows with limited hardware options for heavy-duty model development and multi-agent workloads.

NVIDIA DGX Station for Windows runs on the GB300 Grace Blackwell Ultra Desktop Superchip, delivering 748GB of coherent memory and up to 20 petaFLOPS of FP4 AI compute — enough to run models up to a trillion-parameter scale locally. 

Teams of developers and researchers at AI-native companies, leading research labs and enterprises can build and run always-on AI agents that connect directly to the Windows applications and infrastructure they already use, and fine-tune and inference large models without leaving their primary machine. Linux AI toolchains remain accessible through WSL when needed.

Learn more about NVIDIA DGX Station for Windows and sign up to be the first to know when it’s available.

Follow NVIDIA RTX Spark on X, Instagram, TikTok and Facebook — and stay informed by subscribing to the NVIDIA Local AI newsletter. Follow NVIDIA Workstation on LinkedIn and X. 

See notice regarding software product information.

Telecom operators are increasingly building their AI strategies on open models — and the reasons go beyond mere cost. 

Open models give telcos the ability to trust, control and customize AI across their most critical workloads — from autonomous networks to customer care. 

NVIDIA’s latest State of AI in Telecommunications report reflects this shift, with 89% of respondents reporting that open source models and software are important to their company’s AI strategy. 

For operators, the strategic value of open models is fivefold:

1. They expand access to frontier‑level intelligence at lower cost, allowing operators to reserve closed models for the workloads where they drive the most value. Independent benchmarks such as the Artificial Analysis Intelligence Index v4.3.2 show that leading open models are becoming more competitive across demanding reasoning, coding, scientific and agentic workloads.
2. They support telco‑specific customization, with open weights and training recipes that operators can fine‑tune for their own operations using network, customer and industry data.
3. They enable trustworthy AI by giving telcos greater visibility into and control over model artifacts and behavior, so models can be evaluated, adapted and governed in alignment with regulations and business policies.
4. They enable flexible, secure deployment: teams can size and optimize open models to run across public clouds, private infrastructure and edge environments.
5. They unlock the opportunity for telcos to deliver locally adapted AI services to enterprise and government customers by hosting and fine-tuning open models.

Open Foundations, Tuned for Telecom Operations

The NVIDIA Nemotron family of open models provides frontier-level reasoning performance optimized for agentic workflows, as well as speech capabilities for voice applications, with open weights, training data and recipes. 

SoftBank Corp. illustrates how an operator can use open models as a basis for developing and continuously advancing its telecom-specific AI capabilities.

“Open models allow SoftBank Corp. to build on the rapid progress of global foundation models while applying the network knowledge and operational expertise we have accumulated over many years,” said Rajeev Koodli, principal fellow of SoftBank Corp. and senior vice president of SB Telecom America. “We are using open foundations extensively, including NVIDIA Nemotron models and others, in developing our SoftBank Large Telecom Model, continuously advancing it for telecom-specific use cases such as network operations, design and overall management.”

In addition, NVIDIA is collaborating with partners to help turn open models into practical building blocks for telecom AI.

NVIDIA announced the 30-billion-parameter Nemotron 3 Large Telco Model (LTM), fine-tuned by AdaptKey on open source telecom datasets to deliver accuracy gains for telecom-specific tasks. This gives operators an open baseline that can understand telecom industry terminology and reason effectively through telecom operations workflows such as network configuration and customer incident triage.

To help operators customize the Nemotron 3 LTM and other open models with their own operational data, NVIDIA released the full recipe that walks through the end-to-end fine-tuning pipeline for adapting open models to operator-specific networks, customers and procedures using NVIDIA NeMo open libraries. 

Scale Open Models Into Production Workflows

Open models are a critical building block, but it takes more than models to bring autonomous telecom operations safely into production.

For AT&T, the value of model choice lies in making that flexibility operational in alignment with its business priorities.

“At AT&T, we believe the future of AI is not about choosing a single model; it’s about intelligently matching every workload to the right combination of performance, cost and control,” said Andy Markus, chief data and AI officer of AT&T. “Open models are essential to that approach, and with NVIDIA, we can bring that model-choice strategy into production with the scale, reliability and governance our business requires.”

Operators need pipelines that prepare and protect their data for model fine‑tuning by anonymizing sensitive records and generating privacy‑preserving synthetic datasets. They also need a platform that can turn open models into governed, autonomous agentic workflows.

NVIDIA provides that end-to-end platform, powered by NVIDIA AI Enterprise software and NVIDIA Agent Toolkit, and spanning data pipelines, open models, agent orchestration, secure runtimes and simulation.  

Together with a strong partner ecosystem building at every layer of this platform, telecom operators have a clear path to translate the benefits of open models into production-ready AI workflows.

Create a Platform for Local AI Innovation

As telecom operators build AI infrastructure aligned with national AI strategies, open models are giving them a foundation to deliver AI services tailored to local languages, industries, regulations and data governance requirements. 

Operators across countries can fine-tune open models for local-language and industry-specific needs, as well as host open models on their trusted platforms that customers can consume directly or build on with their own data and applications.

For Indosat Ooredoo Hutchison, one of Indonesia’s largest telecom operators, that means building AI that reflects the country’s own language and culture through its Sahabat-AI family of open source models.

“For countries like Indonesia, the value of open models goes beyond access to powerful AI. It is about adapting that intelligence to our own language, culture, data and real-world needs,” said Chirag Sukhadia, chief data and AI officer of Indosat Ooredoo Hutchison. “Sahabat-AI puts this into practice, using open models as a foundation to build AI that understands Indonesia and can be developed for local applications. This allows us not only to adopt AI, but to build capabilities around it and enable more Indonesians to create with AI on their own terms.”

Learn more about NVIDIA technologies for telecommunications.

Breast cancer is the most commonly diagnosed cancer among American women — yet the gaps in care are wide. A majority of women over age 40 skip the recommended annual screening. Radiologists are reading more mammograms with fewer colleagues. And when a diagnosis arrives, the tests that inform treatment can take weeks to return results. 

Companies in the NVIDIA Inception program for startups are building AI applications to support clinicians at each of these friction points, including imaging, risk assessment and treatment planning.

About 40 million mammograms are performed in the U.S. each year, but a projected shortfall of tens of thousands of radiologists over the next decade is straining the system’s capacity to read them. 

At the other end of the care timeline, treatment decisions often hinge on genomic assays sent off to outside labs — assays that take weeks to process, at a moment when speed and certainty matter most. 

The companies below are addressing both ends of that gap, and everything in between, accelerated by NVIDIA AI infrastructure.

Automated Imaging Delivers AI-Powered Insights

For women struggling to find time or a nearby location for breast cancer screening, access is a practical barrier that translates into missed diagnoses. NVIDIA Inception startup iSono Health was built around simplifying this imaging workflow.

The company’s FDA-cleared ATUSA platform — a wearable, automated 3D quantitative ultrasound system — captures a standardized breast volume in approximately two minutes per breast, compared with up to 45 minutes for a conventional handheld ultrasound. 

The system’s AI, trained on thousands of full-breast scans comprising over 1.5 million ultrasound frames, automates image acquisition. It uses NVIDIA GPU acceleration and open source medical imaging technology to deliver a 3D scan that the company says is 28% more sensitive than a handheld 2D ultrasound.

Handheld ultrasound depends on whoever holds the probe, so a woman’s scans typically can’t be compared from one year to the next. ATUSA captures the whole breast the same way every time, creating a repeatable view of breast tissue that could help clinicians analyze how a patient’s tissue changes across successive scans, while also reducing operator errors and variability.

ATUSA is commercially available through partner clinics across California, Texas, Georgia, Tennessee and Washington D.C., with new sites coming online regularly. 

“Getting the scan closer to the patient is the first breakthrough,” said Neda Razavi, CEO of iSono Health. “Our vision is to make that scan increasingly informative: helping clinicians see what is there, understand what has changed and make more informed decisions.” 

iSono Health has developed AI capabilities for lesion detection, 3D segmentation and lesion classification. It next plans to extend its AI pipeline into multimodal diagnostic intelligence spanning 3D ultrasound, mammography, MRI and clinical information.

The company has a multicenter clinical study with 3,200 patients underway to further validate the platform’s performance, with lead research sites at UC Davis and Vanderbilt University Medical Center. 

Finding Cancers, Cutting False Alarms

Another NVIDIA Inception company, Whiterabbit.ai, develops AI technology for breast cancer screening. Its FDA-cleared WRDensity software automatically assesses breast density from mammograms and has been used in the care of hundreds of thousands of patients.

The company has also developed WRRisk, a clinical decision support software for estimating patients’ long-term risk of developing breast cancer — and is researching a new generation of AI for mammography that could help radiologists detect more cancers while automating the screening of mammograms that are negative for breast cancer. 

The goal is to reduce the burden on a strained radiologist workforce, accelerate results and reduce avoidable callbacks for patients, and lower downstream healthcare costs.

“Every day, breast radiologists face a needle-in-a-haystack problem, trying to find roughly one cancer in every 200 mammograms,” said Jason Su, cofounder and chief technology officer of Whiterabbit.ai. “We hope AI can be a powerful sidekick to radiologists, helping to clear away the hay so they can focus their expertise where it matters most.”

Whiterabbit trains its AI models on a cluster of NVIDIA GPUs housed at Washington University in St. Louis, supplemented by additional GPU capacity in the cloud. Inference runs on NVIDIA GPUs deployed directly in the clinic.

Predicting Which Treatments Will Work

Once a patient is diagnosed with breast cancer, the next question is what to do about it — and the answer depends on predicting how the cancer will respond to treatment. 

Today, these predictions are limited in scope and accuracy, and often require a separate tissue biopsy with a two- to four-week wait. Ataraxis AI is building clinical intelligence that predicts patient outcomes and response to different therapies using digital data, including pathology slides that are already part of the standard patient workup.

“The tools oncologists rely on today to guide therapy decisions were largely trained once, fifteen years ago, and never updated. Our models get stronger every time we acquire more clinical trial data,” said Joseph Cappadona, member of technical staff at Ataraxis AI. “But as we scale our models, the bigger shift is being able to answer more questions to help oncologists personalize therapy across all cancers.”

Ataraxis’s AI models analyze digital pathology slides and standard clinical variables to predict treatment response and recurrence risk. One model predicts whether presurgical chemotherapy is likely to shrink a patient’s tumor to the point of response before the operating room. After surgery, another model estimates a patient’s five-year recurrence risk and the likely benefit of chemotherapy as a next step. 

Both models have been validated across more than 10 institutions and multiple clinical trials, and are in active clinical use. They run on NVIDIA GPUs on premises, in an offsite data center and in the cloud — using PyTorch accelerated by NVIDIA CUDA throughout.

Providing Tumor Insights With 3D Visualization

Another NVIDIA Inception company, SimBioSys, joined NVIDIA on a recent panel marking Breast Cancer Awareness Month. 

The company builds AI-powered precision medicine technology that creates accurate 3D models of breast tumors, veins and other soft tissue, delivering important insights to help guide surgeries and influence treatment plans. It has also built a tool to estimate the risk of breast cancer recurrence based on 3D volumetric data from a patient’s breast MRI, tumor pathology and clinical data. 

“We’re building a platform that now allows us to take multimodal data — imaging exams, pathology results, genomic testing when applicable and other biological inputs — and, using AI, bring that all together,” said Stacey Stevens, CEO of SimBioSys, at the event. “When we do that, it generates new insights beyond what we had from any individual piece.”

SimBioSys uses NVIDIA MONAI for training and validation data, and NVIDIA CUDA-X libraries including cuBLAS and MONAI Deploy for its imaging technology, which runs on NVIDIA GPUs in the cloud. 

“NVIDIA technology gives us the computing power to take hundreds or thousands of images, apply our AI and analyze them quickly,” Stevens said. “That speed matters because patients and physicians need answers quickly. They can’t afford to wait days or weeks.”

Learn more about the NVIDIA Inception program for startups.

Certain technologies described in this article are investigational and have not been approved by the U.S. FDA for commercial use.

Local AI is becoming more useful by the token.

As AI agents move from experiments into everyday development, increasingly capable open models are shrinking to fit on more devices, giving builders more to run locally. 

Coming this month, NVIDIA DGX Spark will be available with 64GB of unified memory from top manufacturer partners — Acer, ASUS, Dell, Gigabyte, HP and MSI — giving developers, researchers and AI enthusiasts a new configuration with DGX OS and the NVIDIA AI software stack ready to use from day one.

The new SKU runs capable local agents on device — privately, without cloud dependency. And when workloads grow, two units can cluster together via NVIDIA Sync Cluster Assistant without any additional setup.

A New Starting Point for Personal AI Supercomputing

DGX Spark combines NVIDIA Grace Blackwell compute, unified memory, NVIDIA ConnectX-7 networking and an NVIDIA CUDA-accelerated AI software stack in one system. It’s a complete local AI platform for agents, inference, fine-tuning, data science and edge development.

The compact, personal AI supercomputer provides a place to experiment with models and developers’ own data without turning to a cloud instance for every task.

The new 64GB configuration, available exclusively from manufacturer partners, keeps the platform at an accessible price point while retaining the GB10 Grace Blackwell Superchip, DGX OS and full NVIDIA AI software stack — same as the 128GB model. It supports up to 100-billion-parameter models and the agentic applications built on them, fully on device. 

Two 64GB units clustered together don’t just double the memory. In NVIDIA’s Qwen 3.8 27B test, two clustered 64 GB systems delivered up to 1.7x performance compared with a single system, with room to keep scaling as workloads demand.

DGX Spark ships ready for agent development from day one — NVIDIA Agent Toolkit, CUDA-X AI libraries, Nemotron open models, and popular runtimes like Ollama, vLLM, and PyTorch with CUDA are all supported out of the box. Developers can go from power-on to running models in minutes.

Blender is among the first major creator application providers to support the platform, with a prebuilt, downloadable installer coming soon.

Scale Up With NVIDIA Sync Cluster Assistant

Developers can start with the memory their projects need today and build on a platform designed to seamlessly scale multi-node clusters for larger workloads as their pipelines grow. 

Every DGX Spark ships with a built-in NVIDIA ConnectX-7 NIC right out of the box. Plus, two units can connect directly with a QSFP cable, pooling their memory to 128GB and expanding model support to up to 200 billion parameters while delivering twice the memory bandwidth and up to 1.7x the performance.

The NVIDIA Sync app configures this multi-node cluster seamlessly. The cluster assistant feature detects connected units, validates device configuration and configures the ConnectX-7 network, so developers can focus on their work rather than the infrastructure. Every node runs the same NVIDIA software stack, so nothing needs to be reconfigured when scaling from one unit to two.

And coming at the end of the month, NVIDIA Sync Model Launcher makes running local AI as simple as clicking a few buttons. Developers can download and launch Qwen3.8 27B on a single DGX Spark system or a cluster, with NVIDIA Sync configuring the model to run across connected devices and making it accessible from users’ laptops. The launcher will also set up OpenCode to use the model, so developers can start coding in their browser.

Developer Use Cases on DGX Spark 

The new DGX Spark 64GB configuration supports practical work from day one. With up to 100-billion-parameter models running entirely on device, developers and enthusiasts can start with a single system for models that fit within its memory, or connect multiple DGX Spark systems with NVIDIA Sync Cluster Assistant for workloads that need more memory and compute. 

Here are three workflow examples:

• Run an AI agent around the clock: Keep a coding or research agent running on DGX Spark, ready to review code, analyze documents or carry out multistep tasks. A cluster provides additional capacity for larger models, longer context windows or multiple agents working at once.
• Power AI apps on your everyday PC: Run a language- or image-generation model on DGX Spark while using an agent or creative application on laptops or desktops. DGX Spark handles the model inference, freeing PCs for other work. 
• Scale when the work grows: When a single task outgrows one unit — running a larger model, a longer context window or concurrent agent requests — two DGX Spark 64GB systems connected over the 200 GbE fabric via NVIDIA Sync Cluster Assistant pool their memory to 128GB. The same workflow that ran on one unit scales to two without reconfiguring the software environment.

Get Started With DGX Spark 

DGX Spark 64GB is available from Acer, ASUS, Dell, Gigabyte, HP and MSI on Friday, Oct. 23, starting at $4,999.

To get started:

• Download a supported inference framework — llama.cpp, Ollama, vLLM or LM Studio.
• Download the recommended local model for the workflow.
• To scale to two units, connect them via their NVIDIA ConnectX-7 ports and launch NVIDIA Sync Cluster Assistant — it configures the network and routes workloads automatically.

For agentic AI playbooks on DGX Spark, visit the NemoClaw, OpenClaw, Hermes Agent and OpenShell pages on build.nvidia.com. 

#ICYMI: More Updates From NVIDIA Local AI

Explore playbooks on build.nvidia.com/spark for DGX Spark. The following playbooks are coming soon to 64GB devices:

• Serve LLMs With vLLM
• Run OpenClaw With a Local LLM
• Connect Multiple DGX Sparks for Distributed Workloads

New Windows PCs powered by NVIDIA RTX Spark are coming this month from Acer, ASUS, Dell, HP, Lenovo, Microsoft and MSI. Sign up for the RTX Spark newsletter to receive future updates. 

Alibaba’s Qwen-Image-2.1 brings image generation and editing together in a lightweight, open-weight model. It runs locally on NVIDIA RTX GPUs, DGX Spark and DGX Station, giving creators more ways to create and refine images on their own hardware.

Follow NVIDIA RTX Spark on X, Instagram, TikTok and Facebook — and stay informed by subscribing to the NVIDIA Local AI newsletter. Follow NVIDIA Workstation on LinkedIn and X. 

See notice regarding software product information.

GPT-6 Astra Ultrafast, running on NVIDIA Blackwell GPUs, is available now in the OpenAI API and to eligible ChatGPT Work and Codex users. 

Accelerated by inference optimizations through OpenAI’s models that tap into the capabilities of the NVIDIA Blackwell architecture, Ultrafast offers up to 8x faster token generation than the Astra Standard mode. For developers, faster generation can shorten coding agents’ edit-test-debug cycles, reduce the time spent generating responses between tool calls and make interactive applications feel more responsive. 

A faster response matters most when it’s repeated across a workflow: an agent writes code, uses a tool, checks the result and decides what to do next. Ultrafast brings Astra’s capabilities into these time-sensitive loops. NVIDIA AI infrastructure helps OpenAI serve more useful model outputs when developers need it. 

“NVIDIA’s deep investment in tooling and documentation has enabled us to make our models exceptionally good at programming Blackwell and Rubin GPUs,” said Philippe Tillet, inference lead at OpenAI. “Astra can turn that knowledge into high-performance kernels that make NVIDIA hardware compelling across the full frontier of latency, throughput and cost. With Astra Ultrafast, that means faster model responses as agents write code, use tools and work through complex tasks.”

Continually Improving Performance

Performance gains don’t stop when a model is deployed. OpenAI is using its own models to help refine the inference software running on NVIDIA GPUs, taking advantage of the platform’s programmability to test and implement improvements. That ongoing work can make model responses faster and deployed infrastructure more productive over time. 

“Our work with NVIDIA is helping us make AI faster and more useful,” said Uday Ruddarraju, chief technology officer of compute at OpenAI. “We used our internal models to optimize inference on NVIDIA GPUs, and NVIDIA’s programmability helped us deliver the acceleration behind Astra Ultrafast.”

A programmable NVIDIA platform allows developers and researchers to reuse infrastructure across training, inference and reinforcement learning as models evolve. That flexibility helps teams repurpose compute resources as demand changes, improving utilization and avoiding overprovision for each workload.

Developers can use GPT-6 Astra Ultrafast through the API today. See the Ultrafast guide for access, pricing and implementation details.

Spooky season is streaming in. Alongside falling leaves, pumpkin spice and everything nice, 25 new games are joining GeForce NOW throughout October, including six ready to play this week.

From a new CONTROL Resonant reward for Performance and Ultimate members to The Witcher 3: Wild Hunt – Remastered joining the cloud, this GFN Thursday is packed with fresh reasons to play.

Claim Rewards

CONTROL Resonant is streaming from the cloud for GeForce NOW Performance and Ultimate members. Explore a warped Manhattan on the brink of paranatural annihilation as Dylan Faden harnesses extraordinary powers to battle the Hiss, the Mold and other reality-bending threats while searching for his sister, Federal Bureau of CONTROL Director Jesse Faden.

Performance and Ultimate members can claim the Third Ice Baseball Cap reward for CONTROL Resonant. After claiming the reward code, launch the Steam or Epic Games Store version of CONTROL Resonant, select “Options” from the main menu, navigate to “Gameplay” and choose “Enter Redeem Code.” The reward can be equipped after completing Act 1.

Cap it off: Claim the reward from Thursday, Oct. 1, through Sunday, Nov. 1, at 11:59 p.m. PT, before it slips into another dimension.

The White Wolf Rides Again

The Witcher 3: Wild Hunt – Remastered launched Tuesday, Sept. 29, and is now streaming for Performance and Ultimate members. Follow the iconic tale of Geralt of Rivia, a monster slayer for hire, as he tracks down the Child of Prophecy in a war-torn, monster-infested open world.

The remastered adventure brings Geralt’s beloved world back at refreshed specs across nearly any device. GeForce NOW handles the rendering, downloads and storage in the cloud, so there’s no expensive hardware upgrade or lengthy install standing between the trail and the next contract.

Ultimate members can take the Continent further with GeForce RTX 5080-class performance in the cloud — high-end horsepower worthy of the White Wolf, with no local hardware upgrade required.

In addition, members can look for the following titles streaming this week:

• Minecraft Dungeons II (New release on Xbox, Sept. 29), available on Game Pass
• Nivalis Nights (New release on Steam, Sept. 29)
• The Witcher 3: Wild Hunt – Remastered (New release on Battle.net and Xbox Play Anywhere, Sept. 29), available on Microsoft Store
• Bookshop Simulator (Steam)
• Megastore: Tidy Up Together (Steam)
• Planet Coaster 2 (Xbox), available on Game Pass

And look forward to the games coming throughout October:

• Gears of War: E-Day (New release on Steam, Oct. 6)
• STAR WARS: Galactic Racer (New release on Steam, Oct. 6)
• Clive Barker’s Hellraiser: Revival (New release on Steam, Oct. 8)
• Order of the Sinking Star (New release on Steam, Oct. 8)
• Silver Pines (New release on Steam, Oct. 8)
• Permafrost (New release on Steam, Oct. 9)
• Deep Dish Dungeon (New release on Steam, Oct. 13)
• Planet Zoo 2 (New release on Steam, Oct. 13)
• Total War: SHOGUN 2 – Complete Edition (New release on Steam, Oct. 13)
• Valor Mortis (New release on Steam, Oct. 13)
• Among the Trolls (New release on Steam, Oct. 14)
• over the hill (New release on Steam, Oct. 14)
• Warhammer 40,000: Boltgun 2 (New release on Steam, Oct. 14)
• Beyond These Stars (New release on Steam, Oct. 16)
• Tenebris Somnia (New release on Steam, Oct. 16)
• Dimraeth (Steam)
• Graveyard Keeper 2 (Steam)
• Solasta II (Steam)
• Tyr (Steam)

More from September

In addition to the games announced last month, 11 more joined the GeForce NOW library.

• The Blood of Dawnwalker (GOG.com)
• Death Stranding Director’s Cut (Epic Games Store)
• DragonSword: Awakening (Steam)
• Dune: Awakening (Xbox)
• Everything Is Crab (Steam)
• LEGO Batman: Legacy of the Dark Knight (Steam)
• Over the Top: WWI (Steam)
• Roadside Research (Steam)
• RuneScape: Dragonwilds (Xbox)
• Stick It to the Stickman (Steam)
• Wild West Pioneers (Steam)

Dates listed above reflect when games are released on their respective stores. GeForce NOW availability may vary, as games are onboarded after they are released and added throughout the week. Keep an eye on GeForce NOW channels and GFN Thursdays for availability updates on announced titles.

Gamers who want to test out high-performance GeForce NOW cloud gaming can start with a day pass to try premium cloud gaming before committing to a membership. Better yet, the cost of the day pass can be applied toward a first monthly membership — making it easy to try GeForce RTX-powered cloud gaming with the latest releases, then level up for more.

What’s on the playlist this weekend? Let us know on X or in the comments below.

AI factories are built by the megawatt, even by the gigawatt. Each megawatt factory costs roughly $60 million, and AI factory operators will only commit capital on that scale with a clear view of the return on investment. Three key things shape AI factory returns:

1. Earning capacity: What the factory could earn in a year if it sold every token it can produce.
2. Useful life: How long its AI hardware keeps earning.
3. Demand: How much demand there is for those tokens.

Strength cannot fully offset weakness in another. High earning capacity counts for little if the factory sells only part of what it can produce. High demand matters little if it stops producing at full capacity in just a year. Nor are the three independent. A factory that can run more kinds of workloads finds more demand, keeping it earning year after year.

NVIDIA AI factories are engineered to maximize all three. They’re:  

• Productive: Delivering the highest throughput per megawatt and the lowest cost per token, which maximizes their earning capacity.
• Durable: NVIDIA GPUs and systems keep earning years after they ship, extending useful life.
• Fungible: They run every type of AI — in every phase and every place — as well as many workloads that don’t involve AI at all, which deepens and broadens the demand they can serve.

Engineering codesign across the full stack maximizes AI factory throughput, and continuous software optimization keeps installed hardware productive years after it ships. NVIDIA CUDA-X libraries let a factory run any accelerated workload. A standardized architecture then puts all of it within reach of any operator, deployable from a validated reference design.

Productive: Highest Tokens Per Megawatt and Lowest Token Cost

Power is the binding constraint on an AI factory. This makes tokens per second per megawatt the number that governs earning capacity. More tokens inside a fixed power envelope means more revenue. Lower cost per token means more margin on it.

SemiAnalysis AgentX data shows NVIDIA Vera Rubin NVL72 systems deliver over 30x higher throughput per megawatt than NVIDIA GB300 NVL72, and up to 45x lower cost per million tokens on the DeepSeek V4 Pro model. Gains of that size come from extreme codesign across the full stack: from models and workloads down through software to compute, networking and memory, all optimized together.

Two questions follow. If every generation makes tokens dramatically cheaper, does demand for compute shrink? 

No, it expands. 

Cheaper tokens make more use cases economical, and those use cases consume more tokens than the efficiency saved. 

The second question is about durability. If each generation is so much better than the last, what happens to the older generations?

Durable: The Installed Base Keeps Earning 

Not every workload needs the newest system. The right fit depends on a workload’s complexity and shape, which is why the last generation keeps earning after the next one arrives. 

The NVIDIA A100 GPU shipped in 2020 and is still in commercial service six years later, demonstrating its continued economic value; CoreWeave recently extended bookings for units first introduced in 2020 through 2029. Over the years, every major operator has extended the depreciation schedule on its servers, which is a guess about when hardware stops earning, and one that keeps moving out. A September 2026 Sprout analysis, “The Productive Life of a Data Center GPU,” tracks how that schedule has shifted across every major operator.

Barkr puts useful life at five to six years for an eight-GPU H100 system and nine to 10 years for GB300 NVL72, based on what those systems resell for. Silicon Data shows a six-year-old A100 GPU is still worth a quarter of what it cost, where a five-year depreciation schedule had it at zero more than a year ago. Ornn Data finds the market paying 80% as much to rent an A100 GPU on a five-year contract as on a one-month contract.

CUDA, the software platform with which NVIDIA GPUs are programmed, runs across generations, so nothing an operator already owns is stranded when a new architecture arrives. Continuous software and kernel optimization keeps improving what existing hardware can do.

The same platform runs machine learning, deep learning, generative AI, reasoning, agentic AI and physical AI. Each new kind of work arrived on hardware that was already installed.

That’s fungibility. The more kinds of work a system can take, the longer it keeps finding work.

Fungible: Every Type of AI, Every Phase, Every Place

A factory built for one kind of work is a bet that the work stays. A factory that runs everything stays useful and revenue-generating even when the work changes.

NVIDIA AI factories run every type of AI model — open and proprietary — across language, vision, biology, physics and robotics. They run every phase, from data processing through pretraining, post-training and inference. And they run in every place, from hyperscale and AI clouds to sovereign programs, enterprise data centers and the edge.

Not all of it is building and running AI models. The same infrastructure runs data processing, scientific computing, simulation, graphics and more.

All of those workloads, AI and non-AI alike, reduce to the same parallel math, and NVIDIA GPUs are built to run exactly that across thousands of cores at once. CUDA is why one chip can simulate light, fold a protein and predict the next token. More than 1,000 ready-made CUDA-X libraries and models sit on top, covering everything from deep neural networks, computational lithography and quantum circuit simulation to vector search and climate modeling, with more than 10 million developers building on them. 

That’s what makes NVIDIA GPUs general-purpose accelerated computing rather than a custom ASIC built for one workload. Being general purpose does not mean being generic: Tensor Cores and the Transformer Engine put AI-optimized hardware inside a programmable architecture, delivering specialization and flexibility in one chip. One architecture running all of it is what keeps utilization high, and the return with it. 

This versatility shows up in production across customers:

• Lilly: Building and running protein, small-molecule and genomics models on a 1,016-GPU, on-premises cluster, plus chatbots and agentic workflows for its own teams.
• Pinterest: Post-training and deploying a vision language model on a hyperscale cloud, across 14,000 GPUs spanning NVIDIA Blackwell, Hopper and earlier architectures.
• Revolut: Processing data for billions of transaction records with NVIDIA cuDF, then training and deploying a foundation model on an AI cloud.
• Runway: Training a world model on NVIDIA Hopper and serving on the NVIDIA Blackwell platform using cloud infrastructure.
• Texas A&M University: Running molecular simulation and AI drug discovery on its supercomputer, at 95-98% utilization across 26 projects and seven institutions.

The same is true beyond AI. Cosm operates a distributed NVIDIA-powered data center providing flexible compute resources across high-resolution video playback, streaming, real-time graphics and synchronized display. Dassault Systèmes powers the virtual twin simulation behind aircraft certification at Wichita State and vehicle design at Lucid Motors. And Unilever builds product imagery from digital twins rather than photo shoots, cutting production costs in half.

No list of examples here would be complete — that’s the point. 

NVIDIA AI factories are engineered to be productive, durable and fungible: more profitable tokens, longer useful life and deeper demand. That’s what maximizes their return. 

Learn more about NVIDIA AI factories by tuning in to NVIDIA founder and CEO Jensen Huang’s GTC Berlin keynote on Wednesday, Oct. 21, at 11 a.m. CEST.

Bringing together the world’s brightest minds and the latest accelerated computing technology leads to powerful breakthroughs that help tackle some of the biggest research problems.

To foster such innovation, the NVIDIA Graduate Fellowship Program provides grants, mentors and technical support to doctoral students doing outstanding research relevant to NVIDIA technologies. The program, in its 26th year, is now accepting applications worldwide.

It focuses on supporting students working in AI, machine learning, autonomous vehicles, computer graphics, robotics, healthcare, high-performance computing and related fields. Awards are up to $60,000 per student.

Since its start in 2002, the Graduate Fellowship Program has awarded over 220 grants worth around $8 million.

Students must have completed at least their first year of Ph.D.-level studies at the time of application.

The application deadline for the 2027-2028 academic year is October 30, 2026. An in-person internship at an NVIDIA research office preceding the fellowship year is mandatory; eligible candidates must be available for the internship in summer 2027.

For more on eligibility and how to apply, visit the program website.

Featured image shows last year’s NVIDIA Graduate Fellowship recipients.

Building on nearly a decade of co-engineering, CoreWeave has built NVIDIA compute, networking and software into a cloud purpose-built for AI that’s still returning on investment across multiple generations of deployment. Now, CoreWeave is bringing the next generation of NVIDIA infrastructure to production.

At CoreWeave Fully Connected, running this week in San Francisco, CoreWeave announced availability of NVIDIA Vera Rubin NVL72 systems with Spectrum-X 102.4T Ethernet networking. Cognition, the applied AI lab behind the Devin AI software engineer, is the first customer running production workloads on Vera Rubin. 

CoreWeave will also offer NVIDIA Vera, the first CPU built for AI agents. In addition, CoreWeave launched CoreWeave Forge, a connected environment for training, evaluating and improving models and agents on NVIDIA accelerated computing.

“NVIDIA accelerated computing delivers value across generations,” said Ian Buck, vice president of hyperscale and high-performance computing at NVIDIA. “CoreWeave’s NVIDIA V100 GPUs are still running customer workloads nearly a decade after Volta launched, even as CoreWeave brings Vera Rubin NVL72 into production. That’s the strength of the NVIDIA platform: infrastructure that keeps earning for years, and the flexibility to put the right GPU on the right workload.”

Cognition Runs on Vera Rubin NVL72 With 4.8x Higher Token Throughput

Cognition runs training, reinforcement learning and production inference for Devin on CoreWeave. The company scaled to thousands of GPUs on CoreWeave in nine months, powering Cognition inference workloads.

Earlier this month, CoreWeave received its first Vera Rubin NVL72 production racks. Shortly after, Cognition benchmarked Vera Rubin’s inference performance against a GB200 NVL72 baseline using a real-world software engineering workload. To generate this workload, it sampled a subset of tasks from FrontierCode and deployed AI agents to solve them.  

In its early tests, Cognition saw Vera Rubin NVL72 deliver up to a 4.8x increase in total token throughput for SWE-2 inference workloads over GB200 NVL72. For Devin, those gains mean faster real-time code generation and more responsive multistep reasoning.

“Agentic coding is a complex workload: long contexts, high concurrency and token volumes where cost per token decides what we can ship,” said Silas Alberti, founding team at Cognition. “Having all of it on one platform, with NVIDIA and CoreWeave engineers who work the hard problems alongside ours, matters more to us than any single spec.”

CoreWeave Announces NVIDIA Vera Rubin Availability on CoreWeave Cloud

CoreWeave announced availability of NVIDIA Vera Rubin NVL72 on CoreWeave Cloud, making it one of the first cloud providers to deliver the platform in customers’ hands.

Early-access customers can put the performance of NVIDIA’s full-stack AI factory platform to work quickly on CoreWeave Cloud. In days, CoreWeave stood up a production Vera Rubin cluster for Cognition, achieved as a result of the codesign and collaboration between NVIDIA and CoreWeave up and down the stack, from infrastructure to tokens served. 

Capacity can be operated through CoreWeave Kubernetes Service, SUNK, CoreWeave Mission Control, CoreWeave Sandboxes and CoreWeave Inference. 

NVIDIA Vera CPU to Come to CoreWeave Cloud, Tests Show More Than 3x Faster Agentic Sandbox Startups

Agentic AI puts pressure on infrastructure from two directions: serving agents demands low-latency compute at scale, while improving them through post-training requires thousands of isolated environments running at once.

NVIDIA Vera CPU is purpose-built for agentic workloads. For agentic AI, a key performance measure is how many isolated agent environments can run at once and how consistent and performant each one stays as that number grows. 

CoreWeave’s deployment of Vera puts 128 CPUs and 11,264 cores in a single rack, enough for more than 11,000 concurrent environments at one core each. With CoreWeave Sandboxes, these environments are hardware-isolated and run alongside the training jobs they support, with Spectrum-X Ethernet switches and BlueField-4 DPUs ensuring secure, high-performance, secure agent communication at low latency. 

In testing, CoreWeave achieved more than 3x faster agent sandbox startup times on NVIDIA Vera CPUs, accelerating and scaling its sandboxes, an execution layer for reinforcement learning (RL), agent tool use and model evaluation that let AI teams run code in isolated environments on CoreWeave. On Terminal-Bench, CoreWeave saw a 1.7x performance gain on Vera CPU across all passing tasks.

CoreWeave Forge: Closing the AI Loop From Production Back to Training

Models and agents improve by running a loop: production behavior informs the next training run, and each evaluation sharpens the next version. This loop has historically been split across tools from different vendors, with signal lost at every handoff.

CoreWeave Forge unifies Weights & Biases, post-training expertise from OpenPipe and the open source marimo notebook project in one connected environment built for continuous model and agent improvement. It stays open across models, frameworks and clouds.

New and expanded capabilities available include:

• CoreWeave ARIA — now generally available — helps users learn, research, code and iterate across the AI loop, analyzing runs, proposing experiments, recommending code changes and storing them in GitHub, and bringing back actionable evidence that analyzes experiment data, surfaces what drove a change and proposes the next experiments to run.
• CoreWeave Agent Lens — a new service — turns production agent observability into continuous improvement and understandable insights. It improves failure detection by 20% and fixes issues at half of the cost, which turns tens of millions of production agent traces into insights that drive fixes.
• CoreWeave Sandboxes — now generally available — let users run agents, tool calls, RL and evaluations in isolated CPU or GPU execution environments, on serverless infrastructure or on the infrastructure they already train on, providing a fresh, isolated environment for every agent tool call, RL run or evaluation.
• Post-training improves model quality and cuts latency and costs harnessing users’ own production signals, with no training cluster required. Serverless supervised fine-tuning and serverless RL let users experiment with their own training recipes. Serverless RL trains 1.4x faster at 40% lower cost than a self-managed setup.

NVIDIA Dynamo, an open source inference framework for AI factories, powers CoreWeave’s managed inference service as well as RL Rollouts, now in private preview. RL Rollouts loads new checkpoints into a live deployment while it’s running, so reinforcement learning continues without redeploys, and post-training gets the same inference efficiency as production.

Canva, Capital One and MasterClass are among the first companies building on Forge.

NVIDIA Nemotron open models give teams on Forge a direct path to customizing and deploying reasoning and multimodal models for agentic workflows.

Proven Impact From Startups to Global Enterprises

AI labs, AI-natives and global enterprises are using the co-engineered NVIDIA and CoreWeave platform to move from prototype to production faster. 

In healthcare, Ennoble Care, a home-based care provider serving about 50,000 high-need Medicare patients across 15 states, selected CoreWeave to run clinical AI inference. It will use reserved NVIDIA RTX PRO 6000 GPU capacity on CoreWeave Kubernetes Service to scale AI agents for clinical documentation, decision support and back-office automation.

CoreWeave has delivered record MLPerf results in every round of training and inference. It’s the only cloud provider who holds the Platinum ranking in SemiAnalysis ClusterMAX 1.0, 2.0 and 3.0, and serves nine of the 10 leading AI labs.

Together, NVIDIA and CoreWeave are giving customers a platform to turn experimental agents into production systems that write software, support clinicians and do useful work in the real world.

Learn more by attending NVIDIA sessions, demos and workshops at CoreWeave Fully Connected.

When COVID-19 emerged, scientists had a crucial advantage: Decades of prior research on coronaviruses meant they understood the virus’ key proteins well enough to design vaccines in record time. The next pandemic may not offer the same head start. 

To help improve the odds, NVIDIA has joined a coalition of global research organizations, including Google DeepMind and the European Molecular Biology Laboratory’s European Bioinformatics Institute (EMBL-EBI), to release predicted 3D structures for the protein complexes of more than 2,800 viruses — openly available to any scientist, anywhere, through the AlphaFold Database.

The structures in the newly released dataset were inferred using AlphaFold2 — Google DeepMind’s AI model for predicting how proteins fold into 3D shapes — with optimization from NVIDIA BioNeMo Inference Runtime. This allowed the team to scale inference to thousands of viral proteomes, predicting the complexes, or groups of interacting proteins, encoded within each virus. 

“Our ambition with the AlphaFold Database has always been to democratize access to foundational biology at scale,” said Risha Patel, life sciences partnerships manager at Google DeepMind. “This collaboration to bring thousands of viral complexes into the database will equip scientists around the world with insights they need to help prepare for future outbreaks.”

NVIDIA is also openly releasing the BioNeMo Structure Prediction Pipeline, the GPU-accelerated workflow used to generate the dataset, so researchers can go from protein sequence to predicted 3D structure for their own targets.

Preparation for the next pandemic must begin now. An analysis by the Center for Global Development estimates a roughly 50% chance of the world facing a pandemic as severe as COVID-19 by 2050.

“When the next pandemic happens, there may be something that comes out of the blue, and we’ll be lacking the knowledge we had for COVID,” said Joe Grove, professor of molecular virology at the Medical Research Council-University of Glasgow Centre for Virus Research and a collaborator on the project. “What we’re trying to do is stockpile some of that knowledge ahead of time.”

About 30% of the protein interactions being added to the database are completely new to science, showing interaction shapes that have never been documented in the Protein Data Bank, the main repository of experimentally determined protein structures. This translates to new insights for the biological community to explore and harness to generate new knowledge.

“This database is an engine for hypothesis generation,” said Chris Dallago, applied research science team lead in digital biology at NVIDIA. “We’re enabling biologists and the AI community to investigate protein interactions, not just as single molecules but as complexes, so the whole field can move forward.”

Predicting Complex Protein Structures

Most proteins don’t work alone — they come together in complexes of multiple molecules to perform sophisticated functions. Those structures are often what a vaccine or drug must target to disrupt viral function. 

Understanding the 3D structure of the COVID-19 virus’ spike protein, for example, proved foundational to vaccine design. For thousands of other viruses, no such structural knowledge exists today. This dataset begins to fill that gap.

Traditional methods for determining protein structures — crystallizing proteins and shooting X-rays at them — can take years and cost thousands of dollars per structure. AlphaFold2, which was optimized with NVIDIA BioNeMo to efficiently run on NVIDIA GPUs, predicts a structure in minutes and can be run in bulk. Scientists can then verify high-confidence predictions through experimental methods. 

For this project, the team systematically worked through the protein structures of viral families known to infect humans, from common-cold viruses to emerging threats like Mpox. 

A Global Collaboration With Global Access

The collaboration spans the Coalition for Epidemic Preparedness Innovations, EMBL-EBI, Google DeepMind, NVIDIA, Seoul National University, Sungkyunkwan University, the Swiss Institute of Bioinformatics and the University of Glasgow. 

The dataset release — coinciding with a United Nations General Assembly meeting convened by the World Economic Forum on pandemic prevention, preparedness and response taking place this week in New York City — contributes to the AlphaFold Database, which now holds more than 260 million protein and protein complex predictions covering nearly every cataloged protein known to science.

“Making this data open is critical for understanding viral diagnostics and developing treatments and vaccines,” said Jo McEntyre, interim director of EMBL-EBI. “The dataset also covers lesser-studied viruses and lowers the barriers for scientists in low-resource settings who are confronting outbreaks firsthand.”

Predictions in the open dataset are labeled by confidence. The structures show what viral complexes may look like and how individual proteins might interact within a viral proteome. 

Overall, the new data represents a major contribution to the information available for scientists across digital biology and disease research.

“When I did my Ph.D., there were no structures for any of the proteins we were investigating. It was like working in the dark — we had to guess what was going on,” said Grove. “This dataset is a powerful tool for all the researchers doing their Ph.D.s now, giving them high-quality structural data that’s going to accelerate fundamental science.”

Explore the viral protein complex dataset on the AlphaFold Database Pandemic Preparedness Portal, predict structures for protein targets with the BioNeMo Structure Prediction Pipeline, and learn more about NVIDIA BioNeMo.

A warped Manhattan is waiting in the cloud this week. Remedy Entertainment’s CONTROL Resonant brings Dylan Faden’s extraordinary abilities and a paranatural crisis to GeForce NOW at launch.

With the release comes the final days of the CONTROL Resonant Ultimate Membership Bundle. Purchase a 12-month GeForce NOW Ultimate membership through Sunday, Sept. 27, and receive CONTROL Resonant at no additional cost.

GeForce NOW support is coming soon to Googlebooks, a new category of laptops bringing together the best of ChromeOS and Android.

To top it off, DragonSword: Awakening leads nine new games this week, bringing an anime-style open world to supported devices.

Take Control on the Cloud

Explore a warped Manhattan on the brink of paranatural annihilation in Remedy Entertainment’s CONTROL Resonant. Harness Dylan Faden’s extraordinary abilities to battle the Hiss, the Mold and other reality-bending threats while searching for his sister, Federal Bureau of CONTROL Director Jesse Faden.

Dylan must master his new powers and the ever-changing forms of the Aberrant, his paranatural melee weapon, to face a cosmic threat that has escaped the Oldest House and swept through the city.

Ultimate members can stream the next chapter of the CONTROL franchise with GeForce RTX 5080-class performance in the cloud, including NVIDIA DLSS 4, ray tracing, NVIDIA Reflex and Cinematic Quality Streaming technologies, with up to 5K high dynamic range on supported devices. Skip the 100GB local install and expensive PC upgrades, and step into the chaos the moment the mission begins.

The clock is ticking: the CONTROL Resonant Ultimate Membership Bundle ends Sunday, Sept. 27.

Cloud Gaming Meets Googlebooks

GeForce NOW support is coming soon to Googlebooks — Google’s new category of laptops, and the first designed for Gemini Intelligence. Googlebook is crafted with premium materials for powerful performance. Secure by design, its Titan C chip and multigrade protection keep viruses out and hackers away. It also lets users jump from phone to laptop without skipping a beat, and access files and apps from phones directly on the laptop.

As Googlebooks begin rolling out this fall, GeForce NOW support will bring the cloud gaming library to the new laptop category, making it possible to stream the latest PC games with GeForce RTX-powered performance — and without local installs. Keep an eye on GeForce NOW social channels for more availability details later this year.

Awaken the Dragon Sword

A dragon has awakened after 60 years, and the legacy of six heroes waits to be uncovered.

Hound13’s DragonSword: Awakening transports players to the radiant, anime-style continent of Orbis, where Lute’s journey to become the hero known as the Dragon Sword begins. Explore meadows, secret cellars, deep seas, caves and dungeons alongside trusted allies and collectible Familiars.

Build a team from 19 heroes, stack Status Ailments with Active Skills and chain Signal Skills into spectacular tag-team combos. Each character carries a story of their own to uncover along the way.

The journey travels across supported devices, including Steam Decks, Firefox browsers and Amazon Fire TV devices, making players ready for the next adventure in Orbis. The Dragon Sword awaits.

In addition, members can look for the following games joining the cloud this week:

• Dune: Awakening (New release on Xbox, Sept. 22, available on Game Pass)
• Stick It to the Stickman (New release on Steam, Sept. 23)
• CONTROL Resonant (New release on Steam, Sept. 24)
• Wild West Pioneers (New release on Steam, Sept. 25)
• DragonSword: Awakening (Steam)
• Everything Is Crab (Steam)
• LEGO Batman: Legacy of the Dark Knight (Steam)
• Over the Top: WWI (Steam)
• Roadside Research (Steam)

Dates listed above reflect when games are released on their respective stores. GeForce NOW availability may vary, as games are onboarded after they’re released and added throughout the week. Keep an eye on GeForce NOW channels and GFN Thursdays for availability updates to announced titles.

Those new to GeForce NOW can start with a day pass to try premium cloud gaming before committing to a membership. Better yet, the cost of a day pass can be applied toward a first monthly membership — making it easy to level up with GeForce RTX-powered cloud gaming.

Which cloud-bound adventure is calling this weekend? Let us know on X or in the comments below.

When Sakeena Fiza describes her work as a validation engineer at NVIDIA, she does so in terms more befitting a detective story than a world-class engineering lab.

“Validation engineers look in the shadows and shine a light into every corner,” Fiza said. “Every time we get a system, our first thought is: how can it break?”

And when it does? 

“I always like to think of it as a mystery to solve,” she said.

At NVIDIA, the systems Fiza and her colleagues in the data center systems engineering lab investigate are the engines of the AI era. Her work begins before the rest of the world knows a product exists — in the lab — when a new system first receives power.

Components are brought up one by one, boards are integrated, firmware and software teams swarm, and engineers watch for the first signs of life. 

One of Fiza’s earliest and most enduring memories of working at NVIDIA is the collective joy she experienced when she saw the NVIDIA Rubin GPU working for the first time at a system level.

“It literally just said, ‘NVIDIA Corporation Device,’” she recalls. “And everyone’s cheering and celebrating because it’s the first time in the world that a Rubin GPU enumerated at a system level.”

Those moments, electric as they are, are only the beginning. From there, the system must be made resilient: from tray to rack to cluster to production line to customer AI factory. 

Fiza describes validation — the process of ensuring a physical device works correctly before mass production begins — as becoming “the first customers for the product,” exercising hardware to its limits in a range of real-world conditions before anyone else has to depend on it.

“The goal is to always catch issues before customers catch it,” she said. 

Fiza arrived at NVIDIA after earning her bachelor’s degree at the University of California, Irvine, where she studied computer science and engineering. Her path into hardware was the result of an accumulating fascination with systems. 

Growing up in Dubai, she was introduced to coding via the Logo programming language, prompting future forays into systems design that included building Mars rovers at a high school robotics camp and working on unmanned aerial vehicles in college.

What drew her to work with data center systems was the chance to work with the whole machine. At NVIDIA, she said, validation sits at exactly that intersection: firmware, hardware, software, mechanical design, thermal behavior, manufacturing and customer experience.

“I get to be a mechanical engineer when I want to be,” she said. “I get to be an electrical engineer when I want to be. I get to be a firmware engineer when I want to be.”

The failures she chases can be immense or microscopic. A rack-scale issue might involve high-speed signaling, thermal margins or power integrity. Another might come down to a screw tightened too far or the level of dust in a customer facility.

“The solution can be elusive,” Fiza said. “We have to follow the clues, ignore the red herrings and know where to look.”

When a log shows how something failed, Fiza’s job is to discover why. Validation engineers reproduce the issue, vary the conditions, investigate firmware, remove mechanical variables, probe signals, study scope shots and narrow the possible causes.

A single board may contain tens of thousands of components; a rack may approach half a million. Those parts must not merely coexist. They must behave as one system under stress, at scale, in the complex realities of production and deployment across diverse AI factory configurations.

“I wish people understood how complex the hardware is that AI needs to run on,” Fiza said.

For Fiza, the pressure of the work is inseparable from the pleasure of it. Bring-up, she said, is “like the Avengers assembling”: architects, designers, software engineers, firmware engineers, validation engineers, all in the room, racing toward a working system.

“One thing I know when I come to work is I’m never alone,” she said. 

To be a validation engineer is to practice a disciplined kind of suspicion: believe a system can work, then try to conceive of every way it might not. The job requires the doggedness of a great detective, as well as the diagnostic abilities of a general practitioner and the temperament of someone who meets catastrophic failure in the way others might a crossword.

Each project brings a new puzzle, a new failure mode and, in turn, a chance to make the next system better.

“With the products we have in the pipeline, I’m so excited,” Fiza said. “They’re going to change the world.”

NVIDIA AI Day Singapore, which takes place Sept. 22-23 at the Raffles City Convention Centre, is offering attendees opportunities to explore the hands-on training, expert-led sessions and advanced tools to accelerate their work in AI and high-performance computing.

At the event, NVIDIA and its partners are showcasing breakthrough AI advancements across the Southeast Asia region at large.

Read more about these announcements below.

NVIDIA Accelerates Public Sector AI from Pilot to Production in Southeast Asia

AI is becoming a matter of national strategy, with governments looking to move from pilots to production and deliver impact at scale, while building trusted AI capabilities that reflect local languages, cultures, priorities and economic needs. 

NVIDIA is working to enable all nations to be AI nations — providing the technology, infrastructure, ecosystem and expertise needed to make this possible.

To accelerate this transition across Southeast Asia, NVIDIA is helping nations move AI from experimentation to production-scale deployment through open models, developer tools and a broad partner ecosystem. 

Together, NVIDIA and its partners are focusing on four key areas: 

• Enhancing government operations and service delivery.
• Developing accessible AI-powered citizen services, and empowering local businesses.
• Strengthening critical infrastructure and public safety.
• Supporting startups, developers and researchers to strengthen national AI capabilities and innovation in each country. 

Singapore’s HTX (Home Team Science and Technology Agency) is embarking on research using the NVIDIA Nemotron 3 Super and Nemotron 3 Nano Omni models to advance AI for public safety. Nemotron Super has the potential to support the agency’s complex reasoning and agentic workflows, while Omni’s unified vision, audio and language capabilities could help HTX develop multimodal applications grounded in real-world operational data. Together, the models could strengthen HTX’s ability to deploy secure, locally controlled AI across Singapore’s Home Team.

NCS is advancing agentic AI adoption across enterprises and the public sector, using Nemotron models and the NVIDIA Blueprint for video search and summarization (VSS), while advancing physical AI for practical humanoid robotics applications, to address security, responsiveness and data governance requirements. ST Engineering is using NVIDIA NeMo tools and NVIDIA cuOpt software to develop its AI Studio platform and deploy agentic AI solutions across its businesses such as Marine MRO.

Beyond Singapore, similar work is already underway across the region. Malaysia’s YTL AI Labs is fine-tuning Nemotron models for enterprise and citizen services, while Viettel AI is doing the same for Vietnamese-language applications. 

In Thailand, the Big Data Institute and iApp Technology, as members of the ThaiLLM Collaboration, are exploring Nemotron as a foundation model. With an initial focus on legal applications, iApp Technology is adapting Nemotron 3 Nano by fine-tuning OpenThai 2.0 Legal with Thai-language legal data using the NVIDIA NeMo framework. 

The model is released as open source for the Thai developer community and serves as the engine for Thanoy, the company’s legal-assistant chatbot, which already serves approximately 43,000 users. 

In Brunei, Antrique built an AI innovation platform to help boost productivity across the nation’s food sector.  

Across the region, NVIDIA Cosmos open world models and the NVIDIA VSS Blueprint are advancing smart city solution development. Malaysia’s ITMAX uses Cosmos with VSS to improve city traffic operations, while Thailand’s AS-TECH applies the same stack to improve passenger flow in airports. 

Learn more about NVIDIA Nemotron and Cosmos models and read about NVIDIA’s participation in the Open Secure AI Alliance.

Southeast Asia Technology Leaders Build With NVIDIA Nemotron Open Models for Region-Specialized AI

Leading enterprises, technology providers and research organizations across Southeast Asia are building region-specialized AI models and applications with NVIDIA Nemotron open models, datasets and libraries — accelerating the development of AI tailored to the region’s languages, industries and communities.

NVIDIA Nemotron provides a foundation for regional AI ecosystems, letting organizations customize, control and own models that address their specific requirements. Nemotron also offers persona datasets that provide locally relevant synthetic data reflecting the region’s populations, languages and workforces.

Across the region, partners are building applications spanning public services, services and healthcare.

NVIDIA Nemotron Adoption Expands in Singapore

Enterprises in Singapore are adopting NVIDIA Nemotron for various use cases. AI Singapore is expanding its SEA-LION model family to include the NVIDIA Nemotron open models and NVIDIA NeMo tools. SEA-LION is an open model family designed for Southeast Asian languages and cultures.

Hummingbird Bioscience, together with LynxKite, is building an explainable Toxicity Knowledge Graph powered by Nemotron 3.5 Lightning and NeMo Retriever with in silico simulations. The collaboration aims to integrate complex public and proprietary data across diverse third-party file formats, creating a comprehensive, unified foundation for robust analysis and reasoning that helps de-risk and accelerate drug discovery and development.

Across Asia Pacific, NVIDIA Nemotron Enables Region-Specific AI

Bitdeer AI co-hosted the Open Models AI Codefest with NVIDIA, providing the GPU cloud infrastructure that enabled developers across the region to use NVIDIA Nemotron open models, datasets and training recipes to accelerate localized applications across critical sectors, including healthcare. 

In Vietnam, Viettel AI has been extensively fine-tuning Nemotron 3 Super for the Vietnamese language and agentic applications. The model achieved the highest ranking on both the VMLU benchmark and the company’s in-house product benchmark, and it’s set to be adopted in Legal AI — an agent harness that will serve both internal Viettel Group employees and external customers.

Also in Vietnam, FPT Smart Cloud codeveloped Nemotron-Personas-Vietnam, an open dataset grounded in Vietnamese demographic and cultural data, and is enabling local developers to post-train and evaluate localized AI models.

Get started building with NVIDIA Nemotron using skills and playbooks that help partners customize Nemotron open models for their languages and domains.

Stay up to date on agentic AI, NVIDIA Nemotron and more by subscribing to NVIDIA news, joining the community and following NVIDIA AI on LinkedIn, Instagram, X and Facebook.  

Explore self-paced video tutorials and livestreams.

Sea the First in ASEAN Region to Adopt NVIDIA Vera Rubin, Scaling AI to Better Serve Communities Across Southeast Asia

Sea Limited, a global technology company founded in Singapore, is the first enterprise in the ASEAN region to adopt the NVIDIA Vera Rubin platform, further strengthening the company’s AI capabilities to better serve and create meaningful economic opportunities for millions of consumers and small businesses across Southeast Asia. 

Serving hundreds of millions of users through its Garena, Monee and Shopee platforms, Sea has already deployed AI across its businesses to make its services more useful and accessible. Now, with NVIDIA Vera Rubin, Sea will build on these efforts, developing and deploying AI models and intelligent agents at greater scale to serve the evolving needs of its communities. 

On Shopee, AI is already helping sellers reduce the time and effort required to create informative product listings, improve product discovery and deepen customer engagement, while enabling better-informed business decisions. These capabilities enable small- and medium-sized enterprises in Southeast Asia, many of which operate with limited resources, to scale their businesses using enterprise-grade AI technologies previously accessible only to large corporations.

Across Monee, the digital financial services division of Sea, AI is being applied in areas such as fraud detection and credit risk assessment, supporting Monee’s ability to deliver simple, accessible and inclusive digital financial services. For small businesses and consumers underserved by traditional financial services, these capabilities can expansively broaden access to financial tools.       

At Garena, Sea’s digital entertainment and video game arm, AI is used to enhance gaming experiences supporting the company’s efforts to create engaging, inclusive and safe online spaces that bring players together.

NVIDIA Vera Rubin will provide the advanced computing infrastructure to build on this foundation — enabling Sea to accelerate innovation, scale AI applications more broadly and deepen its impact for the communities it serves.

Learn more about NVIDIA Vera Rubin.

To build and deploy sophisticated robotics applications that can perceive, reason and act in dynamic environments, developers need new physical AI models and tools.

The ROS open framework is a project from Open Robotics that helps humans build robots. NVIDIA Isaac ROS 5.0 — a collection of GPU-accelerated packages built on ROS, released today at the ROSCon conference in Toronto, Canada — helps humans and AI agents build robots together.

The release introduces new agentic workflows and platform support to help developers build, customize and deploy robotics applications faster.

ROS provides the open source foundation for much of modern robotics development, giving developers common tools, libraries and standards for building and connecting robot applications. 

NVIDIA Isaac ROS brings NVIDIA accelerated computing, physical AI models and production-ready libraries to the nearly 1.3 million ROS users, helping developers build high-performance robotics applications using free, familiar, open source tools.

Bringing AI Agents Into Robotics Development

AI agents are changing how software is built, helping developers automate repetitive tasks, navigate complex codebases and move from ideas to working applications faster. Isaac ROS 5.0 brings these capabilities to robotics development.

Isaac ROS 5.0 introduces support for ROS Lyrical and Ubuntu 24.04, giving developers a path to adopt the latest ROS platform while continuing to accelerate demanding robotics workloads with NVIDIA accelerated computing. NVIDIA worked with the Open Source Robotics Alliance to contribute a standard data-handling interface to ROS Lyrical that helps robotics software work efficiently across different computing hardware, including GPUs. 

Available to the entire ROS community, it gives developers a consistent way to accelerate demanding robotics applications, with CUDA providing a working example for GPU acceleration.

New NVIDIA Isaac skills for setup and manipulation provide reusable workflows that developers and AI agents can use to complete robotics development tasks. Agent-ready documentation also makes it easier for AI agents to understand Isaac ROS tools and workflows, turning developer intent into working applications faster.

Some skills go beyond assisting with individual coding tasks. A new FoundationStereo fine-tuning skill enables an AI agent to help adapt a stereo perception model to a developer’s cameras, environment and robotics application, so developers can easily achieve more accurate perception for a given sensor configuration.

FoundationPose, a foundation model for object pose estimation and tracking, now provides an agent-ready inference library that enables robots to perceive and track the position and orientation of objects up to 5.5x faster. 

In addition, pick and place — a common workflow that connects detection, depth estimation and pose output — is now available as a standalone, agent-ready skill, providing robot developers more flexibility beyond Isaac ROS.

Accelerating the Open Source Robotics Ecosystem

The robotics ecosystem is already extending this agentic approach to development workflows. 

AgenticROS, an open source project sponsored by 3D perception technology company RealSense, connects Isaac ROS with NVIDIA Nemotron open models and NVIDIA NemoClaw blueprints, enabling AI agents to interact with ROS-based robots. RealSense is also optimizing its latest AI-native 3D stereo depth cameras, including RealSense D585 Pro, and an open source software development kit for Isaac ROS and the NVIDIA Jetson Thor edge AI platform, helping developers build perception, navigation and manipulation applications.

Intrinsic’s Open Machine Tending Solution is a reference application for computer numerical control machine tending, part of the newly released Intrinsic Core, an open source suite of preconfigured runtime services and capabilities designed to accelerate industrial robotics applications. It includes built-in compatibility with NVIDIA FoundationPose for out-of-the-box object registration, tracking and pose estimation. Using the FoundationPose perception pipeline, the solution enables robots to dynamically detect and handle parts while reducing the need for rigid, costly physical fixtures and specialized systems integration.

Seeed Studio is using NVIDIA Isaac ROS with reBot Arm, combining accelerated perception, spatial understanding and motion planning on NVIDIA Jetson Thor. This integration gives developers a practical platform for building adaptable physical AI applications, from object localization to collision-aware manipulation and autonomous pick and place.

Magna is using NVIDIA Isaac ROS as a modular, GPU-accelerated foundation for robotic perception, synchronized data collection and NVIDIA Isaac GR00T model deployment, pairing it with Isaac Sim hardware-in-the-loop testing to bring intelligent automation from research to real-world manufacturing and mobility — faster and with fewer risks.

Prefix.dev’s Pixi package-management tool makes it easier to create reproducible robot development environments, bringing together ROS with the NVIDIA CUDA platform to help developers more easily set up and share accelerated robotics workflows. 

As an Isaac ROS Partner, Foxglove helps developers visualize and debug live ROS applications through its web and desktop tools, which are integrated throughout Isaac ROS tutorials and support data such as 3D topics, nvblox meshes and rosbags.

Flexiv is integrating Isaac ROS with its Rizon 4 adaptive robot, giving developers access to NVIDIA-accelerated robotics capabilities and a streamlined path from testing applications in NVIDIA Isaac Sim to deploying them on a physical robot.

Ekumen, a Grid Dynamics Company, is using GPU-accelerated Isaac ROS packages within existing ROS and Nav2 stacks to improve precision docking, 3D obstacle detection, visual localization and real-time motion planning, validating each application in Isaac Sim.

Ekumen uses isaac_ros_cumotion on a GPU to map a collision-free path for a warehouse arm in roughly 2 to 5 milliseconds.

Ouster integrates its Stereolabs ZED stereo cameras with NVIDIA Isaac ROS to deliver GPU-accelerated perception for robotics applications. The integration simplifies the development of real-time object detection, mapping and navigation while maintaining interoperability with the broader ROS ecosystem.

Bringing the Complete Physical AI Stack to the Robot

The applications that developers and agents build ultimately need to run on the robot.

NVIDIA Jetson is a scalable computing platform for running the physical AI stack at the edge with real-time performance, bringing together ROS, accelerated perception and navigation, AI models and application logic on the robot.

Isaac ROS 5.0 supports scalable compute, from entry-level NVIDIA Jetson Orin Nano to high-performance Jetson Thor devices, giving developers a path from development to deployment as robotics workloads become increasingly sophisticated.

Robotics companies are already using this combination to bring more AI processing directly onto their machines.

Mentee Robotics uses NVIDIA Isaac ROS as the perception and AI backbone of its MenteeBot humanoid, enabling the robot to interpret visual information and execute learned behaviors in real time. A shared software foundation across NVIDIA Jetson Orin and Jetson Thor platforms helps Mentee extend its innovations from existing robots to next-generation systems.

Universal Robots has built NVIDIA Isaac ROS into its AI Accelerator software development kit to help integrators deploy advanced perception and motion capabilities faster, without developing complex robotics software from scratch. Powered by NVIDIA Jetson at the edge, the solution enables robots to adapt to parts that are not precisely positioned, reducing reliance on costly fixtures and making manufacturing cells more flexible.

ROBOTIS, which builds the developer-friendly ROS-based TurtleBot3, is integrating Isaac ROS into its AI Worker robot, using GPU-accelerated object perception to enable vision-guided manipulation tasks including picking, placing and alignment.

ROBOTIS performs object manipulation tasks using NVIDIA Isaac ROS CuMotion.

FieldAI’s robot foundation models, which can run entirely on robots without relying on cloud connectivity, are integrating Isaac ROS on Jetson devices to take greater advantage of GPU acceleration and improve the efficiency of the on-robot AI stack.

Noble Machines is using NVIDIA Isaac ROS on Jetson to accelerate the development of general-purpose robots for industrial applications, building on ready-to-use AI and perception capabilities rather than creating them from scratch.

By combining an open robotics ecosystem, accelerated computing and new agentic development workflows, Isaac ROS 5.0 helps developers address both sides of the physical AI challenge: building increasingly capable robot applications and efficiently running them in the physical world.

Available now, Isaac ROS 5.0 is free and open source. Developers can learn more and get started with NVIDIA Isaac ROS on GitHub.

Every AI factory needs power and cooling that fit its computing architecture. As AI infrastructure expands, power, cooling, water, site and grid constraints are shaping what builders can deploy. Choosing products that fit the complete factory design helps builders turn computing capacity into useful AI output.

To help builders make those decisions, NVIDIA is introducing NVIDIA DSX Ready, a qualification program for partner products and solutions that meet applicable NVIDIA DSX AI factory reference design requirements. 

The program launches with two initial categories: battery energy storage systems (BESS) and cooling distribution units (CDUs). Category-specific requirements and review through the program help builders evaluate offerings with greater confidence, reduce integration risk and move toward deployment.

Qualified Building Blocks for Building AI Factory

The NVIDIA DSX AI factory platform unifies AI factory design and operations across compute, networking, power, cooling, facilities and software. It helps partners design and operate the factory as one system to produce more useful AI output within available power, cooling, water and grid constraints.

That system view matters now because optimizing one part of an AI factory can shift the bottleneck elsewhere. Power and cooling suppliers need a clear path from reference designs to qualified offerings, and builders need a clearer way to discover and evaluate those offerings.

DSX Ready connects that selection process to applicable NVIDIA DSX requirements. It gives builders a clear qualification to look for and gives partners a defined way to demonstrate that a specific offering meets the requirements for its category.

Power and Cooling Qualification

At launch, DSX Ready includes qualified BESS solutions Hitachi Energy, LG Energy Solution and Tesla, and qualified CDU solutions from LG Electronics, LiquidStack and Vertiv. These initial categories bring power systems and liquid-cooling infrastructure into a common program, with requirements tailored to each technology. Additional categories across infrastructure and software will be rolled out over time.

For BESS providers, partners run the required qualification tests and submit supporting data for NVIDIA review and approval within a defined qualification boundary. Passing qualification does not replace site-level engineering or imply site-level stability.

For CDU providers, the path uses the CDU self-qualification suite to determine whether a specific offering meets applicable NVIDIA functional requirements.

Teams can then focus on how qualified offerings fit their site, configuration and operating needs. A qualified CDU may meet the relevant cooling criteria, for example, while the builder still evaluates how it will fit the planned facility.

From Platform Design to Partner Selection

The value of a reference design grows when builders can connect it to specific products and informed engineering decisions. DSX Ready makes that connection, bringing partner innovation into the infrastructure choices behind NVIDIA DSX AI factories.

Explore NVIDIA DSX Ready qualification categories and learn how to participate. Connect with the right NVIDIA team to begin qualification for a product or solution.

Physical AI is moving rapidly from research to large-scale deployment. By 2035, ABI Research projects an installed base of 49 million level 3-5 autonomous vehicles (AVs), while Omdia estimates that roughly 60 million industrial robots will be deployed between 2026 and 2035. As these machines enter roads, factories, warehouses and other environments shared with people, safety must scale with them.

Physical AI safety means proving that AI-driven machines — AVs, humanoid robots, industrial robots and more — behave safely when their decisions turn into physical action. That requires safety across the hardware, software, AI, operating environment and deployment lifecycle — not a one-time check before deployment.

Why Is Safety the Key to Scaling Physical AI?

After years of testing and benchmarking, AVs continue to expand commercially. That progress has required developers to demonstrate how automated systems address potential hardware and software failures, limitations in intended functionality and AI-specific risks.

Robotics is approaching a similar inflection point as autonomous machines move into factories, warehouses and other environments shared with people.

Across physical AI, manufacturers, regulators, insurers and workplace safety teams need evidence that hardware, software, AI behavior and operating environments can work together safely without human intervention.

Why Does Physical AI Need a New Safety Model?

Four shifts define new safety standards:

• Dynamic environments require context-aware safety. Roads, factories and warehouses cannot be fully controlled through static zones or physical barriers. Autonomous systems must perceive changing conditions, adapt their behavior and reach a safe state when something unexpected occurs.
• AI behavior requires its own assurance. Testing must assess AI software alongside traditional functional safety, using design-time, runtime and validation-time guardrails. Emerging standards such as ISO/IEC TS 22440 are beginning to address these AI-specific risks.
• Deployment is ongoing. AVs and robots evolve through software and model updates, new tasks and changing operating conditions. Material changes may require additional safety testing.
• Validation at scale requires simulation and synthetic data. The number and complexity of potential scenarios requires real-world testing to be combined with simulation, synthetic data generation and scenario reconstruction.

Together, these shifts require safety to be operationalized across design, deployment and validation, from the underlying hardware to AI behavior and the operating environment. 

What Safety Foundation Has NVIDIA Built for Physical AI?

Physical AI safety requires specialized engineering, data, processes and validation that few companies can reproduce alone. NVIDIA’s safety foundation draws on more than a decade of development in AV safety, building expertise in functional safety, sensor fusion, AI behavior assurance, vision AI, simulation and real-world validation.

NVIDIA Halos is the first and only full-stack safety system for physical AI, helping developers engineer safety across every layer of design, validation and deployment. The principles are shared across AVs and robotics, while the platforms, standards and evidence remain specific to each domain.

For AV development, Halos spans:

• Hardware: NVIDIA DRIVE AGX Thor provides safety-engineered accelerated compute, while NVIDIA Hyperion provides the full-stack vehicle platform and reference architecture for level 4 AVs.
• Operating system and middleware: Halos OS provides a unified software foundation built on ASIL-D certified DriveOS. Halos Core and Halos Middleware support system isolation, monitoring and deterministic communication.
• End-to-end model: NVIDIA Alpamayo offers open reasoning vision language action models that bring explainability to long-tail scenarios.
• Simulation and validation: The NVIDIA Halos Safety Evaluation Framework provides tools and guidelines for generating evidence to support AV safety cases across different levels of automation.

Together, these elements connect cloud-based AI development and simulation with in-vehicle deployment so safety evidence can remain traceable across the vehicle lifecycle.

For robotics, Halos spans:

• Hardware: NVIDIA IGX Thor is an industrial-grade module that combines accelerated computing and functional safety on one platform with a dedicated Functional Safety Island. It’s designed to support systems developed for standards including IEC 61508 and ISO 13849.
• Software: Halos Core for IGX provides the software foundation for safety-related operating functions, including fault detection,  monitoring and reporting, along with the communication and processing capabilities that connects sensors, actuators and other safety components 
• Real-time sensing: NVIDIA Holoscan Sensor Bridge connects sensor data with AI and safety-related processing, helping systems identify invalid information and execute defined safety responses.
• Simulation and validation: NVIDIA Isaac Lab and NVIDIA Omniverse libraries let developers test robot behavior across relevant conditions and edge cases, complementing real-world validation.

• Outside-in safety: The open source NVIDIA Halos Outside-In Safety Blueprint uses external cameras and vision AI agents to extend awareness beyond onboard sensors and support facility-level monitoring and functional safety use cases.

Across both AV and robotics, the NVIDIA Halos AI Systems Inspection Lab turns safety, cybersecurity and AI safety requirements into repeatable inspections and helps prepare Halos integrations for final system-level certification by third-party agencies.

Who Is Building With the NVIDIA Halos Safety Ecosystem?

NVIDIA Halos connects the companies that build, integrate, assess and deploy physical AI solutions, including product developers, software and embedded-system providers, sensor and silicon companies, safety solution developers and certification bodies.

In autonomous vehicles, Geely, Isuzu, Nissan (powered by Wayve software) and Einride are building level 4-ready vehicles on NVIDIA Hyperion, supported by Halos OS.

Uber, Grab, Lyft and other mobility providers are also using Hyperion to scale robotaxi development and deployment. Members of the NVIDIA Halos AI Systems Inspection Lab include AUMOVIO, Bosch, Gatik, Hesai, Lucid, MIRA, onsemi, PlusAI, Sony, Valeo and Wayve, spanning autonomous-driving development, ADAS, sensors, silicon, systems integration, validation and safety assurance.

In robotics, acontis and QNX provide the embedded software needed to run safety functions predictably, while Advantech and NexCOBOT build safety-designed NVIDIA IGX systems. Infineon, NXP, STMicroelectronics and Texas Instruments contribute sensor, safety-microcontroller and other semiconductor technologies. KION Group is developing functional safety agents for autonomous forklifts. Agility is integrating NVIDIA IGX Thor and Halos Core into the safety system for its Digit 5 humanoid.

How Is NVIDIA Halos Independently Assessed?

For AVs, TÜV SÜD certified NVIDIA’s Automotive Product Lifecycle software process and DriveOS 6.0 to ISO 26262 ASIL D, as well as NVIDIA’s automotive engineering processes to ISO/SAE 21434. TÜV Rheinland also performed an independent UNECE safety assessment of NVIDIA DRIVE AV.

For robotics, TÜV Rheinland is inspecting NVIDIA IGX Thor, Halos OS and Holoscan Sensor Bridge for functional-safety certification readiness, building on TÜV SÜD’s inspection of the Thor SoC and Halos Core for ISO 26262.

Across physical AI, ANAB has accredited the NVIDIA Halos AI Systems Inspection Lab as an ISO/IEC 17020 inspection body. The lab inspects scoped Halos integrations and helps companies prepare for final certification by independent third-party bodies.

The companies that scale physical AI will not simply build the most capable systems. They will build systems that can be assessed, certified, deployed and trusted in the real world. Designing functional safety from the start is what separates a prototype from a scalable solution.

Learn more about NVIDIA Halos for AVs and robotics, and explore the full-stack safety architecture for physical AI.