Pi-hole, a network-wide DNS sinkhole, filters ads and tracker domains before any device connects. Unbound, a validating recursive resolver maintained by NLnet Labs, handles the remaining queries by walking the DNS hierarchy from root servers down to authoritative nameservers, with no public resolver involved. DNSSEC (Domain Name System Security Extensions) validation comes built into Unbound’s configuration.

The Lookup Your Router Sends to Google

Every DNS lookup carries three pieces of information the resolver can see: the domain name you queried, the time of the request, and the IP address that made it. Every device on your network that resolves a hostname contributes to that stream: phones, laptops, smart TVs, thermostats. Multiply it across a typical home and the resolver holds a detailed, timestamped record of your household’s internet activity.

According to Google’s DNS privacy documentation, temporary logs, which include your IP address and your query, are deleted within 24 to 48 hours, and ISP-level and city-level location data is retained longer for service improvement. Google also states it doesn’t correlate DNS log data with your Google account information. The data still passes through Google’s infrastructure, and the policy reflects what Google chooses to collect today.

Switching to Cloudflare’s 1.1.1.1 shifts that trust to Cloudflare, which commits to no durable logging backed by independent KPMG audits. Switching to your ISP’s resolver puts the data with your ISP, which in the United States has had legal permission to sell customer browsing data since a 2017 Congressional act. None of these changes removes a third party; they change which third party holds the data.

Encrypting DNS queries with DoH (DNS-over-HTTPS) or DoT (DNS-over-TLS) stops network observers from reading those queries in transit, but the resolver itself still sees and processes every request. Encryption protects the channel, not the endpoint.

Encryption protects the channel, not the endpoint.

How Pi-hole and Unbound Divide the Work

Pi-hole Blocks Before You Connect

Pi-hole installs on the Raspberry Pi and becomes the DNS server for your entire home network. You configure it by pointing your router’s DNS setting to the Pi’s static IP address. When any device asks for a domain, that request hits Pi-hole first.

Pi-hole checks the domain against its blocklists, maintained by the community and covering advertising networks, analytics trackers, and known malware hosts. If a match is found, Pi-hole returns 0.0.0.0, a dead-end address, and the device never makes the connection. The filtering runs at the network layer, so it applies to every device without per-device software installation. A smart TV running its own ad client, an IoT thermostat, a phone with in-app ad calls: all get the same treatment the moment they connect to your Wi-Fi.

Unbound Resolves Without a Third Party

Non-blocked queries get forwarded to Unbound, which runs on the same Raspberry Pi at 127.0.0.1, port 5335. Unbound doesn’t forward those queries further. As the Pi-hole project’s Unbound configuration guide describes, it walks the DNS hierarchy directly: contacting root servers, getting referred to the appropriate TLD (top-level domain) servers, and then querying the authoritative nameservers for the specific domain. Every answer comes from the actual DNS infrastructure, no middleman required.

Results are cached locally. The first lookup for a domain Unbound hasn’t seen can take hundreds of milliseconds, compared to single-digit milliseconds from a warm public resolver with global caches. Subsequent visits pull quickly from Unbound’s local store. NLnet Labs, which maintains Unbound, describes the resolver as designed to be fast and lean while incorporating modern open standards, and query name minimisation is among the privacy features built in by default.

Poisoned Records and the DNSSEC Fix

DNSSEC adds cryptographic signing to DNS records. When Unbound validates those signatures, it confirms that what it received from an authoritative server wasn’t altered in transit. If a signature fails, Unbound rejects the response and the connection fails cleanly rather than landing on a malicious clone.

DNS cache poisoning is the attack DNSSEC directly defends against. An attacker who intercepts communication between a resolver and an authoritative server can insert a forged response, rerouting traffic to a malicious clone without any visible warning to the user. Security researcher Dan Kaminsky’s demonstration of practical cache poisoning at scale is credited as the primary factor behind DNSSEC deployment, showing that standard DNS had no mechanism to verify whether a response came from the legitimate server. The stakes of DNS-level domain attacks are illustrated by a 2016 operation against a Brazilian bank: attackers who seized all 36 of the institution’s domains for about five hours routed roughly 5 million customers to convincing copies of its online portals, ATMs, and point-of-sale systems. That attack involved registrar compromise rather than cache poisoning, a different vector, but it shows what domain-level failures can produce at scale.

Google’s 8.8.8.8 also offers DNSSEC validation, so this check isn’t exclusive to self-hosted setups. Security researchers citing ICANN data put the share of internet users currently protected by DNSSEC validation at less than 30% globally, meaning most households using standard ISP-provided DNS receive no cryptographic guarantee on the records they get.

In the Pi-hole and Unbound configuration, Pi-hole’s own DNSSEC setting is deliberately disabled: that job belongs to Unbound, listening at port 5335. Running both simultaneously creates validation conflicts.

The Setup Has Real Limits

A few tradeoffs don’t get adequate space in most how-to guides:

• Single point of failure. The Raspberry Pi is now your home’s only DNS resolver. A power outage, SD card failure, or physical disruption takes your entire network offline until the Pi recovers, or until you restore a public DNS address in your router settings.
• Cold-start latency. Unbound walks the full DNS tree the first time it encounters any domain. That initial lookup takes hundreds of milliseconds. Repeat visits are near-instant from cache, but the delay on first contact is noticeable, especially on pages that load dozens of third-party hostnames.
• ISP visibility persists. DNS privacy closes one window. Your ISP can still observe the IP addresses your devices connect to, which maps back to domains with reasonable accuracy. A VPN addresses this but sits outside this setup’s scope.
• LAN queries stay unencrypted. Queries from your devices to Pi-hole travel unencrypted across your local network. Unbound’s outbound queries to root and authoritative servers are also plain DNS by default. Encrypting client-to-resolver traffic requires additional configuration on both ends.

The Pi-hole project’s current major release, Pi-hole v6, rewrote the FTL (Faster Than Light) engine and shipped a modernized web interface, addressing several longstanding usability complaints. Software updates and blocklist refreshes still need periodic manual attention to stay current.

A Quiet Drain on Ad-Tech’s Data Pipeline

Pi-hole started in 2014 as a project by Jacob Salmela, conceived as an open-source alternative to a hardware ad-blocker called AdTrap. It now runs on ARM single-board computers, x86 home servers, and inside Docker containers, and it has drawn substantial contributor communities to its blocklists and codebase.

The category has grown around it. AdGuard Home ships native DNS-over-HTTPS support without additional configuration. NextDNS offers a cloud-hosted filter with a free tier capped at 300,000 monthly queries. Of the three, Pi-hole’s community-maintained blocklist ecosystem remains the largest, and that depth is what has kept the project competitive as alternatives have added features.

Browser extensions protect only what passes through a single application. Pi-hole works at the DNS layer, reaching every device that connects through your router regardless of operating system, browser, or installed app. A smart TV running its own ad client, an Android phone’s in-app tracker calls, firmware telemetry from IoT sensors: all resolve through DNS and hit the same blocklist filter.

The advertising industry has partially adapted. YouTube and Twitch moved to first-party ad delivery, serving ads from the same domain as the content, which sidesteps DNS-based blocking entirely. Third-party analytics calls, tracker pixels, and telemetry from installed apps still resolve through separate hostnames and land on blocklists. Users in the Pi-hole community describe the setup’s current primary value as controlling the steady, low-level data exhaust that modern software produces, with ad-blocking as a secondary benefit.

Frequently Asked Questions

Does Pi-hole with Unbound Hide My Browsing from My ISP?

Partially. The setup stops DNS queries from reaching your ISP’s or Google’s resolvers, removing that layer of domain-level visibility. Your ISP can still observe the IP addresses your devices connect to, and those addresses map to domains with reasonable accuracy. A VPN running alongside this setup addresses that gap.

Do I Need a Raspberry Pi, or Will Other Hardware Work?

No specific hardware is required. Pi-hole installs on most Linux distributions, including Debian, Ubuntu, and Alpine Linux. The Raspberry Pi Zero 2 W, with its quad-core ARM Cortex-A53 processor and 512MB of RAM, is a common choice for its low power draw and price under $20, but any always-on Linux device with a static IP address on your network will work.

Will Pi-hole Break Any Websites or Apps?

Some sites detect DNS-based filtering and refuse to load; others break silently. Pi-hole’s web dashboard shows every blocked domain by device, so identifying and whitelisting a false positive takes a few clicks. Streaming services that serve ads from first-party domains can’t be filtered at the DNS layer without also blocking the content itself.

What Happens If the Raspberry Pi Goes Offline?

Your home network loses DNS resolution and, effectively, internet access until the Pi recovers or you restore a public resolver in your router. Most routers support a secondary DNS address; setting a backup public resolver there means devices switch over automatically when the Pi is unreachable.

Does This Setup Protect Mobile Devices Without Installing Apps?

Yes, for any device connected to your home Wi-Fi. The filtering runs at the network level, so phones, tablets, smart TVs, and game consoles all get Pi-hole’s filtering with no app, profile, or VPN configuration required. Devices on cellular data bypass it entirely.

Does Unbound’s DNSSEC Validation Catch All DNS Attacks?

No. DNSSEC confirms that DNS records weren’t modified between authoritative servers and your resolver, but it doesn’t cover registrar-level compromise, BGP rerouting, or DNS hijacking above the resolver. It also provides no encryption of DNS queries in transit. The protection is specific: cryptographic verification of record authenticity, which defends against cache poisoning attacks.

Pi-hole v6, the project’s current major release, updated the core FTL engine and shipped a new web dashboard. The official setup guide for pairing it with Unbound is maintained at docs.pi-hole.net alongside each release.

A Pew Research Center survey of 3,507 U.S. adults conducted in late March found 60 percent of American adults view Israel unfavorably, a nearly 20-point increase since 2022. The committee voted the provision through anyway.

A New Framework in a Must-Pass Bill

The provision was introduced May 26, 2026, as part of the HASC chairman’s mark for the FY2027 NDAA, a $1.15 trillion piece of legislation that serves as the annual blueprint for American military spending and policy. Its language derives from the U.S.-Israel FUTURES Act (United States-Israel Framework for Upgraded Technologies, Unified Research, and Enhanced Security Act), a bill co-sponsored by Rep. Ronny Jackson (R-TX) and Sens. Ted Budd and Kirsten Gillibrand that never passed through regular congressional order. Embedding it in the NDAA grants it statutory permanence that an interagency memorandum or stand-alone pilot program would not carry.

The provision requires the Secretary of Defense to designate an executive agent responsible for synchronizing bilateral cooperation across defense technology research, development, testing, evaluation, integration, and industrial partnership. That agent operates under the Secretary’s authority without requiring Senate confirmation, reducing the number of choke points where a future Congress could intervene.

The FY2027 NDAA as written allocates $750 million in U.S.-Israel cooperative programs, according to AIPAC’s factsheet on this NDAA cooperative funding initiative, broken down across four areas:

• $500 million for U.S.-Israel Missile Defense Cooperation
• $100 million for Counter-Unmanned Systems (C-UxS) cooperation
• $100 million for Subterranean Operations cooperation
• $50 million for Emerging Technologies cooperation, including artificial intelligence, quantum machine learning, directed energy, and biotechnology

The New Republic reported that no other U.S. ally holds a dedicated executive-agent structure of this kind within the Department of Defense, including any NATO partner or AUKUS members the UK and Australia.

Beyond the Annual Aid Check

Since fiscal year 2019, U.S. military support for Israel has run through a 10-year memorandum of understanding (MOU), negotiated in 2016 during the Obama administration, that provides $3.3 billion annually in Foreign Military Financing plus $500 million per year in missile defense funding. The current MOU expires at the end of FY2028. That model is visible: Congress can debate it, the dollar amounts are public, and lawmakers who want to place conditions on aid have a clear procedural vehicle to try.

Ben Freeman, director of the Democratizing Foreign Policy program at the Quincy Institute, argues the provision shifts the relationship into a different channel: Pentagon procurement, licensing agreements, co-production arrangements, and research pipelines where congressional oversight is limited and public accountability is minimal. In a brief titled “The Disappearing Aid Check,” Quincy Institute senior fellow Steven Simon put the structural argument plainly: moving support away from foreign-aid votes and into defense acquisition means it no longer appears on the political scoreboard where constituents can hold members accountable.

The American people don’t want it. Listen to your constituents, not your lobbyists. Strip out Section 224 from the NDAA.

Freeman, speaking to Military.com, made his position unambiguous. In his written analysis for the Quincy Institute, he argued that the initiative “would arguably do more to intertwine the U.S. military with the Israeli military than the more than $200 billion (inflation-adjusted) in military assistance Israel has received from the U.S. since its founding in 1948.”

Two specific phrases in the provision’s text drew the most pointed concern from Freeman: “network integration” and “data fusion.” Both appear without definitions. What they mean in practice, how far any synchronization of military data systems would extend, and whether intelligence data falls in scope are questions the current language does not answer.

Co-Production and the District Lock

Supporters of the provision frame it partly in economic terms. AIPAC’s factsheet says it “encourages U.S.-based co-production and manufacturing partnerships, strengthens domestic manufacturing, and supports American workers.” HASC chairman Mike Rogers (R-AL) described the measure as an efficiency improvement that gives American troops access to superior technology. Ranking member Adam Smith (D-WA) called critics’ characterizations “simply not accurate,” describing the initiative as a coordination mechanism for existing programs.

Freeman’s objection to the jobs framing is specific. He calls the co-production incentive the “real Trojan horse” in the bill: once Israeli-linked defense firms establish co-production facilities and joint ventures inside American congressional districts, those investments create local employment. Jobs become political assets. Lawmakers representing those districts become reluctant to vote against programs connected to local employers, regardless of how public opinion in their constituencies has shifted.

The F-35 program is the relevant model. Lockheed Martin deliberately distributed production across more than 40 states, assembling a coalition of congressional districts with a direct economic stake in keeping the contract funded. Critics of the provision say its co-production framework would replicate that logic specifically for the U.S.-Israel relationship, embedding it in district economies in a way that future floor fights over military assistance would find increasingly costly to disrupt.

AI, quantum machine learning, and advanced cyber research that the U.S. currently restricts even for close partners are also part of the initiative’s scope. Per the A New Policy legislative tracker’s analysis of the provision’s technology domains, those categories currently accessible only to the UK and Australia under AUKUS Pillar II would be opened to Israeli defense institutions under this framework.

The Committee Vote and What It Misses

Khanna offered his amendment at the June 4 HASC markup. The committee rejected it by voice vote. Only Khanna and Rep. Sara Jacobs (D-CA) spoke in favor of removal. Rogers defended the provision from the Republican side; Smith dismissed the critics’ characterizations as inaccurate and described the initiative as a coordination mechanism for programs that already exist.

Smith said he was “frustrated with Netanyahu’s leadership” and with Israel’s role in drawing the U.S. into the Iran war, but he disagreed that the provision represented Congress “bowing to what Netanyahu wants.”

Four days before the June 4 markup, the House passed the Iran war powers resolution 215 to 208, with four Republicans crossing to support it: Thomas Massie (KY), Brian Fitzpatrick (PA), Tom Barrett (MI), and Warren Davidson (OH). It was the fourth time such a measure came to a vote this year and the first time it cleared the House. Massie has separately committed to offering a floor amendment to strip the provision if it survived committee. It survived.

On the full House floor, members from competitive districts can’t shelter behind a voice vote; they register a position. That matters in November midterms where the Administration’s Middle East policy has already produced Republican defection on the Iran war, and where Democratic primary candidates from Michigan to California are running explicitly against U.S. military entanglement with Israel.

What Massie and Khanna Are Betting On

The Polling Numbers Behind the Floor Fight

The coalition Massie and Khanna have assembled spans the libertarian right and the progressive left, held together by opposition to expanded military integration with Israel and by survey data that has shifted sharply since the Iran war began in February. The March 2026 Pew Research Center survey and the May 2026 New York Times/Siena poll together trace the public opinion landscape the floor vote will enter:

The 60 percent Pew figure represents a nearly 20-point increase from where American opinion stood in 2022. The share choosing “very unfavorable” has nearly tripled, from 10 percent to 28 percent. Among Republicans under 50, a majority now views Israel unfavorably, a reversal of what had been a durable bipartisan foreign policy position for most of the past three decades.

AIPAC’s War Chest and the Midterm Math

AIPAC’s allied super PAC enters the midterms with nearly $100 million on hand, up from $35 million in 2022, and has spent more than $221 million on congressional races since it began engaging in competitive primaries, according to campaign finance data cited by The Intercept. Those resources have historically gone toward defending pro-Israel incumbents and defeating primary challengers critical of the relationship.

In the Senate, 40 members backed Bernie Sanders’ April 2026 resolution to block a $295 million bulldozer sale to Israel; just 15 voted for a comparable measure a year earlier. In midterm primaries from New York to California and in Midwestern swing districts, U.S. aid to Israel and AIPAC’s spending in congressional races have emerged as explicit campaign issues, per reporting by the Christian Science Monitor.

Netanyahu weighed in directly on June 1, writing to Rep. Marlin Stutzman to endorse a “new framework of joint defense cooperation, codevelopment, coproduction and mutual investment” as a successor to the current aid relationship. His letter landed during a midterm primary season where U.S.-Israel policy is generating constituent pressure across both parties. Responsible Statecraft described the underlying logic of the initiative as transforming Israel “from a top U.S. aid recipient to a full member of the U.S. defense and intelligence apparatus.”

The Senate Pathway and November

If the provision clears the House floor with the Massie-Khanna amendment defeated, the Senate still has to pass the same language. The agreement between the U.S. and Israel expires at the end of FY2028, creating pressure to formalize the relationship before that framework lapses and before a potentially different Congress takes over in January 2027.

The Arab Center Washington DC, in its analysis of the provision, concluded that even a future Congress committed to a different policy would face, by that point, a set of existing contractual agreements between American and Israeli defense companies, co-production jobs distributed across congressional districts, and Israeli technology embedded in Pentagon acquisition systems. Unwinding each of those structural attachments carries its own political and economic cost.

Whether the Senate’s documented shift on Israel-related military measures translates to opposition against language embedded in an NDAA is a different question from the arms-sale resolutions that have drawn the clearest votes. Those gave senators a narrow, visible statutory argument; NDAA language does not offer that framing.

Massie’s floor amendment will be the provision’s first recorded vote. The midterm election arrives in a political environment where the coalition that wrote Section 224 has shrunk measurably from where it stood when the agreement underpinning U.S. military aid to Israel was negotiated in 2016.

All eight were acquired by distributors at prices that already exceeded their production costs. Ticket sales grew week by week as audiences brought in new audiences. The rankings below cover all eight, with production cost, global gross, and the multiplier each achieved.

Eight Films, Eight Multipliers

None of the eight had major studio financing during production. Six were acquired after festival screenings for sums that already dwarfed what the films cost to make. Artisan Entertainment paid $1.1 million for The Blair Witch Project’s distribution rights at the 1999 Sundance Film Festival; Focus Features paid around $15 million for Obsession at the Toronto International Film Festival (TIFF) in September 2025. Those acquisition costs are separate from the production budgets below, which follow the standard industry convention: cost of making the film against global theatrical revenue.

*Obsession figures as of early June 2026; the film remains in theaters.

The eight films span found footage, supernatural horror, slasher, and torture horror subgenres. All eight were produced for under $1.1 million and independently financed before being acquired by a distributor.

The Slasher Formula, 1968 to 2004

Rawness as the Production Decision, 1968-1974

George A. Romero had spent his early career making documentaries, commercials, and industrial films, including a segment for Mister Rogers’ Neighborhood, before channeling that low-budget discipline into Night of the Living Dead in 1968. Shot for $125,000, the film depicted violence that American theater audiences had no reference for, and its final scene left no ambiguity: Ben (Duane Jones), the film’s Black protagonist, survives a night of zombie siege only to be shot dead by a misidentifying posse, a downer horror ending of a kind that had no precedent in mainstream American cinema. That closing image sent audiences out of theaters in a state they had to describe to everyone they knew. The film returned $30.2 million globally on its $125,000 budget.

Six years later, Tobe Hooper made The Texas Chain Saw Massacre for $140,000 and leaned on the same mechanism through a different technique. The film presented itself as based on true events, and enough of the audience believed that framing to spread it by personal testimony. Because it predated the home video market, it cycled through American theaters for years and accumulated its $30.8 million total slowly, with no studio marketing budget behind it.

Halloween’s Slow Build to $70 Million

John Carpenter shot Halloween in 20 days across suburban Los Angeles on a $325,000 budget. To save money, he composed the film’s score himself. The Michael Myers mask was a Captain Kirk mask purchased for $1.98 and spray-painted white. When the limited prints first toured American cities, ticket sales were weak enough that Carpenter reportedly told producer Debra Hill the film hadn’t made it. Then Roger Ebert awarded it four stars. The positive word-of-mouth cascade that Ebert’s review helped build drove Halloween to $70 million worldwide, making it the most profitable independent film of its era, a title it held until 1999.

Sean S. Cunningham watched Halloween’s trajectory and designed Friday the 13th (1980) around the same economic logic: cheap production, grisly set pieces, no-name cast. The $550,000 film earned $59.8 million globally. James Wan’s Saw, which premiered at the 2004 Sundance Film Festival and was acquired by Lionsgate, closed out the pre-found-footage era of this list at $1 million in production costs against $104 million at the global box office.

How Found Footage Remade the Math

Daniel Myrick and Eduardo Sánchez shot The Blair Witch Project over eight days in October 1997 for $60,000. Before it reached theaters, they had built a promotional website presenting the student filmmakers as genuinely missing, complete with fake police reports and childhood photographs of the actors. IMDb initially listed the cast as “missing, presumed dead.” Artisan Entertainment bought distribution rights at Sundance in January 1999 for $1.1 million. The Blair Witch Project’s production budget breakdown puts the principal photography cost between $35,000 and $60,000 depending on the accounting; the figure most commonly used is $60,000. Against that spend, the film earned $248.6 million globally.

Ten years later, Oren Peli shot Paranormal Activity in seven days at his own home for $15,000. DreamWorks Pictures acquired the domestic rights for $350,000, then Paramount (which had absorbed DreamWorks) shot a new ending for an additional $200,000, bringing total production costs to $215,000. The marketing campaign invited audiences in selected cities to vote online to “demand” a screening, then expanded to a nationwide release after organic interest built through college campuses. The film earned $193 million worldwide.

Both films used pre-release campaigns built on apparent authenticity. Blair Witch’s missing-persons website reached internet users in an era when online content was new enough to be widely trusted at face value. Paranormal Activity’s demand-screening model in college towns generated the same kind of pre-sold audience. In both cases, audiences who saw the film early became recruiters for the next wave.

Obsession’s Three Growing Weekends

Curry Barker, a 26-year-old YouTube sketch comedian, shot Obsession in 20 days in Los Angeles in October 2024. The film centers on a young man who breaks a “One Wish Willow” wishing tree and finds that the consequences are worse than the wish was worth. Barker told interviewers that the wish element was partly inspired by a Simpsons episode involving a monkey’s paw. He recorded his own dialogue as a customer service character in the film while editing from his bedroom.

Focus Features acquired Obsession at TIFF for around $15 million after a bidding war that included A24 and Neon. Jason Blum of Blumhouse Productions signed on as executive producer after the festival premiere, announced alongside the film’s first teaser in December 2025. The $750,000 production opened in American theaters on May 15, 2026, in third place behind two holdovers.

What followed inverted every pattern in wide-release horror. Horror films typically drop 50% or more after their opening weekend. Obsession grew instead:

• $23.9 million in weekend two (a 39% increase), the largest second-weekend gain for any wide-release horror film outside the Christmas season on record
• $27.4 million in weekend three (a further 14% increase), making it the first non-holiday release since Steven Spielberg’s E.T. the Extra-Terrestrial in 1982 to post gains in both its second and third weekends
• $171 million worldwide through early June 2026, with the film still in theaters

Barker has a follow-up feature, Anything But Ghosts, already completed and set for release through Focus Features and Blumhouse-Atomic Monster. A24 has separately signed him to write, direct, and produce a new Texas Chain Saw Massacre film.

The Franchise Penalty

The clearest comparison between an original and its sequel comes from Paranormal Activity. Oren Peli’s original cost $215,000 and earned a 897x return on $193 million globally. Paranormal Activity 2, made on a reported $3 million budget, grossed $177 million worldwide. That is nearly the same box office total. The multiplier was 59x, a drop of 93% from the original’s return rate.

Halloween generated 13 sequels and remakes after the 1978 original. None approached the 215x return that $325,000 had produced. Saw expanded into a 10-film franchise with production budgets scaling up at each installment. Book of Shadows: Blair Witch 2 (2000) cost dramatically more than its predecessor and earned roughly a quarter of what the original had produced. Every franchise on this list followed the same arc: the box office held reasonably well; the per-dollar return did not.

Paranormal Activity 2 opened to $40.7 million in its first weekend, a record for horror at the time. The slow organic build that had carried the original to nine figures globally took months and ran on audience recruitment. The sequel ran its box office course in two weekends. The multiplier is where the difference shows.

Obsession is still in theaters.

Frequently Asked Questions

What Is the Highest Budget-to-Box-Office Multiplier in Horror?

The Blair Witch Project holds the top spot at 4,133x, the highest documented return for any horror film with major theatrical distribution. Paranormal Activity comes second at 897x when measured against its full production cost of $215,000. Obsession, still in theaters as of June 2026, sits fourth on the list at approximately 228x and is still climbing.

How Much Did Paranormal Activity Cost to Make?

Oren Peli shot the original Paranormal Activity in seven days at his own home for $15,000. DreamWorks Pictures acquired the domestic rights for $350,000, and Paramount, which had absorbed DreamWorks by the time of release, shot a new ending for an additional $200,000. That version, with Paramount’s ending, is the cut that reached theaters in September 2009.

How Much Has Obsession Made at the Box Office?

Obsession had grossed more than $170 million worldwide by early June 2026, roughly three weeks after its May 15, 2026 theatrical release by Focus Features. The film remains in theaters, and its final total will exceed that figure.

Why Do Horror Sequels Earn Lower Multipliers?

The Paranormal Activity franchise shows the pattern numerically. The sequel cost $3 million against the original’s production cost measured in thousands of dollars, and returned 59x against $177 million worldwide. The box office totals were nearly the same. The budget was roughly 14 times higher. Audiences who saw Paranormal Activity and returned for the sequel already knew exactly what kind of film they were attending.

Who Directed The Blair Witch Project?

Daniel Myrick and Eduardo Sánchez wrote, directed, and edited The Blair Witch Project together. Both were film school graduates who had conceived the Blair Witch mythology in 1993. Principal photography lasted eight days in October 1997, and the film premiered at Sundance in January 1999, where Artisan Entertainment bought distribution rights for $1.1 million.

What Do the Films on This List Have in Common?

All eight were independently produced before being acquired by a distributor and made for under $1.1 million in production costs. Every one saw its box office grow over multiple weeks, driven by audiences recommending the film to people who had not yet seen it. The subgenres vary considerably, from found footage to slasher to supernatural horror, but the growth pattern holds across six decades of releases.

Every AI frame-generation method in wide use today adds latency by design. The interpolation model underpinning DLSS 3 (NVIDIA’s Deep Learning Super Sampling frame-generation feature), AMD’s FSR 3 (FidelityFX Super Resolution 3), and Intel’s own XeSS MFG (XeSS Multi-Frame Generation, which boosts apparent frames-per-second counts by inserting AI-synthesized frames) requires a second rendered frame before it can generate anything. Extrapolation would skip that wait. Petersen told Gizmodo the technology is “almost” ready for a public showcase, but it wasn’t finished in time for Computex 2026.

Why Frame Gen Keeps Players Waiting

Standard frame interpolation requires two genuine rendered frames. The algorithm analyses motion vectors between them and inserts a synthesized image into the display interval. Output frame rate goes up, at least on paper.

Frame one gets rendered and then sits idle while frame two is produced. The player’s input at the moment of frame one isn’t reflected on screen until the pair is complete and the interpolated image is ready to display. He quantified the cost for PC Gamer during the same Taipei trip: enabling XeSS MFG picks up roughly eight milliseconds of latency purely from that holding period. Hardware Unboxed’s 2022 testing found DLSS 3 frame generation raised total latency in Cyberpunk 2077 from 47ms to 63ms, a 16ms jump alongside a higher displayed frame count; Digital Foundry, testing concurrently, found DLSS 3 “inherently adds input latency due to the fact that it holds a frame.” Hardware Unboxed also recorded an 11ms penalty in F1 2022 under the same conditions.

NVIDIA has partially addressed this with Reflex, a system latency tool that reduces the CPU-to-GPU render queue depth, and more recently with Frame Warp inside DLSS 4, a reprojection technique borrowed from VR headset design that shifts output frames toward a player’s most recent inputs. Both reduce perceived latency without removing the underlying source, which stays tied to the pace of the base render loop.

AMD officially recommends a 60 FPS baseline before enabling FSR 3 Frame Generation. Below that floor, the latency additions aren’t masked and players feel every millisecond of the hold. On a gaming mouse in a reflex-heavy shooter, the extra 10 to 15 milliseconds is noticeable. On a controller in a single-player title, the same delay typically falls within the controller’s own inherent latency, and the player doesn’t register it.

Gaming handhelds sit at the center of that sensitivity gap. A device running at 35 watts rarely holds 60 native FPS in demanding titles at reasonable quality settings. Frame generation exists for those conditions, and the latency cost that comes with it doesn’t lift just because the screen is eight inches wide.

Intel’s Forward Gamble

What Changes with Extrapolation

Standard frame interpolation needs two rendered frames. Extrapolation needs one. It reads current hardware input alongside that single frame and predicts what comes next. The synthesized frame goes to the display while the GPU works on the genuine next render, eliminating the holding period that interpolation requires.

How Petersen Described It

I’ve got one frame. I’ve rastered it. I’m showing it to a user. And while I’m not quite ready to raster a new one, I’m going to predict where he’s going to move his mouse.

Petersen, Intel’s graphics architecture fellow, gave that description to Gizmodo journalist Kyle Barr at a sit-down in Taipei on the eve of Computex 2026. What he outlined is a real-time input-prediction model embedded in the frame pipeline: velocity, direction, and acceleration from the hardware device go in; a synthesized next moment comes out while the GPU processes the genuine render behind it.

He also used the Gizmodo interview to place blame for player skepticism toward AI frame assistance. The reason gamers have become antagonistic toward the technology, he said, is that NVIDIA “fucked it up” during last year’s Multi-Frame Generation rollout. He served as NVIDIA’s director of technical marketing until 2019 before joining Intel.

Handheld Hardware, Where the Stakes Are Highest

The Arc G3 Extreme at Computex

Intel unveiled the Arc G3 Extreme at Computex 2026 alongside those interviews. Built on the Panther Lake architecture, it’s what Intel’s team called “a GPU SoC with a CPU inside it,” meaning the 12 Xe-core Arc B390 GPU takes the design priority and a 14-core CPU cluster works in support. Below 12 watts, the performance CPU cores park entirely, leaving efficiency cores to feed power to the GPU.

MSI’s Claw 8 EX AI+ is the first device carrying the chip. Acer’s Predator Atlas 8, also announced during the show week, uses Arc G3 and G3 Extreme processors in a comparable form factor with an 8-inch 1200p 120Hz display and an 80 WHr battery. Intel also introduced Endurance Gaming at the same briefing, a mode that dynamically caps frame rates and adjusts power; in Forza Horizon 6 at 1080p low settings, Intel claimed it extends battery life from under three hours to nearly six.

• 42%: Intel’s claimed average performance advantage over AMD Ryzen Z2 Extreme at 35W (Intel internal testing; independent benchmarks pending)
• 8W to 35W: Arc G3 Extreme’s configurable thermal design power range
• June 25: MSI Claw 8 EX AI+ launch date, priced at $1,499

Frame Energy Math on Portable Devices

One AI-generated frame costs roughly a quarter the energy of rendering an original, Intel’s team explained at the launch briefing. Four frames can be produced for the energy of about one and a half native renders. On a 35-watt thermal envelope shared between CPU and GPU, that ratio changes what’s achievable at a given battery level without requiring the full power that native rendering at equivalent frame rates would demand.

“You need frame gen to get a good experience” on handheld gaming PCs, he told PC Gamer in Taipei. Visual artifacts from frame generation are also less conspicuous on an 8-inch panel than on a large desktop monitor, which lowers the quality threshold the AI model needs to meet. For the category of compact gaming hardware pushing high refresh rates within tight power budgets, controller players on handhelds are the audience that already leaves frame generation disabled because of the latency cost. Extrapolation’s design goal is to remove that cost.

When Prediction Goes Wrong

The risk is a misprediction the player sees. If the model generates a frame showing the camera tracking right while the player has already moved left, the correcting real frame arrives moments later but the brief visual pop is there. He addressed this in the interview: “The question really is: how often do people change their choices, and how quickly can you respond to the change of choices?”

His position is that most player motion is continuous. The AI model, trained on real input sequences, handles straight lines, parabolic arcs, and sustained thumbstick holds cleanly. The failure case is what he called “a very notable physical change in your hand or your thumb”: a snap reversal, a sudden stop. On a controller, those inputs are rarer and less abrupt than on a gaming mouse. Handheld controllers already carry more than eight milliseconds of inherent signal latency in their communication chain, which means a brief misprediction can land within the margin players already absorb without perceiving it.

Intel hasn’t published prediction error rates, compute overhead, or end-to-end latency figures for the consumer version. The 2024 Intel research paper GFFE: G-buffer Free Frame Extrapolation for Low-latency Real-time Rendering, published by Intel engineers on arXiv, describes the engineering method but carries no consumer performance data. A rasterization trade-off also applies: the more AI frames the system generates per rendered one, the more GPU overhead the extrapolation model consumes from the same chip. On a handheld’s integrated GPU, that overhead competes directly with the render budget frame extrapolation is designed to supplement. How Intel distributes that cost across the Arc G3 Extreme’s power range hasn’t been specified publicly.

Desktop gaming with a gaming mouse is a different test entirely. Players in reflex-heavy competitive titles notice latency differences in the two-to-three millisecond range, well below the 8ms floor that interpolation adds to a handheld. Extrapolation introduces its own latency variable in the form of misprediction, and how often that occurs under competitive desktop conditions is one of the questions the technology hasn’t answered.

From Sydney to Taipei

Intel’s public record on frame extrapolation starts at SIGGRAPH Asia 2023, the graphics research conference held in Sydney that December. There the company introduced ExtraSS, a flow-based neural network that uses a game engine’s G-buffer data (depth, surface normals, geometry) to guide the extrapolation process. The method works where G-buffers are accessible, which excludes mobile platforms, cloud gaming pipelines, and forward-rendering engines that don’t expose that data layer.

The GFFE paper addressed that constraint directly. Intel published the work to arXiv in 2024, describing a recurrent neural architecture combined with optical flow and warping techniques that produce extrapolated frames without G-buffer access. The approach handles occluded regions and shading changes that simpler warping methods fail on, the kinds of artifacts that fast-moving elements like particle effects and transparent surfaces tend to reveal. “G-buffer Free” extends compatibility to engines and platforms that ExtraSS couldn’t reach, including mobile and cloud pipelines that most frame-generation research has historically skipped. The paper targets frame quality within what real-time rendering requires, though it carries no consumer-facing performance figures.

Between Sydney and Taipei, Intel stayed quiet on the consumer side. The Taipei interview was the first substantive public update in roughly two and a half years. PC Gamer’s reporter on the ground noted there had been a real possibility of a live extrapolation demo appearing at the show. It didn’t make the floor.

No release timeline followed the interview. The next milestone is a full public showcase with benchmark numbers, prediction accuracy data, and a working demonstration. Intel first flagged the technology in 2023. The demonstration hasn’t followed.

World Aquatics lifted its participation restrictions on Russian and Belarusian athletes in April 2026, reopening international competition to Rylov for the first time since 2021. He retired instead, his most recent racing having come at the 2025 Russian Nationals.

A Career Built in Backstroke

Rylov registered on the international radar at the 2014 Summer Youth Olympics in Nanjing, winning three golds and a silver and breaking two world junior records within a single afternoon. The 50 backstroke final came less than an hour before a medley relay gold. He was 18 years old.

His first senior World Championships medal arrived the following year in Kazan, on Russian soil: bronze in the 200 backstroke in 1:54.60, a national record at the time. At Rio in 2016, he placed third in the 200 back in 1:53.97, a European record, and sixth in the 100 back in 52.74. Ryan Murphy of the United States won gold in both backstroke events that week.

The 2017 World Championships in Budapest produced his first individual world title. Rylov won the 200 back in 1:53.61, a new European record, with Murphy finishing off the podium. In Gwangju two years later, he defended the title. He entered Tokyo as a back-to-back 200 backstroke world champion with no peer in Europe across either distance.

1. 2014 Youth Olympics, Nanjing: three golds, two world junior records
2. 2016 Rio: 1:53.97, European record, 200m backstroke bronze
3. 2017 World Championships, Budapest: 1:53.61, European record, world gold
4. 2019 World Championships, Gwangju: defended 200m backstroke title, back-to-back champion

Tokyo’s Backstroke Double Was His Peak

Rylov swam 1:53.23 at the 2021 Russian Nationals, a lifetime best and a new European record that qualified him for the Olympic team. At the Tokyo Aquatics Centre, he and his teammates competed under the Russian Olympic Committee (ROC) banner because of a pre-existing WADA ban on Russian sport, the result of an established state-sponsored doping program. His times were his own.

In the 200 backstroke final, he touched in 1:53.27, an Olympic record. Four days later, in the 100 backstroke, he went 51.98, a European record. He won both events in a single Games, an achievement the backstroke had not seen in nearly four decades. He also won a silver in the 4x200m freestyle relay, splitting 1:45.26 on the third leg.

Murphy, who took silver in the 200 backstroke, told reporters afterward the race was “probably not clean,” then said he meant swimming in general and not Rylov in particular. Murphy’s qualification did not fully disperse the remark. Back in Russia, Putin awarded Rylov the Order of Russia. A national poll named him the best Russian athlete of the Games.

In November 2021, Rylov traveled to Kazan for the European Short Course Championships. He finished fourth in the 100 backstroke prelims, third among the Russian swimmers in the field, and did not advance to the semifinals under the meet’s two-per-country selection rule. In the 200 backstroke, he reached the semifinals and placed fourth before withdrawing ahead of the final. It was his last international race.

After the Luzhniki Rally

On March 18, 2022, three weeks into Russia’s invasion of Ukraine, Rylov appeared on stage at a rally at Moscow’s Luzhniki Stadium. He was one of eight Russian athletes at the event, which was hosted by Putin and built around a speech marking the eighth anniversary of Russia’s annexation of Crimea. Rylov wore a jacket bearing the letter “Z,” the symbol associated with Russian military operations in Ukraine, and stood on stage wearing his Olympic medals. He is a staff sergeant of the Moscow Oblast Police in Lobnya, and multiple reports noted that, as a serving officer, he may have had limited ability to decline the invitation.

Three consequences arrived in the days that followed:

• Speedo terminated its sponsorship deal with Rylov with immediate effect and announced the remaining balance of his contract fee would be donated to UNHCR, the United Nations Refugee Agency
• Energy Standard, his International Swimming League (ISL) club, suspended him alongside all Russian athletes and support staff on its roster
• FINA, the international swimming body that rebranded as World Aquatics later that year, opened a disciplinary investigation and issued a nine-month suspension on April 20, 2022, barring him from all internationally sanctioned competition

The nine-month ban became academic within weeks, when FINA extended a blanket restriction covering all Russian and Belarusian swimmers and absorbing his personal suspension.

I don’t understand what I did, but in the end they filed a complaint against me for that offended the feelings of other athletes. You see, I offended them by simply supporting my country, my president. I don’t know how to argue with that.

Rylov gave those comments to Russian sports outlet Sport-Express after Speedo dropped him. He had already announced he would not compete at the 2022 World Championships “as a sign of support” for Russian Paralympic athletes banned from that winter’s Games. FINA announced the comprehensive restriction on all Russian and Belarusian swimmers the same day.

The Restrictions That Outlasted His Ban

The broader restrictions on Russian swimmers changed shape several times over four years but never reopened international competition to Rylov specifically. His personal nine-month suspension ended in January 2023. The framework that replaced it came with conditions he had already said he would refuse.

The Neutral Athlete Framework

In September 2023, World Aquatics allowed Russian and Belarusian athletes to compete at international events as individual neutrals, subject to background checks and a declaration confirming they had not publicly supported the war. Rylov had told reporters he would not sign any such declaration. He did not race internationally under the neutral framework.

The program expanded in December 2024, when Russian relay teams competed at the Short Course World Championships in Budapest as Neutral Athletes B (NAB), winning three relay golds. At the 2025 World Championships in Singapore, Russia’s neutral relay squad won two long-course titles.

The April 2026 Restoration

On April 13, 2026, the World Aquatics Bureau ended the political conflict guidelines for senior Russian and Belarusian athletes, restoring competition rights under national flags, anthems, and uniforms for the first time since 2022. Athletes still needed to pass four successive anti-doping tests and clear background checks with the Aquatics Integrity Unit (AQIU) before competing. “We are determined to ensure that pools and open water remain places where athletes from all nations can come together in peaceful competition,” said the governing body’s president, Husain Al Musallam. European Aquatics asked for a delay in applying the rule within its own competitions until September 1, 2026, to protect the Paris European Championships from immediate effects. The Norwegian Swimming Federation went further, stating it would not host international events if Russian or Belarusian athletes competed under national symbols.

Records He Once Held

While Rylov raced exclusively within Russia, two benchmarks he had set in Tokyo passed to other swimmers permanently.

The 2022 World Championships in Budapest produced a world record of 51.60 from Italian swimmer Thomas Ceccon in the 100 backstroke – the same meet Russian swimmers were banned from attending. Ceccon’s swim erased Rylov’s European record of 51.98 from eleven months earlier and moved Rylov from European record holder to third on the all-time list. The world record still stands.

The 200 backstroke took longer to change. Rylov’s lifetime best of 1:53.23, set at the 2021 Russian Nationals to qualify for the Olympics, held the European record through four years of his absence. Hubert Kos, Hungary’s Olympic champion in the event at the 2024 Paris Games, swam 1:53.19 at the 2025 World Championships in Singapore, four hundredths faster than Rylov had ever gone. The European record Rylov had lowered six times across his career is now Kos’s.

The Open Door in April 2026

At the 2025 Russian Nationals, Rylov finished 13th in the 50 backstroke, 14th in the 100 backstroke, and 4th in the 200 backstroke.

Under the April 2026 World Aquatics ruling, he could have applied this summer to race under the Russian flag in international competition for the first time since 2021. The anti-war declaration that had previously blocked his neutral-athlete eligibility was no longer required under the full restoration. Anti-doping clearance and an AQIU background check remained as conditions, but the political threshold was gone. His 2025 Nationals results were a separate matter the retirement answers by omission.

The Russian Swimming Federation’s tribute praised his “champion’s spirit” and wished him well in his “new chapter of life.” His final career record: four Olympic medals (two gold, one silver, one bronze), eight long-course World Championship medals (two gold, three silver, three bronze), and seven short-course World Championship medals (three gold, two silver, two bronze).

World Aquatics opened the door in April. He walked away in June.

The old Fitbit app enforced a one-device limit, requiring users to unpair one tracker before connecting another. Google removed that constraint on May 19, 2026, when it rebranded the Fitbit app as Google Health, and for the first time a Fitbit Air and a Pixel Watch 4 can stay connected to the same Google account at once.

Two Trackers, One App

The Fitbit Air is a passive health monitor with no display and no standalone apps. Everything it records feeds into the Google Health app on a connected phone. Google announced the Fitbit Air on May 7, 2026, with retail availability beginning May 26. The oval-shaped pebble tracker measures 34.9 x 17 x 8.3mm and weighs 5.2 grams without its band, putting it lighter on the wrist than a Whoop strap. It ships with a fabric Performance Loop band at $99.99, or a Stephen Curry Special Edition at $129.99. Accessory Active silicone and Elevated Modern bands start at $34.99.

The Pixel Watch 4 launched October 9, 2025, at $349 for the 41mm Wi-Fi model and $449 for the 41mm LTE version. It runs Wear OS 6 on a Qualcomm Snapdragon W5 Gen 2 chip with a Cortex M55 co-processor, 2GB of RAM, and 32GB of onboard storage. Both size variants (41mm and 45mm) share a 12.3mm case thickness; the 41mm weighs 31 grams without a band. Its Actua 360 AMOLED LTPO display reaches 3,000 nits of peak brightness across both models.

What Each Device Measures

The Fitbit Air’s Sensor Set

The Fitbit Air carries an optical heart rate monitor running 24/7, red and infrared sensors for SpO2 blood-oxygen readings, a skin temperature sensor, and a three-axis accelerometer with gyroscope. Those sensors enable AFib (atrial fibrillation) heart rhythm alerts, HRV (heart rate variability) tracking, breathing rate monitoring, cardio load and readiness scoring, and automated workout detection. Sleep stages log continuously through the night. A vibration motor handles Smart Wake alarms, and a double-tap gesture blinks an LED to show remaining battery.

The Pixel Watch 4’s Additions

The Pixel Watch 4 carries all the Air’s sensors plus a cEDA (continuous electrodermal activity, a physiological stress response measure) sensor, ECG capability, an altimeter, barometer, magnetometer, and ambient light sensor, along with a compass. Its dual-frequency GPS combines Galileo, GLONASS, and Beidou signals and tracks outdoor routes without a phone. Five capabilities detailed on the Pixel Watch 4’s official specification page aren’t available on the Fitbit Air:

• Loss of Pulse Detection: detects when pulse stops and prompts a call to emergency services
• Satellite SOS: connects to emergency services via satellite when no cellular or Wi-Fi signal is available (LTE models only)
• ECG: on-demand electrocardiogram readings
• Fall and Crash Detection: automatically contacts emergency services after a detected fall or crash
• Gemini Raise to Talk: hands-free access to Google’s Gemini AI by lifting the wrist

Google Health’s New Multi-Device Logic

Before May 2026, the Fitbit app enforced a one-active-device rule. Pairing a Fitbit Air would displace a Pixel Watch from the account, and vice versa; users had to pick one or manage the swap manually each time. The rebranded Google Health app, which went live on May 19, 2026, dropped that restriction entirely.

The app’s sync engine now compares timestamps from both devices, merges activity data, and avoids double-counting steps or active minutes when both trackers are on the wrist simultaneously. Google’s announcement of the Health platform’s data approach describes the app as a hub aggregating inputs from wearables, smart scales, and medical records, with data portability as a design principle. That openness extends beyond Google hardware: Apple Watch, Oura Ring, Garmin, and Whoop can all connect through Health Connect or the Google Health APIs.

The Gemini-powered Google Health Coach draws on that aggregated data to generate readiness scores and personalized workout plans. It requires Google Health Premium at $9.99 per month or $99.99 per year. Both devices include a three-month trial with purchase. Heart rate monitoring, sleep scores, AFib alerts, and workout detection don’t require a subscription after the trial ends.

Does the Fitbit Air Fix the Pixel Watch 4’s Overnight Gap?

The Pixel Watch 4’s battery marks a genuine generational step. The 41mm model runs 30 hours with always-on display active, or 48 hours in Battery Saver mode. The 45mm extends those figures to 40 and 72 hours respectively, and a 15-minute charge delivers 50% battery on the 41mm. For most wearers, overnight charging is still the practical default.

Sleep is where the data gap opens. Overnight skin temperature variation, HRV, sleep stages, and breathing rate all require continuous wrist contact, which the charging window interrupts. The Fitbit Air’s 7-day battery removes that scheduling conflict: five minutes on its proprietary charging connector provides a full day of use, and a complete charge takes 90 minutes. The trade-off is a separate cable. The Fitbit Air uses a different connector from the Pixel Watch 4’s magnetic Quick Charge Dock, so someone running both travels with two chargers rather than one.

Screenless Trackers Before This One Required a Subscription

The closest competitor in the screenless fitness band segment is Whoop. Whoop bundles its hardware into a subscription starting at $199 per year, with no option to buy the strap outright. The Fitbit Air at $99.99 is the first widely available screenless tracker in that form factor that doesn’t require ongoing payment to use its core features.

Both the Fitbit Air and the Pixel Watch 4 include a three-month Google Health Premium trial with purchase. Core health features, including AFib alerts, sleep scoring, and cardio readiness, don’t lock behind the subscription when the trial ends.

• $99 – Fitbit Air, device purchase, no mandatory subscription
• $199/year minimum – Whoop, subscription-only model, hardware not sold separately
• $9.99/month – Google Health Premium, optional coaching add-on for either device after the free trial

The Pairing Case

Three situations favor the Fitbit Air on its own. Someone without interest in a smartwatch gets continuous health monitoring, AFib awareness, and sleep scoring for $99. Someone on iOS can pair the Fitbit Air via the Google Health app (iOS 16.4 or later), since the Pixel Watch 4 is Android-only. Someone already wearing a Samsung, Apple, or Garmin watch can add the Air as a dedicated sleep companion without swapping their primary device. The pattern of Google Pixel hardware building its best features around Android exclusives means the Pixel Watch 4 returns the most to users already running Pixel phones and Android services.

The Pixel Watch 4 is the necessary choice when standalone GPS accuracy matters, when Wear OS app access, Google Wallet, or real-time workout data on a screen are priorities, or when the safety features drive the purchase. Loss of Pulse Detection, Satellite SOS, and Fall and Crash Detection have no equivalent on the Fitbit Air.

For an existing Pixel Watch 4 owner, the addition is a straightforward $99 question. The Air fills the overnight window, charges during the hours the watch already needs power, and the Google Health app consolidates both in a single dashboard. Combined hardware spend comes to around $450 for the 41mm Wi-Fi Pixel Watch 4 and the standard Fitbit Air. Google has built the sync infrastructure to handle both without data gaps or duplicate counts, and that infrastructure is what makes the combination worth considering rather than a simple one-or-the-other pick.

Frequently Asked Questions

Does the Fitbit Air work with iPhone?

Yes. The Fitbit Air pairs with iOS 16.4 or higher via the Google Health app, and a Google account is required to set up the device. The Pixel Watch 4 requires Android 11 or later and has no iOS support.

Can the Fitbit Air and Pixel Watch 4 run on the same account at once?

Yes, as of May 2026. The rebranded Google Health app allows both devices to stay connected to a single Google account simultaneously. The previous version of the Fitbit app enforced a one-device limit and required unpairing before adding a second tracker.

Is a subscription required to use either device?

No. Both the Fitbit Air and the Pixel Watch 4 include a three-month Google Health Premium trial with purchase. After that, heart rate monitoring, AFib alerts, sleep scores, and workout detection remain free. Google Health Premium unlocks Gemini-powered coaching at $9.99 per month or $99.99 per year.

Does the Fitbit Air have GPS?

No. The Fitbit Air has no GPS hardware. Tracking a run or ride with route mapping requires a connected phone. The Pixel Watch 4 includes dual-frequency GPS that operates independently without a phone present.

How does the Fitbit Air battery compare to the Pixel Watch 4?

The Fitbit Air lasts up to 7 days per charge; five minutes on the charger provides a full day of use. The Pixel Watch 4 41mm lasts 30 hours with always-on display active, or 48 hours in Battery Saver mode. The 45mm model extends those figures to 40 and 72 hours respectively.

What health features does the Pixel Watch 4 have that the Fitbit Air doesn’t?

The Pixel Watch 4 adds ECG readings, Loss of Pulse Detection, Fall and Crash Detection, Satellite SOS (LTE models only), Gemini Raise to Talk, a cEDA body response sensor, altimeter, barometer, and dual-frequency GPS. Both devices support AFib alerts, HRV, SpO2, sleep stage tracking, and cardio readiness scoring.

Alvarez Climbs to Sixth in Franchise History

Alvarez entered Saturday as the AL leader in both home runs and RBIs, carrying a season slash of .316/.430/.654. He went 1-for-3, drew a walk, and scored twice in what became Houston’s highest run-scoring game since an 18-1 rout of the Los Angeles Dodgers last July 4th.

The 192nd homer as an Astro moves past one specific name on the all-time list: Bregman’s. Alex Bregman spent his entire career in Houston before signing with the Boston Red Sox in free agency after the 2024 season and is now with the Chicago Cubs. Alvarez passed him in June of his age-28 season.

Next on the franchise list is Jim Wynn, “The Toy Cannon,” who accumulated 223 home runs for Houston between 1963 and 1973. That figure has sat in the record books for more than five decades, and Alvarez needs 31 more to reach it. The Astros-Athletics matchup had looked very different nine months earlier, when Oakland shut out Houston 6-0 in Sacramento and both Alvarez and Altuve went hitless as part of a five-game losing streak that damaged Houston’s 2025 playoff standing.

Above Wynn on the franchise list: Craig Biggio with 291 home runs, Lance Berkman with 326, and franchise leader Jeff Bagwell with 449. Bagwell and Biggio are both in the Baseball Hall of Fame.

Houston third baseman Isaac Paredes hit his 100th career home run earlier in the same week. Alvarez is 28 years old, has played in all but one game in 2026 after missing more than 100 games the year before due to injury, and carries a career 162-game average of 42 home runs.

Altuve’s New Place in Second-Base History

Altuve’s home run Saturday was his fourth of 2026. He is batting .246 on the season. The count he passed Saturday is position-specific: only at-bats logged while Altuve is the listed second baseman apply toward the record. DH appearances and pinch-hit slots do not.

Lou Whitaker played 19 seasons alongside Alan Trammell in Detroit, won five All-Star selections and three Gold Gloves, and took home the 1978 AL Rookie of the Year Award. His career WAR of 75.1 places him, per a Bleacher Report ranking of the 25 greatest modern second basemen, among the strongest Cooperstown cases for any player not yet inducted. His position-specific total of 239 home runs at second base had stood as the seventh-best mark since his 1995 retirement. Altuve tied that figure in a late-May game against the Texas Rangers before surpassing it Saturday.

The players ahead of Altuve in the second-baseman ranking include Rogers Hornsby, Joe Morgan (who spent his early career with the Astros), and Ryne Sandberg. Jeff Kent holds the all-time record at 354 home runs at the position and is set to be inducted into the Baseball Hall of Fame. Altuve’s Baseball Reference career page lists 259 total career home runs, all accumulated with Houston, with the position-specific 240 as a subset of that figure.

He signed with the Astros as an amateur free agent in 2007 and made his major league debut in July 2011. The longest-tenured current member of the organization, he is the only active Astros player who was with the club when it competed in the National League before the franchise’s 2013 move to the AL West. His credentials include seven Silver Slugger Awards (tied for the most ever at second base), the 2017 AL MVP, and two World Series rings. He is also one of four second basemen in MLB history to collect both 250 career home runs and 250 career steals, joining Craig Biggio, Joe Morgan, and Ryne Sandberg in that company. The contract he signed in 2025, five years at $125 million, runs through 2029. Four years remain.

The Record Books After Saturday

With Saturday’s games in the books, two active Houston players hold positions in separate all-time lists. Alvarez sits sixth in Astros franchise history. Altuve sits seventh all-time among second basemen in MLB.

The Astros franchise all-time home run leaders, as of June 6, 2026:

The second-baseman position HR leaders above Altuve, per MLB.com, include four names with Hall of Fame credentials:

• Jeff Kent – 354 (all-time leader at the position)
• Ryne Sandberg
• Joe Morgan
• Rogers Hornsby

Altuve’s 259 total career home runs, all with Houston, place him fourth in the franchise table above. His 240 position-specific HRs are a subset of that career figure, reflecting games in which he appeared as the listed second baseman rather than a DH or pinch hitter.

How Far Is the Top?

Alvarez’s Path to Bagwell

Jeff Bagwell’s 449 home runs as an Astro stretch across 15 seasons from 1991 through 2005. Alvarez needs 257 more to reach that total. At his career 162-game average of 42 home runs, the gap closes in roughly six full seasons, putting the record within reach sometime in his mid-30s assuming consistent availability.

The sourced context from Saturday’s game notes a lower bar also works: 25 or 26 home runs per year from now through age 38. His current production sits far above that floor. Health was the single largest variable in 2025, when he missed more than 100 games. His near-perfect attendance through early June in 2026 resets the projection.

Between Alvarez and Bagwell sit three more milestones: Wynn at 223, Biggio at 291, and Berkman at 326. At 22 home runs through early June, Alvarez is on a pace that could carry him past Wynn’s franchise mark before the season closes in October.

Altuve’s Contract and Kent’s Record

Jeff Kent played 17 seasons across multiple franchises, including a two-year stint with the Astros in 2003 and 2004, before retiring in 2008 with 354 home runs specifically at second base. He won the 2000 NL MVP Award with the San Francisco Giants and is set to enter the Hall of Fame. That position record has stood for nearly two decades. Altuve needs 114 more to match it.

Four contract years remain through 2029. Closing a 114-homer gap in that window requires roughly 28 or 29 home runs per season as a second baseman. Altuve is 36, batting .246 in 2026 with four home runs through early June. His best output came in the 2017-2019 stretch; the years left on his deal will arrive at ages 37 and 38. The math on Kent’s record requires a return to those levels.

He currently ranks 16th all-time in innings played at second base, behind Craig Biggio. Four more seasons in the position will move that ranking higher regardless of the home run trajectory.

The AL MVP Picture

After Saturday’s win, the manager spoke about what Alvarez brings to the Astros’ offensive performance:

In big moments, he rises to the occasion and just finds a way to boost our club. Just a really good hitter. But we had a ton of really good at-bats and a lot of people contributed to our win today.

Joe Espada, Houston’s manager, said after the game at Daikin Park.

Alvarez’s 2026 season through Saturday:

• 22 home runs (AL leader)
• 48 RBIs (AL leader)
• .316 batting average
• .654 slugging percentage

In the second expert MVP poll of the 2026 season, a survey of 35 MLB.com writers and analysts, Alvarez edged Aaron Judge of the New York Yankees as the AL front-runner. That poll was compiled before Saturday’s grand slam.

The Yankees announced Judge suffered a stress fracture in his right first rib cage in a game at Houston in April and would be re-evaluated in four to six weeks before any return timeline is set. His absence removes the most direct AL competition from the field for a significant stretch of the summer.

Alvarez entered June with the AL lead in home runs and RBIs, his franchise ranking just updated, and five names still above him on the list he has been climbing since his 2019 debut.

Davis calls the technique a “green movie.” His Los Angeles production company, Stag Films, has built three features this way since 2008, and the government’s archive is nowhere near exhausted. The puzzle is why so few filmmakers followed him into it.

The Archive That Belongs to Everyone

The National Archives and Records Administration (NARA, the federal body that preserves U.S. government records) holds more than 360,000 reels of film dating from the 1890s through the late 20th century, assembled under the Federal Records Act, which requires agencies to transfer their non-current records when they’re retired. Its online catalog describes approximately 200,000 individual items, but only about 35,000 of those have an accessible remote viewing copy. For the remainder, a researcher must travel to Washington, D.C., and work in person in the Motion Picture, Sound and Video Research Room.

• 360,000+ film reels in NARA’s motion picture holdings, dating from the 1890s
• 200,000 items catalogued in NARA’s online database
• 35,000 with an accessible remote viewing copy
• 40+ military record groups, from WWI Army Signal Corps footage through Air Force reels of the 1980s

The military collections run deepest. The Army Signal Corps record group alone carries more than 12,000 titles covering maneuvers and battles from the First World War through the 1980s. The U.S. Navy’s audiovisual holdings (Record Group 428) run to approximately 13,300 reels spanning from before World War I to the early 1980s. Law enforcement and domestic agency records extend the reach further: training films for FBI procedures, civil-order protocols, weapons qualification, and drug-awareness programs that generated their own peculiar visual vocabulary. Davis found Sonny Bono hosting one of the anti-drug productions. He also found footage of nuclear test towns with mannequins melting under simulated blast effects and medical training sequences that exist in a category of their own.

The legal foundation is 17 U.S.C. 105, which holds that works produced by U.S. federal employees as part of their official duties are not eligible for copyright protection in the United States. NARA’s permissions page for its motion picture holdings confirms the agency charges no royalties and does not license its content. Nuances apply: donated materials from private sources within NARA’s collection may carry restrictions, and the agency cautions that researchers bear responsibility for confirming the status of individual items. For films produced directly by federal agencies, the archive is free.

The International Documentary Association has documented a steady decline in researchers visiting NARA’s film reading rooms, even before the COVID-19 disruption. Davis spent two years inside that archive for one movie.

Making Film in Reverse

The Silent Cut

Davis’s workflow breaks the standard production sequence entirely. He starts with existing footage, combs through it for anything weird, visually striking, or usable, and assembles a rough visual edit before a word of script is written. The result is something like a silent film: a sequence with a loose beginning, middle, and end but no dialogue, no locked characters, no fixed story. He calls this phase “making a film backwards.”

Building a consistent character from government archive material requires creative substitution. Davis identifies someone who appears in multiple government productions, collects every frame of that person he can find, then supplements with stand-ins: other footage shot from behind, or at similar angles, with comparable hair or build, close enough to serve as a cutaway. The cast was assembled this way across a source collection that would ultimately demand far more raw material than working from a feature-length skeleton ever had.

Writing the Script You Don’t Have Yet

Once the rough visual cut exists, Davis writes the script against it. The footage dictates what the story can be. He can’t write a scene he has no footage for, so he writes into what he already has, and the archive keeps offering material that a clean-sheet process would never arrive at.

I can’t imagine something that I don’t have footage for. So I’ll come up with a sort of a loose general plotline and then just put all this footage together in a way that makes some kind of sense. Not a lot of sense, but at least could hopefully work as a story with a beginning and middle and an end with somewhat consistent characters.

Davis, the writer-director behind Stag Films, made the point in an interview with KPBS Cinema Junkie host Beth Accomando.

With the script complete, he brings in voice actors, records the full dialogue track as a standalone audio performance without the footage present, then lays that audio over the existing visual cut, tightening the edit around it. He compares the process to animation production, where voice recording precedes animation and the visuals are built to fit the performance. Here the order reverses: the visuals exist first, the audio is recorded against the script, and the edit reconciles the two. His son worked as editor on the film. The rest of the production was Davis alone, in his home workspace, over roughly two years.

The Stag Films Green Canon

Sex Galaxy (2008), Davis’s first green movie, used the 1968 science-fiction film Voyage to the Planet of Prehistoric Women as a structural skeleton: a feature-length spine with consistent characters onto which he grafted government shorts and additional stock footage. President Wolfman (2012) followed the same logic, pulling the 1973 political horror film The Werewolf of Washington as its base and assembling more than 100 additional source films around it.

For his third film, Davis dropped the spine entirely. Building every scene from fragments without a feature-length anchor required nearly triple the source material of President Wolfman; character continuity came entirely from his editing instincts, with no pre-existing film’s structure to borrow. The film ran the festival circuit through 2025, appearing across the United States and at the Babylon Kino in Ghent, Belgium. It screens tonight at the Art Theater Long Beach. Tubi carries it for free.

Birth of a Nation Reversed

The KKK as Tech Company

The plot opens with a racist white cop shooting a Black man. A modernized Ku Klux Klan and the U.S. military both want the incriminating bullet retrieved from the body, and the Klan’s preferred tool is a high-tech shrink ray. In Davis’s version, the organization has rebranded as something closer to a Silicon Valley company: slick conference presentations, product launch events, promotional videos. An Atlanta mayor goes missing as the FBI and the military converge on the same case.

Davis has been direct about the political origin. The resurgence of organized white nationalism over the past decade and more, and his anger about it, drove the premise. Blaxploitation cinema of the 1970s provided the tonal register: heightened, funky, knowingly over the top, and committed to its point of view. Government training footage from that same era gave him the visual vocabulary to work with: the grain, the flat institutional lighting, the earnest bureaucratic delivery of people who believed in what they were filming.

Two specific archive sources supply the KKK sequences. Part of the footage comes from D.W. Griffith’s Birth of a Nation, the 1915 silent film that was both one of cinema’s technical milestones and an explicit piece of Klan propaganda in which the organization is portrayed as America’s heroic defenders. The film sits in the public domain, confirmed through a legal ruling upheld by the Supreme Court in 1976 and under U.S. copyright law for works published before January 1, 1931. The second source is actual federal surveillance documentation: records of Klan activities that government investigators collected, now held in the federal archives.

Davis is using the primary artifact of cinematic white supremacy as raw material for a film mocking it. The choice of source footage is the argument made visible at the level of method.

Recording Without Watching

The voice cast records in a studio without footage in front of them. Davis directs line by line, the actors take two passes on each, and he lays the finished audio track over the existing visual cut, tightening the edit around what they gave him. Performers don’t need to see what their character is doing, he has found. They need the tone and the context of the scene.

The sync is imperfect, and Davis frames this as a deliberate aesthetic choice. Flat-affect, slightly off-timed delivery makes the absurdist writing hit harder. A punchline delivered with clinical government-film deadpan is already funny; dubbed over footage that doesn’t quite match, it becomes funnier still. The model is Woody Allen’s What’s Up, Tiger Lily?, the 1966 comedy in which Allen redubbed a Japanese spy film to entirely different comic effect, a film Davis cites as a direct influence. Once audiences accept the convention in the opening minutes, the gap between audio and image stops reading as error.

Why Didn’t Anyone Else Follow?

Davis has been explicit since the President Wolfman era that part of the goal was to inspire other filmmakers. The premise of Stag Films’ green movie catalog is demonstrably sound: build a narrative feature with crowds, explosions, military hardware, and multiple locations without a studio, a budget, or a set. A portion of the government archive is searchable online through the Internet Archive’s U.S. government film collection, free from any browser, with no trip to Washington required.

The constraint Davis describes is creative, not financial. You need what he called “a twisted brain” to hold the method’s logic together across hundreds of source films while building a story that can only go where the footage permits. That’s a genuine demand, and it’s the only meaningful cost in the method.

The production cost Davis very little monetarily but took roughly two years of sustained attention. That trade is the method’s core offer: time and ingenuity in exchange for production infrastructure. The green movie approach removes the financing dependency entirely, which has been Davis’s argument since Sex Galaxy proved the concept in 2008.

Davis is already pulling footage for “Hot Heat Hawaii,” his next green movie, a drug-smuggling surfing adventure. The government’s archive isn’t running out.

Competition Across Seven Events

The championship ran across seven event categories on Saturday and Sunday, with swimming held at the Monte Nitzkowski Aquatics Center at Long Beach City College, a new venue this year per SOSC’s Summer Games event information. Track and field, basketball, bocce, and flag football filled the main campus grounds at Cal State Long Beach, joined by Unified versions of both bocce and flag football.

Unified Sports puts athletes with intellectual disabilities alongside partners without intellectual disabilities on the same competition team, matched at similar skill levels. Special Olympics Unified Sports research found that majorities in nine out of ten countries surveyed believed people with intellectual disabilities could not play competitive sport alongside those without them. Unified teams put both groups in the same competitive lineup.

The 2026 Summer Games sport categories:

• Athletics (track and field)
• Basketball
• Bocce
• Unified Bocce
• Flag Football
• Unified Flag Football
• Swimming

Summer Games is the championship of Special Olympics Southern California’s (SOSC) spring training season, with athletes qualifying through regional competitions held across the preceding months. For many competing at Long Beach this weekend, that path started in county-level programs months ago, coaches calibrating each athlete’s event roster to match their training progress. County-level Special Olympics invitationals run the same qualifying-to-championship structure across the country, including in Girard, Ohio last month where 167 athletes from five counties competed at Arrowhead Stadium under the same format.

Seven Disciplines, Zero Cost

The Healthy Athletes stations ran in parallel with competition across the weekend, offering free medical assessments in seven disciplines to SOSC athletes. Volunteer healthcare professionals and students staff the screenings, conducting non-invasive assessments and issuing follow-up referrals when a clinician identifies something worth further evaluation. Coaches could pre-register their teams ahead of the Games to schedule athlete appointments at the stations.

People with intellectual disabilities face multiple barriers to routine healthcare: cost, communication challenges, difficulty finding providers with relevant experience, and a lack of adapted clinical resources in many communities. The Healthy Athletes model addresses several of these at once by bringing screenings to the athlete, in a sports setting they already attend, staffed by volunteer clinicians who accept no payment.

Those barriers carry documented consequences. When Special Olympics first offered on-site health screenings at the 1995 World Games, a significant share of athletes had conditions that day serious enough to require emergency room referrals, a finding that prompted the formal launch of the Healthy Athletes program in 1997.

What the Screenings Cover

SOSC’s event materials named six of the seven disciplines offered at this year’s Summer Games. Each targets a clinical area where people with intellectual disabilities face elevated rates of undetected or untreated conditions, according to Special Olympics Northern California’s clinical program data.

The Fit Feet station addresses podiatric conditions that often go untreated outside a clinical setting. Undetected hearing conditions are another consistent finding at events like Summer Games, where athletes sometimes discover for the first time that routine auditory support could improve how they communicate with coaches and teammates during competition.

A Training Ground for Clinicians

The screenings work in both directions. Athletes move through the stations and receive referral paperwork if a clinician flags something worth following up. The volunteer healthcare professionals and students staffing those stations accumulate hands-on clinical experience with a patient population that most medical and dental training programs address minimally. Data published by Special Olympics Northern California shows that less than 20 percent of healthcare professionals have specific training to treat patients with intellectual disabilities, which means events like Summer Games serve as one of the few structured environments where that clinical training happens at scale.

Athletes who receive a referral at a Healthy Athletes station are connected to local healthcare providers for follow-up evaluation, giving the program a care-continuity function that extends beyond the competition weekend. The cumulative reach of that work, across events like Summer Games around the country and worldwide, shows up in Special Olympics’s global Healthy Athletes program data:

• 2 million+ free health screenings delivered globally since 1997
• 300,000 health professionals and students trained to work with patients with intellectual disabilities
• 50% of Special Olympics athletes have preventable or treatable foot conditions affecting sport performance
• 15% of athletes screened at the 1995 World Games needed urgent emergency care referrals on the day of screening

The Youngest Athletes on Campus

A designated area of the campus served children ages 2 to 7 through SOSC’s Young Athletes program. The activities ran motor skills drills, hand-eye coordination exercises, and sports skills demonstrations, structured as a first experience of organized athletic activity and a preview of what SOSC’s programs look like as children grow older.

The program is designed specifically for early childhood entry, built on the principle that children with intellectual disabilities benefit from structured physical and social activity at the same developmental stages as children without. Families are integral to the design: parents and caregivers participate alongside their children rather than watching from the sideline, giving the format a joint discovery structure. SOSC frames the area as the earliest rung of the competitive pathway that leads to Summer Games competition.

Children who move through the Young Athletes area begin building the physical habits and familiarity with the Special Olympics format that carry into formal competitive eligibility later on. Summer Games is one of the few events where the full range of SOSC’s programming appears in one location, and families attending could move between the competition venues, the health screening stations, and the Young Athletes area, with all three running simultaneously across the campus grounds.

How the Torch Got to Long Beach

The cauldron that opened Summer Games on Friday evening had covered roughly 1,100 miles before arriving at the LSB Financial Credit Union Pyramid. The Law Enforcement Torch Run (LETR) for SOSC left Chula Vista on May 26, with more than 300 law enforcement agencies carrying the Flame of Hope in relay legs across Southern California over two weeks. Friday’s opening ceremony at 5:30 p.m. included the Parade of Athletes before the cauldron lighting, which concluded the relay at its Summer Games destination.

The route that led to Long Beach also served a larger purpose. The relay’s passage through Los Angeles on June 1, which ran from a pep rally at Crypto.com Arena to the LA Memorial Coliseum, was part of the Final Leg of the 2026 Special Olympics USA Games Torch Run. That national relay is headed to Minnesota’s Twin Cities, where the 2026 USA Games will open on June 20, uniting 3,000 athletes for competition through June 26.

In Southern California, the LETR program raised $1.2 million in 2024, with more than 200 law enforcement agencies and stations involved. Every dollar raised directly covers costs like transportation, uniforms, equipment, and venue fees for athletes, ensuring that participation in SOSC’s programs remains free. More than 3,000 officers participated in this year’s Southern California relay.

Those same officers showed up at the Saturday competition, with scores of them positioned along the entry corridor at Cal State Long Beach as athletes arrived for competition day.

A Festival Beyond the Track

SOSC organized a family festival alongside the competition, with live music, interactive games, and activity booths run by partner companies and community groups. A fan zone gave supporters materials to make signs and cheer athletes through their events during the day.

An art display featuring work by SOSC athlete Samantha Roman ran across the weekend in a dedicated exhibition area. Roman is the athlete-featured artist for the 2026 Summer Games, and her display ran through both competition days. SOSC merchandise, including hats, water bottles, shirts, and sweatshirts, was also available on site.

Kelly Pond, president and CEO of Special Olympics Southern California, issued a statement for the event.

The athletes inspire us every day with their energy, determination, confidence and spirit.

Pond’s statement described Summer Games as the direct expression of SOSC’s mission to empower individuals with intellectual disabilities through the life-changing impact of sports.

The Season That Led Here

Summer Games closes the spring competitive season for SOSC, which runs 12 sports across four seasons and reaches nearly 25,000 athletes in community and school-based programs across 11 Southern California counties. More than 50 full-time employees and thousands of volunteers support those programs alongside corporate partners and individual donors.

The organization’s four seasons cover a year-round calendar, with sports ranging from aquatics and athletics to bocce, basketball, golf, and winter sports. Athletes can enter SOSC’s programs at any skill level, with training designed to build toward competition across each season.

The next season opens June 13, when bowling begins and runs through August. Admission to all competition venues at this weekend’s Summer Games was free to the general public.

That 21.70 beats by 0.13 seconds the mark she ran to win Olympic gold at Paris 2024 and falls 0.10 short of her personal best of 21.60, set in 2023. Fellow American Kayla White finished second in 22.07, with Nigeria’s Favour Ofili third in 22.15.

A Time Faster Than Olympic Gold

The Lone Star Grand Prix was the first US-based World Athletics Continental Tour Gold meeting of 2026, a tier below the Diamond League. Saturday’s programme at Cushing Stadium, Texas A&M’s 2,200-seat outdoor track venue, excluded events longer than 800 metres. Temperatures reached 88 degrees Fahrenheit with a 77-degree dew point; the wind reading in the women’s 200m was +0.7 metres per second, legal for records.

The previous 2026 world best in the event had been 21.86, run by Julien Alfred, St. Lucia’s reigning Olympic 100m champion, on April 30 in Austin. Thomas cut that by 0.16 seconds. Saturday’s 21.70 was the first sub-21.80 clocking in the women’s 200m this season. She arrived in College Station five weeks after running 10.95 for 100 metres at the Botswana Golden Grand Prix in Gaborone, the first wind-legal sub-11 clocking of her career. She seized the lead on the bend and held the gap to the line, with White 37 hundredths of a second back.

I’m a little surprised by the time, but I’ve been training hard. I had a great little tour in Africa and now the hard work is paying off. I’m enjoying having an ‘off’ year where there’s not the pressure of an Olympics or World Championships, so I’m just having fun this season.

Thomas’s coach throughout this comeback has been Tonja Buford-Bailey, who guided her through rehabilitation and back to Saturday’s start line.

The Year She Couldn’t Reach Tokyo

An Achilles tendon injury that surfaced in May 2025 and was re-aggravated in July defined the whole of last season. Thomas forced herself to a third-place finish at the USATF Outdoor Championships in August, qualifying for the World Athletics Championships in Tokyo by one-thousandth of a second with a time of 21.95. Then she and Buford-Bailey decided the Achilles could not safely carry her through championship rounds; Thomas announced her withdrawal in early September.

While Thomas sat out, Melissa Jefferson-Wooden swept the sprint double at the World Athletics Championships in Tokyo, running 10.61 in the 100m to set the championships record and collecting the 200m title alongside it. Thomas had collected three golds at Paris 2024, the 200m, the 4x100m, and the 4x400m relay, along with a bronze from the Tokyo 2020 Olympics. An individual World Championships title had never arrived; silver in Budapest in 2023 was the closest she had come.

“The toughest thing about having to drop out of Tokyo last year was just the frustrations of knowing that I worked all year to run really well in Japan,” Thomas told Athletics Weekly in April. She said the experience had made her more resilient: “You can bounce back from it and become an even better athlete afterwards.”

She did not race in any Diamond League events after the USATF Outdoor Championships, ending 2025 ranked fourth in the world at 200m, behind Julien Alfred (21.71), Jefferson-Wooden (21.84), and McKenzie Long (21.93).

Rebuilding Across Africa

Thomas opened 2026 carefully. At the Texas Relays in Austin on April 4, she ran 11.00 seconds for 100 metres, equalling her personal best in the event, and anchored a 4x400m relay leg. She posted afterward that there was “lots of work to do this season.”

What followed was her first time competing in Africa, a three-stop circuit she chose to build competition fitness away from the season’s highest-profile meets before bringing the times home. The legs ran in sequence:

1. Addis Ababa Grand Prix, April 18 (Ethiopia): Won the 100m in 11.13 and the 200m in 22.15, both victories at altitude on a new track facility.
2. Kip Keino Classic, April 24 (Nairobi, Kenya): A 1.1 metres-per-second headwind in the 100m held her to 11.01, but she still won the race.
3. Botswana Golden Grand Prix, April 26 (Gaborone): Ran 10.95 for 100m into a slight headwind, her fastest effort of the African leg.

American sprinter Cambrea Sturgis, who trained alongside Thomas through the circuit, finished behind her at each stop. Thomas had said she wanted to be fully recovered from the injury before returning to the US circuit; four wins across three African cities later, the Achilles was holding up.

Tate Taylor and 19.97

The inaugural Lone Star Grand Prix replaced the USATF NYC Grand Prix, which last ran in 2024. After USATF could not make the finances work in New York, the meeting moved to College Station with Texas beef producer 44 Farms as a lead backer. The programme excluded events longer than 800 metres, a format suited to the Texas summer heat. Saturday was the meet’s first edition.

Thomas was not the only story at Cushing Stadium. In the men’s 200m, 18-year-old Tate Taylor ran 19.97 seconds, becoming only the second American high schooler in history to break 20 seconds in the event. The first was Erriyon Knighton, who set his high school best of 19.49 in 2022 before turning professional. Taylor, committed to Texas Tech University for the fall, finished second; Zimbabwe’s Makanakaishe Charamba, an Olympic 200m finalist, won in a personal best of 19.88. The time moved him to equal sixth on the world under-20 all-time list, and he also holds the US high school 100m record at 9.92.

Elsewhere in the programme, American sprinter Trayvon Bromell won the men’s 100m in 9.85, but a wind reading of +3.8 metres per second placed the result outside legal record territory. World 100m champion Oblique Seville of Jamaica finished third in that race. Jamaica’s Sabrina Dockery, 19, won the women’s 100m in a personal best of 10.92, breaking 11 seconds for the first time.

Budapest in September

Thomas confirmed in April that she intends to run the 200m at the World Ultimate Championships in Budapest, scheduled for September. The meet carries a $10 million prize pool, with $150,000 for each event winner, and will draw the sprint field that matters most in a season without a World Athletics Championships or Olympics.

The Diamond League circuit occupies the weeks ahead, giving Thomas competitive rounds against world-ranking fields. She also co-owns Athlos, the women-only track league founded with investor Alexis Ohanian, one of several commercial ventures she has built alongside her competitive career. Jefferson-Wooden, the reigning world sprint champion, will almost certainly share the track with Thomas in September.

Her personal best of 21.60 sits 0.10 seconds ahead of Saturday’s mark. The championships are in September.