In February 2026, OpenAI announced that ChatGPT had crossed 900 million weekly active users. By June, Reuters confirmed the app had reached one billion monthly active users, the fastest in history to reach that figure. The lab behind that product closed a $122 billion funding round in March at an $852 billion post-money valuation, led by SoftBank, Amazon, and Nvidia. It generates $25 billion in annualized revenue. In July 2026, it has 670 open roles across nine offices on four continents. Twenty-eight of them are remote-eligible.
That final number is the editorial frame for this piece. OpenAI is not a remote-first company. Sam Altman called remote work "a mistake" in 2023, and the policy has not changed: employees near an office report in-person three days per week, Monday through Wednesday. New hires who are not local receive 45 days to relocate. The case for OpenAI in The Remote Issue is not about remote eligibility. It is about what the nine-office global footprint does to talent access at scale.
The Workforce Architecture
OpenAI's headcount was approximately 4,500 as of March 2026, per the Financial Times, up from roughly 1,700 at the start of 2024. That is the fastest net-new headcount build in frontier AI over a comparable window. The company has stated plans to reach 8,000 employees by December 2026, implying a net-new hiring pace of approximately 12 employees per business day through H2. ENTRA estimates the July 2026 headcount at approximately 5,000 to 5,500.
The physical footprint spans nine cities: San Francisco (Mission District headquarters), New York City (Hudson Yards), Seattle, London, Dublin, Tokyo, Paris, Brussels, and Singapore. Each location serves a distinct function. Dublin anchors EMEA legal, enterprise sales, and compliance. London houses the largest non-SF research cluster. Tokyo services Asia-Pacific enterprise partnerships. Paris was added as OpenAI moved to access the Francophone ML research community. Brussels is a direct EU regulatory engagement outpost. Singapore anchors APAC regulatory relationships and regional partnerships.
Compensation Landscape
OpenAI converted its Profit Participation Units to standard RSUs in January 2026, following the company's conversion to a public benefit corporation in October 2025. The structural effect on recruiting was immediate. Competing labs, particularly Anthropic, had used PPU complexity as a consistent offer-stage objection since 2023. That objection is now gone. OpenAI competes on the same equity instrument as every other frontier lab, with the liquidity argument sharpened by an IPO Sam Altman has publicly targeted for Q4 2026.
Per Levels.fyi data as of July 2026: total comp ranges from $249,000 at L2 to $1.27 million-plus at L6. L5 software engineers earn approximately $1.15 million TC, composed of a $336,000 base and $774,000 in annual equity. The ENTRA-estimated company median sits at approximately $608,000. Forward Deployed Engineers, the fastest-growing role category in H1 2026, run $350,000 to $550,000 TC. Research scientists at the L5 equivalent earn $600,000 to $1 million.
At the L5 equivalent, OpenAI pays an estimated 20 to 30 percent premium over Anthropic's $575,000 to $725,000 TC range for software engineers, per ENTRA salary-survey aggregation against Levels.fyi. Google DeepMind's London L5 research scientist band runs $450,000 to $650,000 at UK purchasing parity, a wider gap from OpenAI SF rates than the headline numbers suggest.
Global Hiring Corridors
London is the most productive non-SF talent corridor. OpenAI has drawn from the UCL, Imperial College, and Oxford graduate pipelines, alongside the researcher cohort that previously sat in Google DeepMind's London cluster. The office competes directly with DeepMind and Wayve for ML research talent and benefits from a graduate supply that no other frontier lab can replicate outside the US.
Dublin provides the EU hiring corridor without the legal complexity of a continental EU entity. Enterprise sales and solutions engineering roles serving EMEA accounts sit here, alongside the compliance infrastructure that EU AI Act enforcement requires. The Irish operation functions as a regulatory bridge more than a research hub.
Paris was added in late 2024 as OpenAI moved to access the INRIA, ENS, and Paris-Saclay research networks. The rationale was partly defensive. Mistral AI holds a home-court advantage in Francophone ML recruiting. The Paris office narrows that gap by giving OpenAI a local presence in the hiring conversations that Mistral's founders are also running.
The Remote Issue Lens
The canonical question in The Remote Issue is: how much of this company's hiring is genuinely accessible to someone who cannot or will not relocate? For OpenAI, the answer is 4 percent. Of the 670 roles tracked by ENTRA as of July 21, 2026, 28 list no office requirement. The remote-eligible pool clusters around policy, trust and safety, and a small number of infrastructure engineering roles. The research and applied ML spine is entirely office-anchored.
The more productive frame for ENTRA readers is the satellite office argument. If you can relocate to London, Dublin, Paris, Tokyo, or Singapore, you gain access to OpenAI hiring without moving to San Francisco. That is a global talent access argument, not a remote work argument.
The contrast with Anthropic is instructive. Anthropic runs 8 percent remote eligibility against 411 open roles, with deliberate concentration in interpretability research, policy, and trust and safety. OpenAI runs 4 percent against 670 open roles, with a wider office network and a more rigid in-person requirement. Anthropic's distributed strategy is functionally remote. OpenAI's is geographically satellite.
ENTRA Talent Index
ENTRA Talent Index: 78/100 (A)
Component breakdown: Hiring Velocity 90, Compensation 100, Retention 62, Mission Alignment 60.
The compensation score of 100 is the one unambiguous data point in OpenAI's profile. No frontier lab in ENTRA's tracking pays better at equivalent levels. The retention figure reflects a two-year retention rate that SignalFire's 2025 State of Talent report placed at 67 percent, below Anthropic at 80 percent and Google DeepMind at 78 percent. The RSU conversion and the $852B valuation trajectory into a 2026 IPO are improving that picture in real time. The mission alignment score reflects ongoing internal debate around the PBC conversion and several prominent departures, including Mira Murati and Ilya Sutskever, that cited mission concerns as contributing factors.
What's Next
OpenAI's H2 2026 priorities are public and consequential for the talent market. The Forward Deployed Engineer program is scaling from hundreds toward the low thousands as enterprise deployment becomes the primary revenue growth lever. The o3 reasoning model is extending to longer-context tasks. Operator, the company's agent product, is in active enterprise rollout. An IPO in Q4 2026 would require the pre-listing build in finance, legal, compliance, and people operations that every public-company transition demands. Several of those roles will carry remote eligibility that the current job board does not yet reflect. The 4 percent figure will move.
ChatGPT user figures: 900M WAU announced February 2026 (OpenAI); 1B MAU reported June 2026 (Reuters). Valuation and funding: $122B raised, $852B post-money, closed March 2026 (Bloomberg, CNBC, Forbes). Revenue ($25B ARR) and headcount (4,500, March 2026) per ENTRA Intelligence briefing "OpenAI Hiring 2026: Talent Strategy After the PBC Shift" (June 7, 2026), citing The Information and Financial Times. Headcount target (8,000 by December 2026) per Financial Times. ENTRA July 2026 headcount estimate (5,000–5,500) is a midpoint projection based on stated pace. PBC conversion: October 28, 2025. RSU conversion: January 2026. Open-role count (670) and remote-eligible count (28, 4%) reflect ENTRA tracking of OpenAI careers page as of July 21, 2026. Compensation figures from Levels.fyi public data, July 2026, and ENTRA salary-survey aggregation. Retention rate (67%) from SignalFire 2025 State of Talent report. Anthropic and Google DeepMind compensation comparison from Levels.fyi and ENTRA salary-survey aggregation. ENTRA Talent Index (78/100, A) reflects ENTRA composite methodology applied to July 2026 data.