OpenAI Has 712 Open Roles. Nearly All the New Ones in London and Delhi Have Nothing to Do With Research.
500 Seats in King's Cross: What the London Lease Actually Signals
On April 13, 2026, OpenAI announced it had signed a lease for an 88,500-square-foot space at Regent Quarter in King's Cross, London. The office, spanning Jahn Court and the Brassworks Building, will have capacity for more than 500 team members — more than double the roughly 200 staff OpenAI currently employs in the city. It is expected to open in 2027.
The location is deliberate. King's Cross has become a cluster of AI companies including Google DeepMind, Meta, Synthesia, and Wayve. OpenAI is planting its flag in the middle of that ecosystem.
In February, before the lease was signed, OpenAI had already announced it would make London its largest research hub outside the United States. The new office gives that claim physical form. "The UK has an incredible depth of talent and a strong track record in AI," said Phoebe Thacker, London site lead at OpenAI. "London is already a key hub for our research and teams, and this new office gives us the space to keep building here."
The timing carries a wrinkle. The London announcement came just days after OpenAI paused its UK Stargate infrastructure project, a data center buildout the company shelved over high industrial energy costs and regulatory friction. Discussions with project partner Nscale are still ongoing. The pause was widely read as a blow to the UK government's AI infrastructure ambitions — industrial energy prices in the country remain among the highest in the world.
The office lease signals that OpenAI is separating two bets: infrastructure and talent. The Stargate pause says the UK is expensive for compute. The King's Cross commitment says the UK is worth it for people. OpenAI's London workforce already spans research, engineering, customer support, policy, and sales. Under a memorandum of understanding with the UK government, OpenAI has introduced data residency for UK customers and launched an SME Accelerator programme.
Zero G Talent's board currently lists OpenAI with 40 roles added in the past seven days, including an Account Director for Large Enterprise based in London and a Recruiter for AI/ML Research EMEA — a hiring pace that suggests the office buildout is already driving headcount planning.
Why New Delhi — and Why Now
India has the second-largest ChatGPT user base on Earth, behind only the United States. That single fact is driving OpenAI's decision to open its first permanent office in New Delhi.
The company's careers portal lists three Account Director roles for India: Digital Natives (seven years of SaaS sales minimum), Large Enterprise (ten years, $2M+ annual targets), and Strategics (fourteen years, senior partnership leadership). All three are hybrid, all three are sales-focused, and none are research positions. Zero G Talent's board shows an AI Success Engineer role posted for Delhi alongside the enterprise sales hires — a pattern that mirrors what OpenAI is doing in London, scaled for a market where the buyer landscape is fragmented across thousands of mid-size firms and a handful of massive conglomerates.
OpenAI has also moved on the infrastructure side. The company has partnered with India's Tata Group to secure 100 megawatts of AI-ready data center capacity, with plans to scale to 1 gigawatt, according to TechCrunch.
The policy angle matters too. In June, OpenAI signed a Memorandum of Understanding with the government-backed IndiaAI Mission to launch the OpenAI Academy, offering AI learning resources in English, Hindi, and other Indian languages. The deal also includes up to $100,000 in API credits for 50 startups or fellows selected by the IndiaAI Mission.
The competitive pressure is real. Google, Microsoft, Meta, and Amazon already employ thousands of engineers across India. Anthropic's careers board shows active APAC hiring out of Singapore. OpenAI is late to the geography but is making up for it with a focused enterprise sales push.
The Hidden Signal: Sales Engineers, Not Researchers
OpenAI's careers page lists 712 open roles. A close look at where those roles are concentrated tells you about the company's current priorities.
The London office alone has openings for Solutions Engineer, Large Enterprise; AI Deployment Engineer, Large Enterprise; Forward Deployed Engineer; and Manager, Forward Deployed Engineering. Delhi has Solutions Engineer; AI Success Engineer; AI Deployment Engineer, Codex; and Partner AI Deployment Engineer — AWS. These are go-to-market roles, and they dominate the hiring pipeline for both new offices.
The pattern holds across the full board. OpenAI is hiring Account Directors for Large Enterprise in London, Stockholm, Madrid, and Seoul. It is hiring Forward Deployed Engineers in London, Dublin, Munich, Paris, Singapore, Sydney, Tokyo, Abu Dhabi, and Madrid. Solutions Engineers for Digital Natives are posted in Dublin, Paris, and San Francisco.
Compare that to the research side. Almost every Researcher, Research Scientist, and Research Engineer role is based in San Francisco. A single "Researcher, Training" position sits in London. The EMEA research recruiting presence is exactly one role: Recruiter, AI/ML Research EMEA, based in London. OpenAI is building out its commercial engine in Europe and South Asia while keeping its research core concentrated at headquarters.
This matches what the job descriptions themselves demand. The Solutions Engineer, Large Enterprise role in London requires "5+ years of experience in a technical pre-sales role, managing C-level technical and business relationships with complex global organizations." These are sales engineering jobs with deep technical bars.
The shift makes sense given the revenue trajectory. OpenAI's annualized revenue reached $20 billion in 2025, up from $6 billion in 2024, with more than one million organizations now using its technology, according to industry reports citing OpenAI's enterprise data. At that scale, the bottleneck is no longer model capability — it is deployment.
What the Roles Actually Demand: Reliability, Safety, and Security at Scale
OpenAI's 2025 enterprise report states: "Enterprise problems present the hardest technical challenges for frontier intelligence, requiring reliability, safety, and security at scale." That is the engineering bar the company is staffing against.
The April 2026 update to OpenAI's Agents SDK added native sandbox execution — agents now run in isolated environments with their own files, tools, and dependencies, accessible through providers like Cloudflare, Vercel, E2B, and Modal. Karan Sharma, who works on OpenAI's product team, told TechCrunch the goal was to let users "go build these long-horizon agents using our harness and with whatever infrastructure they have." The SDK also shipped an in-distribution harness for frontier models, giving developers a standardized way to deploy and test agents.
Oscar Health used the updated SDK to automate a clinical records workflow that previous approaches couldn't handle reliably. Rachael Burns, Staff Engineer and AI Tech Lead at Oscar Health, said the difference was "not just extracting the right metadata, but correctly understanding the boundaries of each encounter in long, complex records."
OpenAI's Frontier Governance Framework, published in May 2026, defines systemic risk as foreseeable material risks of severe harm — specifically, scenarios where a model contributes to more than 50 fatalities or causes $1 billion in property damages from a single incident. It categorizes threats across cyber offense, chemical/biological/radiological/nuclear risks, harmful manipulation, and loss of control, with distinct capability tiers for each. The framework maps to the EU's General-Purpose AI Code of Practice and California's Transparency in Frontier AI Act.
For the solutions engineers and enterprise sales staff OpenAI is hiring, this translates into a specific set of customer-facing demands. The company's own Solutions Engineer postings require 7+ years of experience managing relationships with large, global, complex organizations. The Technical Success team owns the customer experience across both the API platform and ChatGPT Enterprise.
The compensation range for OpenAI's research roles on the Zero G Talent board runs $295,000–$555,000 USD/year, a figure that reflects the scarcity of people who can operate across both the harness-engineering layer and the enterprise deployment surface.
The Talent War: Anthropic, Google, and Everyone Else
OpenAI's London and New Delhi hiring push isn't happening in a vacuum. The same enterprise go-to-market roles it's racing to fill (Applied AI Architects, Solutions Engineers, Customer Success Managers) are the exact positions Anthropic, Google DeepMind, and others are staffing at the same moment.
Anthropic's enterprise hiring is already massive. The company's careers board lists 373 open roles. Its Sales department alone has 62 open positions. Applied AI Architects are posted in London, Dublin, Munich, Paris, Tokyo, Bangalore, and Sydney. Customer Success Managers are split across San Francisco, New York, Seattle, Washington DC, London, and Singapore.
The Applied AI team (27 open roles) is where Anthropic's enterprise ambitions are most visible. The job titles are specific: "Applied AI Architect, Enterprise Tech," "Applied AI Architect, Government Technology," "Applied AI Security Architect" in London. This is a sector-by-sector sales engineering buildout.
And the research side isn't shrinking. Anthropic still has 60 open roles in AI Research & Engineering, including positions in London, Zürich, and its core US offices.
Google DeepMind's enterprise push is harder to read from job boards alone, but the revenue numbers tell the story. Google Cloud's Q1 2025 revenue hit $12.3 billion, up 28% year-over-year, with AI services cited as a direct driver. Google was named a Leader in the 2025 IDC MarketScape for Worldwide GenAI Life-Cycle Foundation Model Software, specifically for the Gemini model family delivered through Vertex AI.
Google's advantage is distribution. It already has the cloud infrastructure relationships, the enterprise sales teams, and the compliance frameworks that OpenAI and Anthropic are still building. But the talent pool for people who understand both frontier models and enterprise IT is finite, and Google is drawing from the same one.
What this means for the talent market is straightforward: the same profiles are in demand everywhere. An Applied AI Architect with experience deploying large language models in regulated industries can choose between OpenAI in London, Anthropic in Dublin or London, or Google DeepMind's enterprise teams. Anthropic's posted roles in the US range from $140,000 to $850,000 depending on seniority and function. OpenAI's board shows similar bands.
The scramble is global. Anthropic has open roles in Tokyo, Seoul, Singapore, Sydney, Zürich, and Brussels. These are commercial beachheads, and the people being hired there will determine whether frontier AI companies can convert model capability into recurring enterprise revenue.
What This Means for Engineers and Operators
OpenAI's London and Delhi expansion is a concrete signal of where the frontier AI industry is heading — toward deployment at scale, inside organizations that have compliance departments, legacy IT stacks, and procurement processes.
For engineers and operators weighing career moves, the pattern is worth reading carefully. The money is on both sides of the house, but the geographic spread tells the story: OpenAI is hiring where the customers are, not just where the research happens.
A 365 Data Science analysis of 1,000 AI engineer job postings found that only 2.5% of positions targeted candidates with 0–2 years of experience. Employers want people who have already deployed models in production, not just trained them in a notebook. The most in-demand skills were PyTorch (37.7% of postings), TensorFlow (32.9%), and Kubernetes (17.6%) — the tooling of deployment, not experimentation.
This tracks with what OpenAI itself is signaling. The company has hired forward-deployed engineers to embed on-site with enterprise clients, a model borrowed from Palantir. The job is less about model architecture and more about making AI systems work inside organizations that have data governance rules, uptime requirements, and legal review cycles. An Interview Query analysis of OpenAI's enterprise hiring noted that trust and compliance are the core challenges when AI agents access internal systems and operate across sensitive unstructured data like emails and reports.
For researchers, the picture is more mixed. AI researcher salaries still command a premium — averaging $150,000 to $350,000+ in the US according to AI People Agency benchmarks — but the hiring volume is smaller and more concentrated in a handful of labs. The Bureau of Labor Statistics projects 26% growth for computer and information research scientists through 2033, but the bulk of that demand is in applied roles, not pure research.
The practical takeaway: if you're an engineer or operator looking at the AI job market right now, the highest-demand, highest-leverage roles sit at the intersection of ML engineering and enterprise deployment. That means MLOps, cloud infrastructure, API development, and the work of monitoring, retraining, and securing models in production. It also means being willing to work where the enterprise customers are (London, Delhi, Singapore, Frankfurt) not just where the research labs are.
The researchers who will thrive are the ones who can ship, not just publish. And the engineers who understand why a model fails inside a bank's compliance environment will be worth more than the ones who can squeeze another benchmark point out of a training run.
Working in AI? Zero G Talent tracks the openings: browse AI jobs, openings at OpenAI and Anthropic, and the people building the field.