OpenAI London staff averaged £2.3m in 2025, including shares

Accounts show OpenAI UK Ltd paid its 110 London employees an average of £2.3m each in 2025, a sharp rise from the previous year.

Threat Times Systems DeskOctober 18, 20262 min read

Share LinkedIn Reddit

Illustration: A sleek office desk in a London high-rise holds a smartphone displaying a blank notification screen, next to a stack of gold coins and a stylized graph trending sharp — AI-generated illustration: Threat Times

Key points

OpenAI UK Ltd reported £293m in turnover and £20.3m profit for 2025.
The average compensation package rose to £2.3m from £1.1m the prior year.
London remains OpenAI’s largest international research hub with 110 staff.

OpenAI UK Ltd employees in London received an average compensation package of approximately £2.3 million each in 2025, according to company accounts reported by The Guardian. This figure includes both salaries and equity shares for the 110 staff members based in the city. The Next Web reported these details, noting that the publication has not independently verified the financial data.

The total compensation consisted of £87 million in salaries and £168 million in shares distributed among the workforce, The Guardian said. This average represents a significant increase from the £1.1 million average package recorded the previous year. The 2025 figure is roughly 58 times the median earnings of a full-time employee in the UK, based on Office for National Statistics data.

Financial Context and Operations

OpenAI UK Ltd more than tripled its turnover in 2025, reaching £293 million, while recording a profit of £20.3 million, according to The Guardian. The company maintains its primary office in Westminster and is preparing to open a new headquarters in King’s Cross, an area that also hosts Google DeepMind. London serves as the organization’s largest international research hub.

The London office houses not only AI researchers, who receive the largest compensation packages, but also teams focused on government affairs, sales, and marketing, The Guardian reported. Pay varies widely across the organization, with significant differences between the highest and lowest earners. OpenAI declined to comment to The Guardian regarding these specific financial disclosures.

Broader Industry Landscape

These figures emerge during a period of intense competition for artificial intelligence talent, The Guardian noted. Reports indicated that competitors such as Meta were offering multi-year compensation packages worth several hundred million dollars during the same period. George Osborne, the former Conservative chancellor, joined OpenAI in December of the previous year to serve as head for countries.

OpenAI is currently seeking new funding at a $1.4 trillion valuation ahead of a planned stock market listing, according to The Next Web. Chief Executive Sam Altman has stated that the company will prioritize its mission over its share price. Earlier this summer, OpenAI disclosed that its agents had hacked Hugging Face, highlighting ongoing security challenges in the sector.

What should you do?

Security professionals should monitor official company filings and verified news reports for accurate financial and operational data, avoiding unverified social media claims. Organizations should review their own compensation structures to ensure they remain competitive in the AI talent market without compromising security protocols.

WATCH FOR:

Verify financial claims against official regulatory filings or direct company statements.
Monitor industry trends for shifts in talent acquisition strategies and compensation models.
Stay updated on security incidents involving AI agents, such as the reported Hugging Face breach.
Review internal access controls to prevent unauthorized data exfiltration by automated agents.

What to do and how to stay safe: OpenAI

Sudden role changes in output signal a successful override of system instructions.
Hidden metadata in images or documents often carries invisible injection payloads.
Immediate isolation of the affected session prevents lateral movement within the AI workflow.

Prompt injections exploit the model’s ability to follow instructions by hiding malicious commands in input data. Isolate affected sessions immediately and implement external guardrails to filter inputs before they reach the model.

Step-by-step guide: Prompt Injection Warning Signs: How to Spot AI Manipulation

General security guidance from the Threat Times newsroom. It is not confirmed advice from the organisations named in this story.