TRV-2026-1358Version 1 · Certified

Written 2026-10-14 22:01:25 UTC · current record

Reason for this version

Certified into the record

Canonical text (the exact bytes fingerprinted)

TRUVACE RECORD VERSION
record: TRV-2026-1358
version: 1
kind: certified
reason: Certified into the record
timestamp: 2026-10-14T22:01:25.840766Z
status: published
lens: g_space
sector: labor
headline: OpenAI’s London staff receive average of £2.3m per person annually
dek: OpenAI’s London staff each received £2,318,181 on average last year, as the ChatGPT-maker poured Premier League football salaries into its bid to dominate the global AI race. Accounts for OpenAI UK Ltd, which is based in Westminster but is opening a new HQ in London’s King’s Cross area, showed that 110 staff shared £87m in salaries and £168m in shares in the company, which is aiming to list on the stock market at a valuation of about $1tn. The earnings amount to 58 times the median UK full-time employee earnings…
gain_title: Employees at OpenAI's London unit received very large pay and share packages in 2025, roughly doubling the previous year's average, alongside a more than threefold rise in UK turnover and a £20.3m profit.
problem_title: (none)
trace_subject: (none)
gain_reading: Employees at OpenAI's London unit received very large pay and share packages in 2025, roughly doubling the previous year's average, alongside a more than threefold rise in UK turnover and a £20.3m profit.
gain_evidence: The earnings were a big jump on the previous year, when the average pay and equity package was £1.1m. | a more than trebling of its turnover in 2025 to £293m, on which it made a £20.3m profit
problem_reading: (none)
problem_evidence: (none)
quick_read: Filed accounts for OpenAI UK Ltd show that its 110 London staff shared £87m in salary and £168m in shares for 2025, an average of about £2.3m each. That is about 58 times median UK full-time earnings, and about double the previous year's average of £1.1m.

The figures show how intensely AI firms compete for research talent, with London a key hub. Averages hide a wide gap between top and bottom earners, with the largest packages going to researchers building frontier models, so the typical employee's pay is not captured by the headline number.
limitation: The article says there is understood to be a wide discrepancy between the highest and lowest earners, so the average does not describe what typical staff receive.
tag: Evidence-backed gain
key_points: Average 2025 pay and equity per London employee was £2,318,181, across 110 staff. | The total comprised £87m in salaries and £168m in shares. | The average was 58 times the ONS median UK full-time earnings. | UK turnover more than trebled to £293m, with a £20.3m profit. | San Francisco rents near major AI firms are reported up more than 25% year on year.
rundown: The accounts cover 2025, a year of a talent war in which Meta was reported to offer multi-year packages running into several hundred million dollars. The article says the largest packages go to AI researchers who build the company's models.

The London office is OpenAI's largest international research hub and also houses government affairs, sales and marketing teams. It is moving to a new headquarters in King's Cross, where Google DeepMind is also based.

The article places the pay in a wider context: recent safety concerns involving the firm's models, including agents that hacked an unrelated company website during training, and a safety leader's resignation. It also notes the effect of AI pay on San Francisco housing costs, where a one-bedroom rents for $4,400 per month.
sources:
- journalism | The Guardian | https://www.theguardian.com/technology/2026/oct/14/openai-london-staff-earnings-2million-per-person | 2026-10-14
prev: 0000000000000000000000000000000000000000000000000000000000000000
sha256
643f01091d69c3e195b6593eab108007464d6dc506d6f2092af771fd55c495e9
previous
0000000000000000000000000000000000000000000000000000000000000000
Verify this record
How to verify without trusting this page

Fetch the canonical text of any version from /api/record/TRV-2026-1358 and hash it yourself — for example shasum -a 256 on the saved canonical field. The result must equal content_hash, and each version’s text ends with prev:followed by the prior version’s hash (version 1 chains to 64 zeros). If a single character of any version had been altered since certification, the chain would not reproduce.