Openai doubles headcount while meta and amazon slash 31,500 jobs
While Silicon Valley hands out pink slips by the thousands, OpenAI is writing offer letters. The ChatGPT maker will swell from 4,500 to roughly 8,000 employees before the 2026 ball drops, three people close to the plan told the Financial Times.
The hiring blitz is a direct counter-punch to Anthropic, the Amazon-backed rival that has poached researchers and enterprise clients at a steady clip. OpenAI’s recruiters are now scouring Stanford, MIT and every European ai lab for product builders, sales closers and reinforcement-learning zealots who can squeeze more dollars out of Fortune 500 pilots.
San francisco real estate sends its own signal
Signs of the arms race sit two blocks from the company’s current Mission-district headquarters. OpenAI quietly leased a second tower, 625,000 square feet, enough desks for the incoming battalion. No remote-first slogans here: desks, whiteboards and kombucha taps await, five days a week if you want the RSUs to vest.
The move flips the post-pandemic script. Meta just shed another 1,500 roles in late April, lifting its two-year total past 21,000. Amazon has vaporized 30,000 positions since October, many in Alexa and retail engineering teams now “automated” by generative scripts. Google, Microsoft and Tesla have all used similar language—ai efficiency—as they thin the herd.
Lo que nadie cuenta es que OpenAI’s burn rate is already $7 billion annually, mostly on GPU rentals and exorbitant talent packages. Doubling payroll will push operating losses past $10 billion next year, according to two investors briefed on the numbers. The bet: enterprise subscriptions and a rumored $200-per-month “Pro” tier will cover the tab before cash runs dry.
The desktop trapdoor
Internally the expansion is code-named “Atlas,” the same nickname given to a forthcoming desktop app that bundles ChatGPT, Codex and the image generator under one shortcut. Engineers who saw the alpha say it pipes context straight from Excel, VS Code and Slack, turning Anthropic’s separate web claude into a browser tab you forget to open.
Recruiters are dangling front-loaded equity and a promise: help us ship Atlas before Thanksgiving or watch Anthropic eat the market. The clock is audible—Altman himself has been cold-messaging researchers at DeepMind and character.ai, offering seven-figure packages if they start within 30 days.
La ironía is brutal. The same models that let Meta fire content moderators and Amazon drop procurement staff now force OpenAI to hire humans by the stadium. Scaling laws still hunger for data, but they devour specialized brains even faster. The company burning cash to replace everyone else’s jobs needs people—8,000 of them, give or take—to keep the illusion fully automated.
Come December, either Atlas launches and revenue hockey-sticks, or OpenAI becomes the most expensive job-creation program in tech history. The interviews are already scheduled; the new building’s lights stay on past midnight.