OpenAI Cuts $800 Billion From Spending Plans — Silicon Valley Discovers Math
In a development that has shocked absolutely no one with a calculator, OpenAI has quietly informed investors that its grand plan to spend $1.4 trillion on computing power by 2030 was, shall we say, enthusiastically optimistic. The revised figure? A mere $600 billion. That’s not a budget cut. That’s a budget bereavement.
To put this in perspective: OpenAI earned $13.1 billion in revenue in 2025. It then proposed to spend $1.4 trillion. This is the financial equivalent of earning a junior doctor’s salary and announcing you intend to purchase Switzerland.
Sam Altman’s Trillion-Dollar Typo (That Lasted Six Months)
The original $1.4 trillion pledge — confidently announced by CEO Sam Altman to anyone with a microphone — was described at the time as a commitment to build 30 gigawatts of computing resources. Which sounds impressive until you remember that a company pulling in $13 billion a year was promising to outspend the GDP of a medium-sized country.
When investors politely asked how this maths was supposed to work, Altman reportedly snapped during a podcast: “If you want to sell your shares, I’ll find you a buyer.” That’s the corporate equivalent of flipping the Scrabble board because someone questioned your spelling of “trillion.”
According to Futurism, the company has now “reset” its expectations. In Silicon Valley, “reset” is the preferred term for what the rest of the world calls “admitting you made it up.”
The AI Budget Bubble — A Comedy in Three Acts
Act One: Sam Altman announces $1.4 trillion in infrastructure spending. Journalists applaud. Investors nervously nod. Accountants quietly weep.
Act Two: Revenue comes in at $13.1 billion — respectable by any sane measure, but not exactly the launchpad for spending more than the entire GDP of Spain. Investors begin asking awkward questions at awkward hours.
Act Three: The number magically shrinks by $800 billion. Press releases are issued. Everyone pretends this was always the plan.
The Data Centre Dynamics coverage noted, with commendable restraint, that $600 billion is “still an enormous number.” Indeed. It’s roughly equivalent to the entire global semiconductor industry’s annual revenue. The fact that this is now being presented as the conservative option tells you everything you need to know about the AI spending era we are living through.
OpenAI’s Revenue vs. Ambition: A Visual Guide for the Financially Baffled
Imagine your mate Dave earns £40,000 a year. Dave announces he plans to spend £4.3 million redecorating his flat. When his bank manager asks him to clarify, Dave revises his estimate down to £1.8 million. Dave then expects a round of applause for his fiscal responsibility.
This is, in essence, what has happened at one of the most valuable private companies in human history.
The company is simultaneously raising over $100 billion in fresh funding — with Nvidia reportedly chipping in $30 billion, presumably to ensure their chips remain the chips of choice for the chips that power the chips. SoftBank and Amazon are also along for the ride, because nothing says “disciplined investment” like a funding round where 90% comes from strategic partners who also happen to be suppliers.
The Stargate Project — Or: How to Name a Spending Plan After a Sci-Fi Show and Still Run Out of Money
Much of OpenAI’s infrastructure ambition has been channelled through the Stargate Project — a joint venture with Oracle and SoftBank designed to build data centres across the United States at a pace that would make a Roman emperor blush. As of late 2025, Stargate had expanded to five new US sites, boasting nearly 7 gigawatts of planned capacity.
Now, with the budget “recalibrated,” data centre operators who had been planning expansions based on OpenAI’s original projections are quietly reassessing their own spreadsheets. According to Fintool, analysts flagged that OpenAI could run out of cash by 2027 without additional capital — which, for a company currently valued at $730 billion pre-money, qualifies as a spectacular form of irony.
What $600 Billion Actually Looks Like (A Brief Tourism Guide)
In case you’re struggling to visualise $600 billion, here is a helpful comparison:
- It is more than the entire GDP of Switzerland, Norway, and Belgium — combined.
- It is roughly what Google earns in advertising across nearly a decade.
- It is still, somehow, being called a “more realistic” number.
The Meridiem — with the studied seriousness of a publication that has never eaten an entire pizza and then recalibrated its dietary expectations — declared this moment “the AI industry grows up.” One imagines the AI industry had slightly more hair before all this.
Sam Altman: The Man, The Myth, The Moving Goalpost
To be fair to Altman — a sentence that requires considerable effort to type — his company did exceed its 2025 revenue target, burning through $8 billion rather than the anticipated $9 billion. In startup terms, spending slightly less than you planned to lose is practically a profit.
ChatGPT now supports 900 million weekly active users. The company’s coding tool, Codex, has surpassed 1.5 million weekly users. CNBC reports that consumer and enterprise revenue is expected to contribute equally to a projected $280 billion annually by 2030 — a figure that requires OpenAI to grow its revenues by roughly 2,000% in five years. Which is ambitious. Which is also, perhaps, why we are having this conversation.
Meanwhile, rivals Anthropic and Google DeepMind are watching events unfold with the calm composure of people who already knew the dessert menu was too long.
Lessons From the Great AI Wallet Implosion of 2026
What the OpenAI budget saga teaches us — besides the obvious lesson that press releases are not the same as balance sheets — is that the AI infrastructure arms race has officially entered its slightly-more-sober phase.
According to ZeroHedge, the broader tech industry is now watching with interest as hyperscalers — Google, Amazon, Microsoft — collectively guide toward $700 billion in capital spending this year. Which means even if OpenAI has found its sensible shoes, everyone else is still wearing rocket boots.
The lesson, for those keeping score at home: ambition is excellent. Revenue helps. A basic grasp of the difference between the two is, it turns out, entirely optional in Silicon Valley — at least until someone starts asking about the IPO.
Auf Wiedersehen, amigo!
In February 2026, OpenAI — the San Francisco-based company behind ChatGPT, founded in 2015 and valued at approximately $730 billion — announced it was revising its planned compute spending from $1.4 trillion down to $600 billion by 2030. The original figure, touted by CEO Sam Altman in late 2025, had raised alarm among investors given that the company earned only $13.1 billion in revenue in 2025 while burning through $8 billion in cash. The revision came as OpenAI finalised a new funding round exceeding $100 billion, with Nvidia, SoftBank, and Amazon among the strategic investors. The move was widely interpreted as an attempt to reassure investors that the company’s spending ambitions were tied to realistic revenue projections rather than the unbridled optimism that characterised its earlier announcements.
Alan Nafzger was born in Lubbock, Texas, the son Swiss immigrants. He grew up on a dairy in Windthorst, north central Texas. He earned degrees from Midwestern State University (B.A. 1985) and Texas State University (M.A. 1987). University College Dublin (Ph.D. 1991). Dr. Nafzger has entertained and educated young people in Texas colleges for 37 years. Nafzger is best known for his dark novels and experimental screenwriting. His best know scripts to date are Lenin’s Body, produced in Russia by A-Media and Sea and Sky produced in The Philippines in the Tagalog language. In 1986, Nafzger wrote the iconic feminist western novel, Gina of Quitaque. He currently lives in Holloway, North London. Contact: editor@prat.uk
