OpenAI has a bold plan to pay for its $1 trillion spending spree: Ads, personal assistants, and cheaper subscriptions
Reports suggest OpenAI is exploring all options to fund its huge infrastructure investment
OpenAI has lined up more than $1 trillion in spending – and now it's trying to figure out how to pay for it all.
Over the last few weeks, the AI leader has signed a $300 billion contract with Oracle as well as multi-billion dollar deals with Nvidia, AMD, and Broadcom.
Reports of this huge spending spree have echoed across the AI industry, fueling concerns that the bubble is about to burst. If bullish AI companies can't pay their bills, infrastructure providers could find themselves out of pocket.
And no wonder there are concerns. Though these deals are staggered over several years, OpenAI last year posted revenue of about $10 billion.
While that's expected to double this year, that's still a far cry from being able to cover such bills – and don't forget that this time last year OpenAI was scrounging up $6 billion from investors to stay in business.
How is OpenAI paying for this?
OpenAI has plans to pay for it all, according to reports from the Financial Times, which note that about 70% of OpenAI's $13 billion in annual recurring revenue comes from consumer versions of ChatGPT.
Indeed, in January CEO Sam Altman said the tech giant was "currently losing money on OpenAI Pro subscriptions" because people used them so heavily.
Sign up today and you will receive a free copy of our Future Focus 2026 report - the leading resource for IT decision-maker insight on priorities and investment areas in AI, security and more.
Given that, the first goal is to double the number of paying subscribers from just 5% of the total 800m or so globally, perhaps by offering cheaper access in more places beyond India.
Another revenue boost is the addition of a checkout in ChatGPT, letting OpenAI take a slice of purchases. Advertisement revenue is another consideration, though CEO Sam Altman has said OpenAI would only do ads "with great caution".
Beyond leveraging ChatGPT's popularity with consumers, the FT reported that OpenAI plans to make use of its expertise to help develop AI infrastructure improvements, something that could prove popular with the companies it's signed all those deals with.
The tech company is also working with ex-Apple designer Jony Ive on an AI-powered personal assistant device.
Beyond those ideas, OpenAI is hoping it can continue to raise funds from investors – and believes expensive compute costs will fall thanks to competition in the sector.
Circular deals
Another solution is to be found in the deals themselves. The partnerships with Nvidia and AMD have been both described as “circular”. For example, Nvidia is investing $100 billion in the company to let it buy Nvidia’s highly sought-after chips.
Meanwhile, AMD has offered up 160 million shares for a penny apiece to help fund its shopping spree.
The FT cited an unnamed source who had advised the company, who suggested the current deal setup with a host of industry stakeholders makes perfect sense given the bold ambitions the company has.
"I don’t view them as drunken sailors going to bars and laying down IOUs everywhere,” they told the publication.
“It might look and feel that way, but this is actually a strategy backed up by technology, products, business plans and visibility into what is happening."
A complete account of OpenAI deals
All told, the FT estimates OpenAI has signed on for more than 26 gigawatts of computing capacity at a cost of more than $1 trillion. However, that covers the next several years of deployments and many of the plans are tiered, so they can be clawed back or ditched if not needed.
Last week, OpenAI and Broadcom announced they were working on 10 gigawatts of compute together; though financial details weren't disclosed, CNBC estimated that each 1 gigawatt costs a whopping $50 billion at the moment.
Elsewhere, OpenAI confirmed a $300 billion deal with Oracle over the next five years as part of the $500 billion infrastructure project known as Stargate.
Alongside that, Nvidia and OpenAI announced a partnership that could see the chipmaker invest as much as $100 billion into the AI company, which will buy hardware from Nvidia to build up to 10 gigawatts of compute.
Shortly after, OpenAI and AMD announced a deal worth as much as $100 billion over several years, that could see OpenAI take a 10% stake in AMD as it deploys 6 gigawatts of the chipmaker's GPUs.
Earlier this year, OpenAI signed a five-year, $12 billion deal with CoreWeave for infrastructure as it started to step away from Microsoft. OpenAI even tried to buy AI coding startup Windsurf for $3 billion, but lost out to Cognition instead.
Make sure to follow ITPro on Google News to keep tabs on all our latest news, analysis, and reviews.
MORE FROM ITPRO
- OpenAI just revealed what people really use ChatGPT for
- Everything you need to know about OpenAI's new open weight AI models
- OpenAI wants to simplify how developers build AI agents
Freelance journalist Nicole Kobie first started writing for ITPro in 2007, with bylines in New Scientist, Wired, PC Pro and many more.
Nicole the author of a book about the history of technology, The Long History of the Future.
-
Poor business context is scuppering enterprise AI adoption – here’s why that mattersNews Research from Alteryx has found that more than half of organizations can't effectively operationalize the business knowledge AI needs
-
Manchester University: from AI initiators to AI integratorsThe institution where Alan Turing first addressed the topic of artificial intelligence is rolling out Microsoft 365 Copilot access to 65,000 staff
-
‘Real-time data remains the voice that every leader must hear’: This one engineering role could be the key to building a truly data-driven enterpriseNews AI has increased the expectation for fast, or even instant, decisions. Data streaming engineers could help leaders keep pace
-
This new technique could improve AI output accuracy by 80% – and tackle hallucinations once and for allNews Researchers believe the new method will improve accuracy and help reduce costs
-
Reports: UK could consider regulation amidst growing 'rogue AI' concernsNews New AI minister said regulation is on the table, but only if public safety is at risk from AI
-
UK firms are automating roles, but nowhere near ready to outright replace themNews ONS figures show that organizations are automating tasks, rather than entire jobs
-
Two-thirds of workers are so fed up with ‘AI slop’ that they ‘feel nostalgic for pre-AI work’News A survey has revealed that dealing with low-quality 'AI slop' is making jobs feel less meaningful and more repetitive
-
3 ways to get your data AI-readySponsored Data placement, preparation, and orchestration make or break an AI project – and the Dell AI Data Platform reduces the strain on enterprise IT teams across both storage and the data foundation

