Last November, OpenAI investor Brad Gerstner pressed Sam Altman on a podcast about how a company with $13 billion in revenue could commit to $1.4 trillion in spending. Altman bristled.
"If you want to sell your shares, I'll find you a buyer," he said. "Enough."
Three months later, OpenAI is aiming to raise $100 billion in its latest funding round — a sign that, even amid mounting questions, Altman can find buyers.
Amazon, SoftBank, and Nvidia are all reportedly considering investments that could run into the tens of billions. It would be the largest fundraise in history, surpassing the last recordholder: the company's March 2025 fundraising round of $40 billion. And there are strong signals that it's weighing an IPO for later this year.
For all that momentum, the chatbot pioneer can't shake the central concern: It is spending money it doesn't have, at a scale that could overwhelm its backers if revenue doesn't offset its costs.
The urgency is mounting. On Monday, OpenAI began testing ads inside ChatGPT for free and low-paying users in the US, a sign it's looking for more ways to monetize. Rival Anthropic seized the moment, spending millions on Super Bowl ads the day before mocking the decision.
OpenAI's view is that scale itself will become an advantage, one large enough to overwhelm competitors and cement OpenAI's position at the center of the AI industry.
Whether it can make good on that all-important bet is the trillion-dollar question — and one that will be in front of millions of investors if the company goes public this year.
A 'cash incinerator'
For some investors, the uncertainty is reason enough to stay away.
"The company is not yet profitable and doesn't have a plausible pathway to near-term profitability," Charles Jaskel, founder of secondaries firm New Vintage Partners, told Business Insider. He wouldn't invest in this round if given the chance, he said.
"Markets change, and there is no guarantee that a technological edge, especially in this AI-driven end of the market, will endure."
Others have been more blunt. George Noble, a former Fidelity portfolio manager, likened OpenAI to a "cash incinerator" on X. Sebastian Mallaby, a senior fellow at the Council on Foreign Relations, wrote in a New York Times op-ed in mid-January that OpenAI will run out of money in the next 18 months.
OpenAI's losses will total $143 billion between 2024 and 2029, the "largest startup losses in history," Deutsche Bank analysts wrote in a December 4 note. HSBC researchers said in a late November report that they expect OpenAI to have a $207 billion shortfall by 2030, even when modeling for significant boosts in revenue.
In September, The Information reported that OpenAI was telling investors it would burn $115 billion of cash by 2029, more than three times the company's previous estimate.
Its largest corporate partner is feeling the pressure. Last month, Microsoft said nearly half of its cloud computing backlog was tied to OpenAI. The software giant's reliance on OpenAI helped wipe $440 billion from Microsoft's market value amid concerns that Altman's company may not deliver on its obligations.
At the center of OpenAI's spending is compute: the processing power required to train and run AI models. It has signed agreements for more than 30 gigawatts of capacity in the coming years — nearly a third of what JLL estimates the entire industry used last year.
"You should expect OpenAI to spend trillions of dollars on data center construction in the not very distant future," Altman reportedly told journalists at a dinner in August.
The case for confidence
For OpenAI and those investing in it, that spending isn't reckless — it's necessary.
In a January blog post, chief financial officer Sarah Friar said the company's relative lack of computing power had been holding back monetization. OpenAI had about 1.9 gigawatts at the end of the year, far short of what it needs to serve demand, she said.
"Compute is the scarcest resource in AI," Friar wrote. "Three years ago, we relied on a single compute provider. Today, we are working with providers across a diversified ecosystem. That shift gives us resilience and, critically, compute certainty."
Some investors say that the headline spending figures overstate OpenAI's true exposure.
Ethan Choi, an OpenAI investor at Khosla Ventures, wrote on X last month that of the $1.4 trillion in obligations cited by Altman, about $600 billion would be spent directly by OpenAI. The remaining $800 billion, he said, would be covered by partners building the data centers OpenAI will rely on.
Choi told Business Insider that the $600 billion figure is less daunting than it sounds — spread over a decade, it could be matched by the revenue OpenAI expects to generate over the same period.
Choi noted that OpenAI and its rival Anthropic have shown they can generate about $10 billion in annual revenue for each gigawatt of computing power. OpenAI should have about 14 gigawatts online in three years or so, Choi said, adding that the figure was his estimate based on public information. That means $140 billion in potential annual revenue.
Choi trusts Altman and Friar on how to deploy capital. They "have the most data in front of them," he said. "Whenever there is a big wave like this, there are always naysayers, or folks who don't see the big picture. I think we just need to step back and think about how we can apply AI to our lives, and it's endless."
An impressive run — and its limits
ChatGPT's launch was a watershed for consumer technology, hitting 100 million users in two months to achieve the fastest consumer adoption of any product in history. For a time, OpenAI seemed unstoppable.
It went on one of the most impressive growth runs in business history. The company hit $20 billion in annual revenue in roughly three years, torrid growth even in the blitzscaling world of tech.
Revenue is only half the equation.
"It basically did what Google did in seven years, in two," Rob Enderle, an independent technology analyst, told Business Insider. "So that's great, except for the problem of their capex."
"The infrastructure they're buying is undergoing massive advancements, with performance doubling, tripling, quadrupling," he said. "You can't build a data center fast enough to ensure that it's not going to be obsolete by the time you finish it."
One way to fund that burn: Go public. Multiple outlets, including Business Insider, have reported there are signs OpenAI is aiming to IPO later this year. That would give OpenAI access to a far deeper pool of capital than private markets can offer, but it also means quarterly earnings calls, public scrutiny of its losses, and increased pressure to show a path to profitability.
A competitive market for IPOs adds to the pressure. Last week, Elon Musk merged xAI with SpaceX, a company widely expected to go public this year. Reporting also suggests Anthropic, OpenAI's closest competitor, is eyeing going public by the end of the year.
A tightening race
There are also questions about focus and competition. Over the past year, OpenAI has shifted between expansion and retrenchment.
Last May, OpenAI named Instacart chief Fidji Simo as CEO of applications, signaling how serious the company is about expanding beyond its core chatbot. In December, Altman called a "Code Red" in an internal memo, telling employees the company had spread itself too thin and needed to refocus on improving ChatGPT.
Two months later, the company is pushing deeper into search and ecommerce with a new checkout feature, considering adult content, working on a personal device with Apple's former design chief, and rolling out ads — something the "Code Red" memo explicitly hit pause on. It's also exploring robotics and designing its own chips.
Meanwhile, rivals are closing in. On the consumer side, Google's Gemini has been steadily eating into ChatGPT's lead, according to data from Apptopia. On the enterprise side, Anthropic has carved out a strong position, particularly with software developers and businesses looking for specialized AI coding tools.
It's no longer clear that ChatGPT's consumer dominance alone will be enough to generate the kind of returns OpenAI's valuation demands.
Andrej Karpathy, an OpenAI founder who is now building his own startup, fawned over Anthropic's Claude on X.
"This is easily the biggest change to my basic coding workflow in ~2 decades of programming and it happened over the course of a few weeks," he wrote. "I'd expect something similar to be happening to well into double digit percent of engineers out there."
OpenAI's version, Codex, has not received the same level of adulation.
There's also competition on the model approach. Open-source models from China and elsewhere are delivering rapid performance gains at a fraction of the cost.
"Our biggest concern is competitive intensity," Kyle Qi, an investor at Llama Ventures, which backs Anthropic, Thinking Machines Lab, and xAI, told Business Insider. "Google and leading Chinese LLMs are rapidly catching up and, in some areas, already outperforming on cost, latency, and specific benchmarks. That convergence compresses OpenAI's technological moat at a time when expectations for durable dominance are priced in."
Jeremy Abelson, the founder and CEO at Irving Investors, which has investment exposure to OpenAI, told Business Insider that "OpenAI needs to spend a large amount of money on capex now because key competitors like Google are investing in-line to larger capex at this point to remain at the top of the LLM leaderboard."
He added that if the Altman-led company doesn't keep ahead of the innovation curve, "it risks decreasing market share and losing relevance."
'A lot of things need to go right'
Until fundraising is finished, it's unclear what OpenAI's valuation will be, but estimates reported by other news outlets range from $750 billion to $830 billion. Whether its ultimate valuation makes sense may depend less on the brilliance of OpenAI's models than on the discipline of its capital allocation. The margin for error is thin.
"At this valuation, a lot of things need to go right," said Steve Brotman, managing director and founding partner at Alpha Partners, who said he would pass on OpenAI at this stage for less risky opportunities. "Use the product if you enjoy it. That doesn't mean you need to buy the stock."
Altman himself has been upfront on the stakes.
"Someone is going to lose a phenomenal amount of money," he said at the August dinner. "And a lot of people are going to make a phenomenal amount of money."
Now the market will decide.
