OpenAI's $14 Billion Hole: Can the World's Most Valuable AI Startup Survive Its Own Ambition?

The Scale of the Burn

OpenAI — valued at approximately $730 billion as of early 2026, according to Al Jazeera — has raised more than $168 billion from investors. It spends that money faster than almost any startup in history.

Financial documents obtained by The Wall Street Journal and reported by Fortune in November 2025 showed that OpenAI projected roughly $22 billion in total spending against $13 billion in sales for 2025, resulting in a net loss of approximately $9 billion. That means the company spent approximately $1.69 for every dollar of revenue generated in 2025.

The trajectory worsens from there. Internal projections, also reported by The Wall Street Journal through Fortune, show OpenAI anticipates its cash burn remaining at 57 percent of revenue through 2026 and 2027. For 2026 specifically, the company's own forecasts — as reported by The Information and separately analyzed by RD World Online — project losses of approximately $14 billion, roughly three times the losses estimated for 2025.

By 2028, OpenAI projects operating losses that amount to roughly three-quarters of that year's revenue, driven primarily by escalating compute costs, according to the Wall Street Journal documents reviewed by Fortune.

Where the Money Goes

The primary cost driver is computing infrastructure. OpenAI has committed to spending up to $600 billion on compute by 2030, according to figures the company shared with investors in February 2026, as reported by CNBC. That figure represents a revision downward from the $1.4 trillion Sam Altman had been citing publicly; OpenAI told investors the updated number was intended to more directly tie projected compute spend to expected revenue growth, per CNBC.

The company is spending "almost $100 billion on backup data-center capacity alone," according to Wall Street Journal documents reviewed by Fortune — essentially insurance against demand surges from future products.

An OpenAI spokesman, quoted by The Wall Street Journal through Fortune, described the rationale: "Demand for AI exceeds available compute supply today. Every dollar we invest in AI infrastructure goes to serving the hundreds of millions of consumers, businesses, and developers who rely on ChatGPT to get more done."

Analyst Richard Windsor, described by The Telegraph as an independent tech analyst, estimated that OpenAI's spending on computing has risen at the same pace as its revenues for the last three years — meaning the company has not improved its economic efficiency as it scaled.

Microsoft, which had committed $13 billion in funding to OpenAI, had funded $11.6 billion of that total by September 30, 2025, according to Microsoft's own financial disclosures as reported by Where's Your Ed At. Nvidia CEO Jensen Huang said in early March 2026 that Nvidia would invest another $30 billion into OpenAI but characterized it as potentially the last such investment before an IPO, per Al Jazeera. Huang also said a previously touted $100 billion infrastructure commitment was "not in the cards," according to Al Jazeera.

The Revenue Problem

OpenAI's ChatGPT counted more than 900 million weekly active users as of March 2026, according to CNBC's reporting on an internal all-hands meeting led by Fidji Simo, OpenAI's CEO of Applications. But an estimated 95 percent of users use the free service, according to The Telegraph's analysis, meaning the vast majority of that user base generates no direct revenue.

The company sells monthly subscriptions for access to more powerful models and additional features, but the math remains unfavorable. To justify its projected valuations and planned spending, OpenAI needs to generate $200 billion in annual revenue by 2030, according to financial analyst George Noble, quoted in Al Jazeera. OpenAI's own internal projections show it expects more than $280 billion in revenue by 2030, with roughly equal contributions from consumer and enterprise segments, per CNBC.

Sebastian Mallaby, a senior fellow at the Council on Foreign Relations, told Al Jazeera that OpenAI lacked the capital reserves needed to sustain its current build-out phase until it reaches the revenue threshold required for profitability. Mallaby had previously written an op-ed in The New York Times forecasting the company could run out of money within 18 months from that writing.

Aleksandar Tomic, associate dean for strategy at Boston College, told Al Jazeera: "I'd say the only thing worse than losing money with OpenAI is being left behind entirely." He described continued investment as partly driven by fear of missing out on the technology rather than confidence in the business model.

The Sora Shutdown: A Case Study in Failed Product Strategy

OpenAI's decision to shut down Sora, its AI video generation app, in late March 2026 crystallized the company's financial dilemma. The app had skyrocketed to the top of the App Store within days of its September 2025 release, according to The Atlantic. Disney had announced a $1 billion investment in OpenAI as part of a licensing deal to bring its characters to Sora — and then retracted those investment plans following the shutdown announcement, per The Atlantic.

Bill Peebles, who led the Sora team, said the product's economics were "completely unsustainable," according to The Atlantic. Forbes had previously estimated Sora was costing OpenAI millions of dollars daily in compute costs, a figure OpenAI did not dispute at the time, per The Atlantic. Generating video is significantly more compute-intensive than generating text.

The shutdown was not isolated. The Atlantic documented a broader pattern: OpenAI launched a shopping feature allowing purchases inside ChatGPT, then killed it and pivoted to product discovery. The company said in January 2026 that its consumer hardware device was on track for that year; court filings weeks later revealed it would not launch before 2027. OpenAI declared in 2024 that advertising would be a last resort, then launched an ad initiative in early 2026 targeting free and lower-tier users.

In December 2025, OpenAI declared what it called a "code red" effort to improve ChatGPT amid stiffening competition from Google and Anthropic, temporarily pulling back investment in health, shopping, and advertising initiatives, according to CNBC.

The IPO Pressure

OpenAI is targeting an IPO as soon as the fourth quarter of 2026, according to a person familiar with the matter cited by CNBC. CFO Sarah Friar is building out the company's finance team ahead of the offering: she hired Ajmere Dale, former chief accounting officer at Block, and Cynthia Gaylor, former CFO of DocuSign, who will oversee investor relations, according to CNBC and LinkedIn posts reviewed by the outlet.

In a March 2026 all-hands meeting reported by CNBC, Fidji Simo said OpenAI was "orienting aggressively" toward high-productivity enterprise use cases. "Our opportunity now is to take those 900 million users and turn them into high-compute users," Simo said, according to a partial transcript reviewed by CNBC. "We'll do that by transforming ChatGPT into a productivity tool."

CFO Sarah Friar separately said the company had healthy margins and could break even if it chose to, according to Fortune's reporting. She highlighted rapid growth in OpenAI's enterprise business as the core path to profitability.

An IPO would expose OpenAI's financials to public scrutiny for the first time. Public market investors would see a company projecting $14 billion in losses for the current year and $74 billion in operating losses by 2028 — before an expected pivot toward profit in 2029 or 2030, per the Wall Street Journal documents reported by Fortune.

The Competitive Context

OpenAI's financial position looks notably worse relative to its main rival. Anthropic — which also burns cash at high rates — projects dropping its cash burn rate to roughly one-third of revenue in 2026 and down to 9 percent by 2027, per Fortune's reporting on the Wall Street Journal documents. Anthropic expects to break even by 2028, the same year OpenAI projects its operating losses will peak.

Anthropic's strategy deliberately avoids compute-heavy products like video generation, focusing instead on corporate customers who account for about 80 percent of Anthropic's revenue, according to Fortune. Google and Microsoft — both with entrenched enterprise sales forces — are competing for the same business customers OpenAI is targeting ahead of its IPO.

OpenAI is carrying approximately $100 billion in debt as of early 2026, according to Al Jazeera. The cumulative cash burn is expected to reach $115 billion through 2029, based on projections reported by The Information.

Michael Ashley Schulman, partner and chief investment officer at Running Point Capital Advisors, told Al Jazeera: "To become profitable, they really need to transition from what is essentially a subsidised research laboratory to an enterprise software juggernaut."

What the Numbers Require

For OpenAI's trajectory to make sense, several things must happen simultaneously: compute costs must fall faster than they have historically, enterprise revenue must scale dramatically faster than consumer revenue, and market demand for AI services must remain strong enough to absorb pricing increases. The company's own projections assume roughly 15-fold revenue growth — from approximately $13 billion in 2025 to more than $200 billion by 2030 — while still spending $115 billion cumulatively before generating positive cash flow.

Richard Windsor, the independent tech analyst cited by The Telegraph, framed the core problem simply: if compute spending rises at the same pace as revenue, the fundamental economics do not improve. Breaking even requires either extracting significantly more revenue per user or dramatic reductions in the cost of running AI models.

The company's planned IPO, if it proceeds in 2026, will force public investors to decide whether that bet — and the losses required to pursue it — is worth taking.