OpenAI's Own Numbers Show It Runs Dry By 2028

Published: September 21, 2026 | Updated: September 21, 2026 | Category: AI | By Mahesh

OpenAI has told investors it expects to burn through $278 billion in cash between 2026 and 2030 as it aggressively expands computing infrastructure, according to Reuters' September 18 report, citing a Financial Times account of an internal company presentation.^[1] The figure represents cumulative negative free cash flow over five years, meaning OpenAI's own spending plans outpace even its aggressive revenue growth targets by that amount, according to the same presentation.

The Two Numbers That Actually Define This Story

Reading the cash burn figure alongside OpenAI's own spending and revenue projections clarifies exactly what is driving it. According to BigGo Finance's detailed breakdown of the internal documents, OpenAI forecasts spending roughly $856 billion on computing power and infrastructure through 2030, its single largest expense category by a wide margin, while cumulative revenue over the same period is projected to reach $840 billion.^[2] Annual revenue itself is expected to climb nearly tenfold, from $36 billion this year to $350 billion in 2030, according to PYMNTS' reporting on the same presentation.^[3] Even with that scale of revenue growth built into the company's own projections, the infrastructure bill still outpaces it, producing the negative cash flow figure over the full five-year window.

The Runway Math That Should Worry Anyone Reading This Closely

The most consequential detail in this reporting is not the $278 billion figure itself, but what it implies about OpenAI's existing cash reserves relative to its own spending timeline. According to Runtime Wire's detailed analysis, OpenAI's March financing, which closed $122 billion in committed capital at an $852 billion post-money valuation, is on pace to be exhausted in 2028 at the company's planned spending rate.^[4] That is two years before OpenAI's own revenue trajectory is projected to reach its peak, and two years before the company's internal projections show it achieving positive cash flow at all. In plain terms, OpenAI's own math shows the company running out of money from its largest-ever funding round before the spending that round was meant to finance even produces the revenue growth needed to sustain it.

Why This Timing Directly Undercuts a Story Depth Grid Covered Days Earlier

This disclosure lands in the same window as, and adds hard numbers to, a story Depth Grid covered in detail on September 16, when OpenAI was reported to be negotiating a new funding round at a $1.5 trillion valuation just four days after Sam Altman cited AI safety concerns as his reason for delaying the company's IPO. Runtime Wire's reporting confirms that same $1.2 trillion figure directly, noting investors have discussed a valuation of at least $1.2 trillion, with OpenAI pushing for more, representing a 41% increase from its March post-money valuation. Reading the two stories together, the safety-based IPO delay Altman announced publicly and the internal $278 billion cash-burn projection now surfacing are not separate developments: they describe the same underlying situation from two different angles, a company that needs enormous, continuous private capital infusions and has a specific financial reason, not just a stated safety rationale, to prefer private fundraising over the disclosure obligations a public listing would bring.

The One Piece of Genuinely Good News in the Presentation

It would be inaccurate to read this disclosure as purely negative, and the reporting is explicit that the trajectory has actually improved recently. According to Runtime Wire's analysis, OpenAI's projections have gotten measurably better since May, when the company expected $305 billion in cumulative negative free cash flow over the same period, a $27 billion improvement in the revised forecast. The reporting attributes part of that improvement to recent model releases, which reportedly helped lift annualized revenue by about 20% in July. That improvement trend is worth taking seriously: a company whose own burn projections are shrinking quarter over quarter, even while remaining enormous in absolute terms, is telling a different story than one whose losses are actively widening.

What the Infrastructure Spending Is Actually Buying

The $856 billion infrastructure figure is not an abstract number, and BigGo Finance's reporting connects it to specific, named projects already underway. That spending includes the previously announced Stargate project, encompassing five new US data centers targeting 10 gigawatts of AI capacity. This connects directly to Depth Grid's own coverage of the broader AI infrastructure buildout, including Amazon's recent warrant-linked supply deals for backup power equipment, and to the pattern of equity-linked infrastructure financing Depth Grid has tracked across the industry throughout September. BigGo Finance's reporting notes the projections underscore the enormous capital demands facing the world's most prominent AI developer as it races to secure scarce computing capacity against rivals including Google, Anthropic and Meta, framing OpenAI's spending not as an isolated bet but as a competitive necessity given how the rest of the frontier AI industry is investing at comparable scale.

What to Watch Next

The clearest near-term signal will be whether OpenAI's negotiations for a new funding round actually close at the $1.2 trillion valuation investors have proposed, or the higher figure OpenAI is reportedly pushing for, since the size and terms of that round will determine how much runway the company buys itself beyond the 2028 exhaustion point its own March round is currently projected to hit. Also worth tracking is whether OpenAI's next round of model releases continues the revenue-acceleration pattern that helped narrow the cash-burn projection between May and this presentation, since sustained improvement on that front is the clearest evidence that the company's underlying bet, spending aggressively now to outrun declining AI prices later, is actually working as intended.

Read More on Depth Grid

• OpenAI Wants $1.5 Trillion. It Just Called Itself Unsafe.
• Amazon Just Gave a Generator Company a 40% Stock Pop
• Anthropic Hired a $70B Consultant to Check Its Own Safety
• Every Major AI Lab Has Now Confessed to the Same Thing
• An OpenAI Model Told Its Successors to Lie and Hide It

Article by Mahesh | Depth Grid