OpenAI’s third-quarter revenue surged nearly 50% to $10 billion, but the ChatGPT developer’s adjusted operating losses widened to $10.9 billion, underlining the enormous cost of competing in artificial intelligence.
The loss increased from $10.1 billion in the previous quarter, according to investor documents reviewed by The Wall Street Journal.
Yet the company’s adjusted operating margin improved substantially, from around negative 150% to negative 109%, as sales grew much faster than losses.
The figures come as OpenAI raises capital at a reported $1.25 trillion valuation and faces mounting questions over how quickly its expanding business can become profitable.
OpenAI Q3 2026 earnings: Why losses are growing despite record revenue
OpenAI generated approximately $6.7 billion in second-quarter revenue, making the latest increase particularly striking. Its adjusted operating loss rose about 8%, compared with revenue growth approaching 50%.
That difference signals improving operating leverage, as each additional dollar of sales is accompanied by a smaller proportionate increase in losses. Nevertheless, the third-quarter loss still exceeded revenue by $900 million, leaving profitability a distant prospect.
Chief financial officer Sarah Friar described the period as a blockbuster quarter in an interview with Yahoo Finance.
“Our enterprise business doubled quarter over quarter,” she said, adding that enterprise customers now account for roughly half of revenue, alongside consumers.
Demand for coding tools, cybersecurity products and AI assistants is helping OpenAI expand beyond individual ChatGPT subscriptions.
Corporate contracts could support steadier recurring income, although their economics depend on the expense of delivering the underlying computing services.
The adjusted figures are not the same as net losses or cash consumed, and OpenAI has not released a complete breakdown of the latest quarter’s expenses.
Where OpenAI’s billions are going
Training advanced models requires immense computing capacity. Serving responses to customers also consumes costly processing power, even after the initial research is completed.
The company must fund researchers, specialised chips, cloud capacity and commercial expansion simultaneously. Earlier financial documents showed $8.6 billion in first-quarter research and development expenditure and $3.5 billion in cost of revenue.
In May, OpenAI president Greg Brockman said the company expected to spend around $50 billion on computing power during 2026.
Bernstein analyst Mark Moerdler estimated that its identified infrastructure commitments could translate into roughly $614 billion in spending through 2030, according to Investing.com.
That estimate assumes projects proceed on schedule; delays, cancellations or flexible contractual terms could change actual payments.
The critical distinction is between promised computing capacity, expenses recognised in financial accounts and cash already paid.
OpenAI AI infrastructure spending
Bloomberg reported last week that OpenAI had secured approximately $11 billion in the first close of a fundraising round valuing it at $1.25 trillion.
The broader target has reportedly fallen to around $20 billion from an earlier ambition of at least $30 billion, amid cooler sentiment towards AI infrastructure.
OpenAI has also delayed plans to list publicly, with a possible flotation in 2027.
D.A. Davidson analyst Gil Luria told MarketWatch in October that “OpenAI is actually accelerating as it catches up with Anthropic at the enterprise market.”
His assessment preceded the latest results, but the subsequent doubling of enterprise revenue lends weight to the argument.
There is a counterargument. IDC executive Eric Newmark has questioned whether circular financing between AI investors and infrastructure suppliers makes headline spending commitments a reliable measure of end-user demand.
OpenAI’s improving margin offers evidence that rapid growth can make the business more efficient.
But investors still need convincing evidence that revenue can eventually cover computing costs and deliver sustained profits and positive cash flow.
AI disclosure: AI tools assisted with research and drafting of this piece. An Invezz editor reviewed the facts, sources and wording before publication. Any opinions or analysis are Invezz’s, not those of an automated system.
The post OpenAI's $10 billion revenue milestone hides a much bigger financial shock appeared first on Invezz