OpenAI reported Q1 2026 revenue of USD 5.7 billion and cash burn of USD 3.7 billion, both approximately tripling year-over-year. The company posted an operating loss of USD 9.3 billion in Q1, with R&D expenses reaching as high as USD 8.6 billion, while gross margin improved slightly to 39%. Of greater concern is OpenAI’s committed spending on cloud computing capacity, which amounted to USD 665 billion as of the end of 2025—most of which remains off-balance-sheet—and will be a key risk scrutinized by investors ahead of its IPO.
The stronger the AI demand, the faster the cash burn.
According to a June 16 report by The Information, documents disclosed by OpenAI to shareholders show that in the first quarter of 2026, the company generated $5.7 billion in revenue, while its cash burn reached $3.7 billion during the same period—equivalent to burning through more than half of its revenue. Both figures were roughly triple those of the same period a year earlier.
The timing of this data release is particularly sensitive—OpenAI announced last week that it had secretly filed an IPO application but simultaneously indicated that, pending regulatory review, it might not rush to proceed with the listing. Against this backdrop, these financial figures will become a focal point for scrutiny by public market investors.
As of the end of Q1, OpenAI held over $73 billion in cash and marketable securities on its balance sheet, a significant increase from $40 billion as of the end of December, primarily driven by a large-scale funding round completed in late March. At its current cash burn rate, the company does not need to raise additional capital in the near term, which somewhat alleviates pressure to accelerate its IPO timeline.
OpenAI forecasts its total cash burn for this year will reach $25 billion, rising further to $57 billion next year.
Behind the staggering loss figures: bloated accounting expenses mask the real pressure—operating losses
The net loss for Q1 exceeded $21.3 billion—a shocking figure—but approximately $12.4 billion of this consisted of non-cash accounting charges stemming from fair value adjustments related to convertible interest rights and warrant liabilities, tied to equity issuances to investors and the establishment of the OpenAI Foundation.
Excluding these accounting adjustments, OpenAI’s operating loss for Q1 stood at $9.3 billion, including over $2.3 billion in employee stock-based compensation expenses—more than double the amount recorded in the same period a year earlier.
This means that even without accounting adjustments, OpenAI’s actual operational losses remain close to the $10 billion mark per quarter.
Gross margin edged slightly higher, but R&D spending continues to consume all available room
Q1 revenue costs (primarily model inference costs) amounted to USD 3.5 billion, resulting in a gross margin of 39%, an improvement from 33% in the same period last year. This indicates that as scale expands, the per-unit service cost is declining, and operational efficiency is marginally improving.
However, above gross profit lies a major challenge: Q1 R&D expenses reached USD 8.6 billion, covering core investments such as model training.
A simple calculation shows: USD 5.7 billion in revenue minus USD 3.5 billion in cost of revenue yields approximately USD 2.2 billion in gross profit—yet R&D alone consumed USD 8.6 billion. This explains why, despite improving gross margins, operating losses remain substantial.
A USD 665 billion 'off-balance-sheet time bomb'
For potential IPO investors, a figure outside the balance sheet may warrant even closer attention.
OpenAI disclosed to shareholders that, as of the end of 2025, its estimated commitments to cloud providers for computing capacity are as high as USD 665 billion, extending through 2030. The majority of these commitments are not reflected on the balance sheet.
This means OpenAI has effectively pre-committed to massive computing bills for the coming years—regardless of whether AI demand grows as anticipated. Should AI adoption slow or model efficiency improvements fall short of expectations, this will become a significant fixed burden.
Editor/Lambor