OpenAI Posts $21.3B Net Loss in Q1, $3.7B Cash Burn Still Exceeds Half of Revenue
0xBroomberg
OpenAI posted $5.7 billion in Q1 revenue — triple the year-ago quarter — yet burned through $3.7 billion in cash, more than half of what it earned. The AI demand boom has not delivered a matching improvement in profitability, a hard constraint on the company's looming IPO valuation.
Revenue tripled — so why is it still losing money?
Q1 revenue hit $5.7 billion, tripling year-over-year, but net loss reached $21.3 billion.
This means → revenue is expanding fast, but costs are expanding faster; scale has not yet outrun the cost curve.
Nearly $12.4 billion of that loss is a non-cash accounting charge — fair-value swings on convertible interest rights and warrant liabilities tied to equity issued to investors and the OpenAI foundation. No cash actually left the building.
Strip that out and operating loss still stands at $9.3 billion, including over $2.3 billion in stock-based compensation — more than double the year-ago figure.
How fast is the cash actually burning?
Cash burn for the quarter: $3.7 billion — for every dollar earned, $0.65 went out the door.
The company previously guided full-year burn of roughly $25 billion, rising to $57 billion in 2027. Annualizing Q1 gives about $14.8 billion — below its own forecast.
In plain terms = one quarter of slower burn does not set the trend. Model-training spend swings heavily between quarters.
Gross margin improved — what does that actually tell us?
Gross margin rose from 33% a year ago to 39%, with cost of revenue — mostly model inference costs — at roughly $3.5 billion.
This means → scaling is delivering some efficiency; the marginal cost of serving each user request is falling.
But R&D expense — which includes model training — hit $8.6 billion, dwarfing cost of revenue. The real spending weight is in building new models, not running existing ones.
What do $665 billion in off-balance-sheet commitments mean?
As of end-2025, OpenAI's compute-purchase commitments with cloud providers totaled roughly $665 billion, extending to 2030.
In plain terms = this money does not appear on the balance sheet, but it is a locked-in, long-term bill. However the business evolves, the payments are contractually due.
This reflects a reality public-market investors will scrutinize: the true financial burden on AI companies far exceeds what the income statement shows.
How much cash is on hand, and how long does it last?
Cash and marketable securities topped $73 billion at quarter-end, up sharply from $40 billion at end-2025 — largely from a major funding round closed in late March.
This means → at the current burn rate, the cash pile covers roughly two years. No near-term need to raise again.
That also eases pressure to rush the IPO — the listing timeline is in the company's hands, not forced by a cash crunch.
What is the biggest test for the IPO?
OpenAI filed a confidential listing application with regulators last week, but said it will pace next steps by the review timeline.
The core tension: revenue growth is extraordinary, but the path to profitability remains unclear and off-balance-sheet commitments are enormous.
Put simply = public-market investors must answer one question — what is a company worth when it is on track for $20 billion-plus in annual revenue, yet burns tens of billions a year and carries a $665 billion tab it cannot walk away from?
Content is for reference only, not financial advice.