OpenAI Shocking Losses: Net Loss Surges to $38.53 Billion in 2025

·12 min read·CleanTechnica / Financial Times

💡 Tool TipFollowing AI industry developments? Try Evergreen Tools' JSON Formatter, Markdown Editor, and CSV to JSON — all free, no login required, runs locally in your browser!

July 4, 2026 AI News: Leaked audited financial statements reveal OpenAI's net loss surged from $5.09 billion in 2024 to $38.53 billion in 2025, a 650%+ increase, raising deep questions about AI business model sustainability.

According to CleanTechnica, leaked audited financial statements from OpenAI reveal a shocking figure: the creator of ChatGPT posted a net loss of $38.53 billion in 2025, up from $5.09 billion in 2024 — an increase of over 650%. The report, published by Ed Zitron and verified by the Financial Times, has once again pushed the AI industry's cash-burning problem into the spotlight.

Looking at the 2024 figures specifically: OpenAI generated $3.7 billion in revenue that year, but costs and expenses totaled $12.48 billion. This included $7.81 billion in R&D spending, $1.11 billion in sales and marketing, $907 million in general and administrative expenses, and $2.65 billion in cost of revenue. The operating loss reached $8.78 billion, with a net loss of $5.09 billion.

The 2025 figures are even more alarming. While revenue grew, the rate of cost and expense growth far outpaced revenue increases. Massive GPU procurement costs, data center construction expenses, top talent compensation, and global marketing investments combined to push losses to unprecedented levels. The $38.53 billion net loss means OpenAI was losing over $100 million per day on average.

These figures have sparked deep discussion about the sustainability of AI business models. Critics point out that while “AI” has become the hottest buzzword in today's business world, with every company touting its AI strategy, if even the industry leader is burning cash at this rate, where exactly is the profitability outlook for the entire industry?

Notably, OpenAI isn't the only AI company facing massive losses. Competitors like Anthropic, Google DeepMind, and xAI are all investing heavily in infrastructure and talent. The entire AI industry is in a “capture market first, discuss profitability later” phase, with investors betting on the possibility that AI will fundamentally transform every industry in the future.

However, market sentiment is undergoing subtle changes. More and more investors are questioning whether such massive investments can truly generate matching business returns. Has the capability improvement of large language models reached a point of diminishing marginal returns? Are enterprise customers willing to pay high enough prices for AI services to cover these astronomical operating costs?

Looking ahead, OpenAI is considering delaying its IPO until 2027, targeting a valuation of up to $1 trillion. But in the face of such enormous losses, whether this valuation target is achievable remains questionable. The AI industry stands at a critical crossroads: either find a sustainable profitability model or face the risk of a bubble burst. The second half of 2026 will be a crucial period for testing AI commercialization.

📌 Frequently Asked Questions

Why are OpenAI's losses so severe?

The main reasons are massive GPU procurement, data center construction, R&D investment, and talent compensation. Training and running large language models requires enormous computing power, with infrastructure costs far exceeding current revenue.

Could OpenAI go bankrupt because of this?

Unlikely in the short term. OpenAI has raised substantial funding, with investors betting on long-term potential. But if losses continue expanding while revenue growth slows, financial pressure will intensify.

What's the overall profitability outlook for the AI industry?

The entire industry remains in a massive investment phase. Enterprise AI applications are growing, but whether they can cover astronomical operating costs remains to be seen. 2026-2027 will be a critical validation period.