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

In early July 2026, leaked audited financial statements from OpenAI caused an uproar in the tech industry. According to financial data published by Ed Zitron and verified by the Financial Times, OpenAI's net loss surged from $5.09 billion in 2024 to $38.53 billion in 2025, an increase of 657%. This shocking figure not only stunned investors but also raised deep questions across the industry about the sustainability of artificial intelligence business models. In today's world where AI concepts are wildly追捧, does OpenAI's financial predicament sound an alarm bell for the entire industry?

Financial Data Breakdown: The Shocking Jump from $5B to $38.5B

According to the leaked audit report, OpenAI's financial situation in 2024 was already quite severe: $3.7 billion in revenue, $12.48 billion in total costs and expenses, $8.78 billion in operating losses, and $5.09 billion in net losses. Among these, R&D spending reached $7.81 billion, sales and marketing expenses were $1.11 billion, and general administrative costs were $907 million. However, the 2025 figures are even more staggering. While specific revenue numbers haven't been fully disclosed, the net loss jumped directly from $5.09 billion to $38.53 billion, meaning OpenAI's losses increased by $33.44 billion in just one year.

This astronomical increase in losses stems from several factors: first, massive investments in data centers and computing infrastructure. Training and running frontier models like GPT-5.6 requires enormous GPU resources — AWS recently raised rental prices for some NVIDIA GPUs by approximately 20%, reflecting the continued strain on AI computing demand. Second, soaring compensation costs from talent competition, with top AI researchers commanding annual salaries in the millions. Third, global expansion and marketing expenses, as OpenAI builds data centers and sales teams worldwide.

IPO Delay: OpenAI's Capital Dilemma

Facing such massive losses, OpenAI's IPO plans have also encountered major uncertainties. According to Forbes, despite having confidentially filed S-1 documents, the company is considering delaying its listing until 2027. This decision is driven by multiple considerations: first, the current financial situation may not support the desired valuation; second, market enthusiasm for AI companies is cooling, with investors paying more attention to profitability and cash flow; third, regulatory uncertainty is increasing, with governments strengthening oversight and control of frontier AI models.

Meanwhile, OpenAI's competitor Anthropic has actually surpassed it in valuation. Anthropic's recent private market valuation reached $965 billion, exceeding OpenAI's private valuation, and has also filed its own S-1 documents. This means the competitive landscape in the AI industry is undergoing subtle changes — OpenAI is no longer the undisputed leader. Anthropic's Claude Mythos 5 and Fable 5 models have performed excellently in multiple benchmarks, and despite facing government regulatory challenges, their technical capabilities have been recognized by the market.

Concerns About AI Bubble Bursting

OpenAI's massive losses have sparked widespread discussion about whether the AI bubble is about to burst. Media outlets like CleanTechnica have bluntly pointed out that people are becoming skeptical about whether AI can truly justify its enormous costs. Building massive data centers, hiring top engineers, purchasing expensive GPUs — can these investments generate sufficient revenue to justify them? Looking at OpenAI's financial data, the answer appears to be no. The $38.53 billion loss in 2025 means that even with rapid revenue growth, it falls far short of covering costs.

However, there are also views that these losses are strategic and temporary. The AI industry is in an infrastructure building phase, similar to the massive investments made by early internet companies. Once models are trained and infrastructure is built, marginal costs will drop significantly. Additionally, AI application scenarios are expanding rapidly — from ChatGPT's consumer applications to Codex's enterprise coding tools, to various vertical AI solutions, revenue sources are diversifying. OpenAI's Codex grew its active user base more than 5x in the first half of 2026, with non-developer adoption rising 137x for individual users.

Advertising Business: A New Revenue Attempt

To alleviate financial pressure, OpenAI has started selling ads in ChatGPT. While this move may bring new revenue sources, it also raises concerns about privacy and data usage. When users ask for product recommendations or seek advice in ChatGPT, will AI responses be influenced by advertisers? This business model is similar to search engine advertising, but AI's personalized recommendation capabilities make it even more influential. Enterprises need to establish clear rules about what data can be collected by AI tools and how that data shapes what users see.

The launch of advertising also reflects OpenAI's exploration on the path to commercialization. Beyond subscription revenue (ChatGPT Plus and Enterprise), the company is experimenting with multiple monetization approaches. However, there's a potential conflict between the advertising model and AI assistant objectivity and credibility. Users expect AI to provide neutral, accurate advice, but if responses are driven by commercial interests, that trust could be undermined. Finding the right balance between commercialization and user experience will be an important challenge for OpenAI.

Industry Insights: The Future of AI Business Models

OpenAI's financial predicament provides important lessons for the entire AI industry. First, pursuing technical leadership without considering commercial viability is unsustainable. Even the most advanced AI models will face existential crises if they cannot generate sufficient revenue to cover costs. Second, AI companies need to be more transparent about their financial状况, allowing investors and the public to understand the true economic state of the industry. Third, regulators and investors should maintain rationality regarding AI company valuations, avoiding resource misallocation caused by excessive hype.

Looking ahead, the AI industry may undergo a shakeout. Companies that cannot achieve commercialization will be eliminated or acquired, while those that truly find sustainable business models will stand out. This may include: more efficient model training techniques to reduce computing costs, more precise application scenarios to improve paid conversion rates, and more reasonable pricing strategies to balance revenue and user growth. Meanwhile, the development of open-source AI models will also pressure closed-source companies, forcing them to demonstrate the value of their premium services.