Nvidia Margins Squeeze OpenAI as AI Capex Hits $280B
Nvidia and Micron hold 75-85% margins while OpenAI reports a $21B loss, creating a fragile dependency in the AI supply chain.
Key points
- OpenAI lost $21 billion in 2025 and expects to burn $280 billion by 2030 on infrastructure.
- Nvidia and Micron hold 75% and 85% gross margins, increasing costs for AI developers.
- The S&P 500 CAPE ratio is at 41, near the 1999 dot-com peak of 44.
The S&P 500 and Nasdaq Composite are trading near all-time highs, supported by a cyclically adjusted price-to-earnings ratio of 41. This valuation level approaches the 1999 dot-com peak of 44, yet the current market is driven by profitable infrastructure leaders rather than unproven startups. According to reporting via AOL.com, the sustainability of this rally depends on whether frontier AI labs can convert massive hardware purchases into sustainable consumer revenue.
The financial structure of the AI sector shows a stark divergence between suppliers and buyers. Infrastructure providers like Nvidia and Micron maintain gross margins of 75% and 85% respectively. In contrast, leading model developer OpenAI reported a loss of $21 billion against $13 billion in revenue for the full year 2025. This gap indicates that the high costs of computing power are not yet being offset by end-user monetization.
OpenAI faces massive capital burn
OpenAI’s internal financial documents reveal a critical cash flow challenge. Management expects to burn through approximately $280 billion in cash by the end of 2030, with the majority allocated to computing power and infrastructure. The company’s current R&D expenditures are driven by the need to maintain technical leadership against competitors, creating a persistent pressure on its balance sheet.
The high gross margins retained by hardware suppliers like Nvidia directly increase the cost base for AI developers. This dynamic makes it difficult for frontier labs to achieve sustainable unit economics. As Chinese rivals close the technical gap, the premium pricing of Western infrastructure may become less justifiable for AI companies struggling with negative free cash flow.
Market valuation exceeds historical averages
The CAPE ratio of 41 significantly exceeds the historical average of 17.4, signaling elevated risk. While the current market is underpinned by established firms with real profits, the underlying demand is tied to the speculative success of AI applications. If the monetization of these algorithms fails to materialize, the revenue foundation for infrastructure providers could erode.
Historical precedents suggest that buying at the peak of such valuations can lead to prolonged negative returns. For example, an investment in the Nasdaq on March 10, 2000, did not regain its initial value until April 2015. The current market faces similar structural risks, compounded by external geopolitical pressures that may further disrupt supply chains or investor confidence.
Geopolitical risks threaten trade stability
Political and economic instability adds a layer of uncertainty to the market. Inflationary policies, including wide-ranging tariffs on major trading partners, are increasing operational costs for global technology firms. Additionally, the partial blockage of the Strait of Hormuz due to military operations in Iran poses a direct threat to energy and shipping logistics, which are critical for data center operations.
These external factors create a volatile environment for companies relying on stable global supply chains. The combination of high domestic valuations, intense competition from low-cost foreign rivals, and geopolitical friction could accelerate a correction in the AI sector. Investors are increasingly scrutinizing whether the current capital expenditure cycle can be sustained without significant revenue growth from end-users.
