OpenAI's Financial Crisis and AI Market Share Decline

OpenAI is facing significant financial challenges, with a Q1 2026 operating margin of -122%, meaning it spent over two dollars for every dollar earned. The company is projected to undershoot its ad revenue targets by 90% and has seen its market share in key areas decline. Sam Altman, OpenAI's CEO, acknowledged these difficulties, stating, "OpenAI has not had our best 12 months ever, and it's mostly my fault."

The Shifting Landscape of AI Market Share

Despite ChatGPT having over 900 million weekly users, OpenAI's dominance is eroding. The initial tipping point was not Anthropic, but Google's Gemini 3, which outperformed GPT-5.1 across benchmarks in November 2025, prompting a "code red" at OpenAI. While OpenAI still boasts a large user base (1.1 billion users compared to Gemini's 662 million and Claude's 245 million), its control in lucrative markets is shrinking.

In May, ChatGPT's global assistant share dropped below 50% for the first time, settling at 46%, with Gemini at 28% and Claude at 10%. The situation is more dire in enterprise spending, which is a major revenue driver.

Enterprise and Coding Market Decline

OpenAI's enterprise share has plummeted from 50% in 2023 to 27%, while Anthropic's share has grown to 40% (up from 12% in 2023). Google also saw substantial gains, increasing its enterprise share from 7% in 2023 to 21% in 2025.

This decline is particularly evident in the coding market, where Large Language Models (LLMs) have shown significant returns. Anthropic now commands 54% of the coding market, leaving OpenAI with just 21%.

Computing Power Constraints

A contributing factor to OpenAI's struggles is the delay and restraint of computing power. Approximately 40% of US data centers planned for 2026 are already delayed, and transformer lead times have extended to five years in some regions. OpenAI has also canceled planned sites in the UK, Norway, and Lordstown, Ohio.

The Unfulfilled Promise of Ad Revenue

OpenAI, like other AI companies, has looked to advertising as a potential savior, especially for monetizing its vast free user base (95% of its 900-910 million weekly active users). OpenAI had set an ambitious target of $100 billion in ad revenue for 2030.

However, OpenAI's ad business is on track to miss this target by 90%. Standalone AI chatbots, including ChatGPT, Gemini, and Copilot, are collectively projected to generate less than $1 billion in ad revenue in 2026, with a market-wide forecast of only $5.41 billion by 2030. OpenAI's individual 2030 target is roughly 20 times larger than eMarketer's estimate for the entire US chatbot ad market.

Challenges in AI Advertising

Several factors contribute to the poor performance of AI advertising:

• Irrelevant Prompts: Only about 2% of ChatGPT prompts involve purchasable products, making it difficult to display relevant ads.
• Low Click-Through Rates: ChatGPT ads have a click-through rate of 0.91-1.3%, significantly lower than Google search ads (around 29.2%), where searches are inherently more commercial.
• Primitive Ad Tech Stack: Advertisers describe the current AI ad tech stack as "primitive," and CPMs (cost per mille) have collapsed from $60 to as low as $25.
• Infrequent Usage: 80% of users sent fewer than 1,000 messages in 2025 (an average of less than 3 prompts per day), limiting exposure to ads.

Financial Losses and Capital Commitments

OpenAI is facing substantial financial losses. It forecasted a $14 billion loss for all of 2026, with total losses reaching $44 billion from 2023 to 2029. The Q1 2026 results showed a $6.95 billion operating loss on $5.7 billion in revenue, resulting in a -122% operating margin. Including stock compensation, the GAAP operating loss for Q1 was $9.3 billion, leading FutureSearch to forecast a $33 billion GAAP loss for 2026.

While revenue grew from $6 billion to roughly $25 billion ARR over 17 months to May 2026, it has since plateaued. OpenAI's internal target for mid-2027 is $62 billion, but the forecast median is $44 billion. ChatGPT's weekly active users have also plateaued around 900 million, with only 50 million paying subscribers.

Massive Capital Commitments

OpenAI has made enormous capital commitments:

• Oracle: $300 billion over 5 years for cloud capacity (4.5 GW).
• AWS: Initially $38 billion over seven years for Nvidia GB200 and GB300 GPUs, expanded by another $100 billion over eight years.
• Microsoft Azure: $250 billion for cloud services.
• Broadcom: An estimated $350-$500 billion for 10 GW of custom AI accelerators, where OpenAI will design its own chip.

In total, OpenAI's CFO, Sarah Friar, stated that current commitments amount to $750 billion through 2030, with more spending beyond that.

Funding and Runway

Despite these massive bills, OpenAI is also receiving significant funding. A recent $110 billion funding round saw Amazon invest $50 billion, Nvidia $30 billion, and SoftBank $30 billion. OpenAI's cash reserves and marketable securities stand at $73 billion, up from $40 billion at the end of 2025.

Worldwide AI spending is projected to reach $2.59 trillion in 2026, a 47% increase over 2025. The "Magnificent Seven" tech companies plan over $700 billion in AI capital expenditure for 2026, up from $400 billion in 2025.

The Commoditization of AI

Sam Altman's comments at the BlackRock Infrastructure Summit in March 2026 hinted at a significant shift in the AI industry, drawing parallels to the commoditization of electricity and the internet.

Lewis Strauss's 1950s prediction of "electrical energy too cheap to meter" for nuclear power never materialized. Instead, electricity became a tightly regulated commodity with thin margins, with value accumulating in infrastructure providers like turbine makers and transmission networks.

Similarly, the internet became a commodity, where a gigabit from one provider is indistinguishable from another. Value shifted from infrastructure equipment manufacturers (like Cisco, which was briefly the most valuable company in 2000) to consumer-facing services like Google and Netflix, which leverage user data and high switching costs.

AI Following a Similar Path

The AI industry is exhibiting a similar pattern:

• GPU Scarcity, Model Abundance: In 2023, both GPUs and AI models were scarce. By 2026, GPUs remain scarce, but AI models are not, and their performance gaps are narrowing. The gap between the best closed model and the best open-weight model on Chatbot Arena fell from 8.04% in January 2024 to 1.70% in February 2025.
• Value Down the Stack: The scarce asset is no longer the models themselves, but the ability to run them efficiently (throughput, latency, cost per token, routing, caching, fine-tuning, evaluation, and guardrails). The value is accumulating at the bottom of the stack, in semiconductors and compute.
• Nvidia's Dominance: Nvidia's data center revenue hit $75.2 billion, up 92% year-over-year, with a 75% gross margin, indicating an annualized run-rate of approximately $300 billion. The money is in selling GPUs, semiconductors, and compute, not necessarily in "intelligence."
• Thin Margins for AI Apps: While there is some value in third-party, consumer-facing software built on AI, these AI apps often have lower margins than traditional SaaS products. Amazon, for example, is reportedly scaling back its AI model, Nova, suggesting a recognition that this layer of the industry may not be the primary money-maker.

OpenAI finds itself in the "least valuable layer" as AI and intelligence become commoditized. As more comparable models emerge and prices decrease, paying its massive bills will become increasingly challenging. The cost of a single token has fallen by about 90% since 2023, yet overall AI spending is growing, indicating that the value is shifting to the underlying infrastructure and compute.