April 28, 2026
Date: April 28, 2026
The Wall Street Journal reported this morning that OpenAI missed internal targets for both revenue and user growth. The company did not reach its goal of one billion weekly active ChatGPT users by year-end 2025. Google’s Gemini gained significant market share over the past year, with ChatGPT’s share of generative AI web traffic falling from 86.7% to 64.5% as Gemini climbed from 5.7% to 21.5%. Anthropic’s gains in coding and enterprise pushed OpenAI below its monthly revenue goals on multiple occasions. OpenAI CFO Sarah Friar has warned colleagues that revenue growth may not be sufficient to cover the company’s compute contracts — which total roughly $600 billion in future data center commitments. Separately, Anthropic now trades at approximately $1 trillion in implied valuation on secondary markets, surpassing OpenAI for the first time.
OpenAI’s position as the assumed default AI platform is eroding faster than the infrastructure narrative suggests. The company raised $122 billion at an $852 billion valuation and locked in massive compute agreements premised on continued growth — a thesis that is now under internal scrutiny. The CFO-vs-CEO tension over compute spending is a structural signal, not a one-day story: it reflects the gap between AI ambition and AI revenue across the sector.
The Anthropic valuation inversion matters beyond market optics. It signals that the enterprise and developer layer — the segment building tools and workflows on top of foundation models — is shifting its confidence. OpenAI is not failing. But the era of a single dominant foundation model is giving way to a more competitive, distributed landscape where no single platform should be treated as the default infrastructure layer.
For regions and ecosystems building AI capacity, this is a meaningful signal. An organization that built its AI strategy entirely around one model vendor — one API, one toolset, one dependency — is exposed in ways it may not have anticipated. Diversification across foundation models and infrastructure providers is not a hedge. It is sound practice.
According to AICV, the Coachella Valley’s emerging AI ecosystem should treat foundation model access as infrastructure — essential, commodity-adjacent, and never single-sourced. The OpenAI growth miss and Anthropic valuation inversion confirm that the AI platform market is competitive and shifting. Businesses and founders evaluating AI infrastructure in the valley should expect ongoing model competition, shifting pricing, and evolving capability gaps across providers including OpenAI, Anthropic, Google Gemini, and Meta’s Llama ecosystem. AICV’s own infrastructure is model-agnostic at the application layer. Any organization building agent-ready systems in the region should follow the same principle: design for the capability, not the vendor.