The AI industry’s race for profits is now existential

The AI industry's era of limitless capital investment is colliding with the harsh reality of profitability. According to a Decoder podcast analysis by The Verge's Nilay Patel and senior AI reporter Hayden Field, 2026 is a make-or-break year for giants like Anthropic and OpenAI. Facing immense pressure to go public, these companies must now generate more cash than they burn, with leaked Wall Street Journal projections outlining a path to hundreds of billions in revenue by decade's end. The catalyst for this urgent shift is the explosive, and expensive, adoption of AI agents.

Products like Anthropic's Claude Code and OpenAI's Codex are valuable to customers but consume compute resources at an unsustainable rate for standard subscription plans. This has triggered a series of drastic cost-cutting measures that signal a new phase for the industry. OpenAI abruptly killed its Sora video-generation app, walking away from a potential $1 billion licensing deal with Disney, to reallocate precious GPU capacity to its more lucrative Codex agent. Similarly, Anthropic has essentially banned the open-source OpenClaw agent framework from its standard Claude subscription, forcing heavy users onto significantly more expensive pay-as-you-go pricing.

These moves represent the first major glimmers of an industry-wide monetization cliff. After building on hundreds of billions in investor capital and forward-looking infrastructure spend, the largest AI firms must now prove their business models are viable before the bubble risks popping. The compromises being made—canceling flagship products and restricting customer access—highlight the intense financial calculus now dictating product roadmaps. As both companies barrel toward historic IPOs, their every decision is increasingly framed by the race to show profitability, marking a pivotal inflection point from pure research and growth to sustainable business operations.