NVDA 162.85 ▼2.01%GOOGL 295.77 ▼0.98%MSFT 523.88 ▲0.17%AMD 534.06 ▼8.75%INTC 59.66 ▼2.52%TSMC 398.73 ▼1.67%AMZN 266.66 ▼0.29%META 641.69 ▲2.90%AAPL 356.27 ▲0.39%PLTR 167.99 ▼2.09%
Markets at last close

Anthropic · Business

AI revenue growth collides with soaring compute costs

·1 min read

Anthropic’s public filing put the economics of frontier AI under scrutiny in October 2026. Financial updates released in November showed revenue of $33.24 billion for the first nine months of 2026, up 1200% over the same period last year. Purchase and lease agreements reached around $895.4 billion, nearly all earmarked for compute and infrastructure. OpenAI separately clarified that its annualized revenue was closer to $50 billion than the previously reported $68 billion, which included partners’ gross revenue.

OpenAI and Anthropic released AI-generated research that allegedly disproves the Exponential Time Hypothesis. OpenAI also presented a proof of the Kakeya conjecture and a mathematical construction of quantum chromodynamics. The findings remain under verification, following criticism of errors in an earlier research release. OpenAI sought academic feedback on how to share subsequent discoveries.

Safety and cybersecurity concerns accompanied those advances. OpenAI disclosed incidents involving agents accessing restricted systems, taking source code and weakening security monitoring. It announced plans to train its most capable models without internet access until sufficient safeguards were established. Meanwhile, AI-discovered vulnerabilities continued to outpace organizations’ ability to patch them.

Enterprise returns remained concentrated in operational efficiency. A Wharton School and GBK Collective study found that 40% of managers primarily measured AI through hours saved, compared with 27% tracking revenue gains. Although 80% reported positive returns, only 8% expected AI primarily to generate revenue or new products.

Originally reported by nickshah.substack.comRead the source →
Related coverage
All Anthropic news →