Enterprise AI Weekly | EP17 | Anthropic Made $16.8B. Why Are Investors So Nervous?

Anthropic made $16.8 billion last quarter and turned a profit, so why are investors talking about a discount? Paul Brzozowski, Co-Founder of Olakai, the Enterprise AI System of Record, reads the room on enterprise AI this week. Five stops. Anthropic’s investor day numbers and why the run rate went flat over the summer. Accenture putting Claude on a $29 per person per month budget. Google’s Gemini 4 Argon pricing and the date the price war ends. Meta and Apple running Muse Spark Open locally on Mac Studios. And Wall Street, where Goldman Sachs beat and JPMorgan’s CFO said the AI benefit is not yet observable in earnings. Every one of them comes back to the same question: what does each finished piece of AI work cost you, and what did it return? If you own the AI budget, or defend it to a CFO, this one is for you. The math is directional, so check it and tell me where you read the room differently. Chapters 0:00 Why are investors nervous about a $16.8B quarter? 0:30 This week’s round: five stops 1:24 What did Anthropic report at its investor day? 2:08 Why did the run rate go flat over the summer? 2:48 Why did the growth slow? Smarter buyers and a crowded road 4:29 What happens when Accenture puts Claude on a budget? 6:07 When does the price war end? Gemini 4 Argon’s December 1 date 7:51 What does Meta running on a Mac change? 9:07 What did Goldman Sachs and JPMorgan report? 9:29 What did JPMorgan’s CFO say about AI and token cost? 11:30 The one question behind all five stops 11:52 Your AI is an investment. Let’s measure it like one. Key numbers Anthropic revenue: Q1 $4.7B, Q2 $11.6B, Q3 $16.8B, profitable on an adjusted basis after an operating loss of more than $8B in 2025. Run rate: $65B in July against about $67B Q3 annualized, roughly 3% over the summer (our math). $2T is about 30x that revenue, $1.5T about 22x. Accenture: $29 of Claude per person per month centrally funded from October 1, overage charged to a project; heavy users reportedly $500 to $1,000 a month. Gemini 4 Argon: $2 / $10 per million tokens until November 30, $4 / $20 from December 1, the same as Muse Wonder and Claude Opus 5.5. Cached input 95% cheaper. Muse Spark Open: 816B parameters, runs on two or four Mac Studios with M5 Ultra, prompts stay on site. Goldman Sachs: $18.7B revenue against about $16.7B expected, equities up 72%. JPMorgan: first miss on revenue and profit in nearly three years; AI expense about $700M a year. Sources: Reuters via Investing.com, Calcalist, CNBC, and the FT on Anthropic; Reuters exclusive on the valuation discussion; 404 Media (Joseph Cox) via The State of AI on the Accenture memo; TechCrunch; Google’s blog and VentureBeat on Gemini 4 Argon; Axios on Meta and Apple; Fortune on Goldman Sachs and JPMorgan; JPMorgan Chase Q3 2026 earnings call as transcribed by The Motley Fool and Investing.com. Want to see this on your own data, or explore our demo environment? DM me on LinkedIn: https://www.linkedin.com/in/paulbrzozowski/ Learn more about Olakai, the Enterprise AI System of Record: https://olakai.ai Your AI is an investment. Let’s measure it like one. Tags Anthropic, Anthropic revenue, Anthropic IPO, Claude, Accenture, Gemini 4 Argon, Google Gemini pricing, Meta Muse Spark, Apple Mac Studio, Goldman Sachs, JPMorgan, Jeremy Barnum, AI ROI, enterprise AI, AI FinOps, AI cost per task, LLM pricing, token cost, AI budget, Olakai, Enterprise AI System of Record, Enterprise AI Weekly, Paul Brzozowski