Subject: OpenAI Cracks at the Top — and Its CFO Didn't Get the Memo
Hey Operators,
Today's AI news is dominated by corporate drama at the top of the industry's power structure. A Wall Street Journal report reveals both OpenAI and Anthropic have been shopping profitability projections to investors — while The Information separately broke the story that CEO Sam Altman has quietly excluded CFO Sarah Friar from key financial meetings as IPO pressure intensifies. On the product front, Google just made a quiet but significant statement by putting Gemma 4 directly on your phone with full on-device privacy, and Microsoft's legal team may have just published the most honest AI product disclaimer in the industry — buried inside its own terms of service.
Operation Check
Tech stocks under moderate pressure Monday as ongoing tariff uncertainty around U.S. trade policy continues weighing on risk assets; semiconductor and device sectors finding partial relief amid ongoing exemption discussions for consumer electronics.
Bitcoin hovering near $92K as the tariff-driven macro environment keeps pressure on risk assets — institutional demand has prevented a deeper sell-off, but sentiment remains cautious.
AI captured $242B — 80% of all Q1 2026 global venture capital in a quarter that shattered every funding record; $300B total poured into startups worldwide, raising real questions about whether revenue justifies the capital outlays.
Operation Dive
OpenAI CEO Cuts His CFO Out of the Room
Sam Altman has reportedly excluded OpenAI CFO Sarah Friar from key financial meetings, including a high-level investor discussion about server procurement, according to The Information. Friar — who has been restructured to report to Fidji Simo rather than directly to Altman — raised internal concerns that a 2026 IPO timeline is premature given the scale of organizational preparation still required and the company's $200B projected cash burn before profitability. The move to sideline the CFO ahead of a landmark public offering raises pointed questions about internal governance at one of the most closely watched companies in tech. The Insight: When a CEO starts excluding his finance chief from the numbers right before an IPO, it usually signals the numbers tell a story someone doesn't want told.
The Race to IPO: OpenAI and Anthropic Both Showed Investors Their Books
The Wall Street Journal reports that both OpenAI and Anthropic have presented profitability projections to potential investors — showing scenarios both with and without model training costs stripped out. The parallel roadshows underscore how both companies are racing toward public markets even while burning cash at extraordinary rates: OpenAI is projected to lose ~$14B in 2026, while Anthropic targets positive free cash flow by 2027. With each company watching the other's IPO timeline closely — fearing the first mover absorbs all pent-up retail demand — the pressure to present a clean financial story is acute. The Insight: Both companies are showing investors the version of the numbers that looks most compelling — the real test will be which story actually holds up at IPO.
Operators in Focus
Google's Gemma 4 Now Runs Fully On-Device — No Cloud Required
Google released the AI Edge Gallery app for Android and iOS today, bringing its Gemma 4 models entirely on-device. The app handles 4,000-token agentic workflows in under 3 seconds, supports image, audio, and text inputs across 140+ languages, and runs all inference locally no internet connection or cloud subscription needed. Developers can build on top of it through a new LiteRT-LM library. The Insight: On-device multimodal AI that can run autonomous agents offline and preserve full user privacy is no longer a research demo — it's now a free app on the App Store.
Microsoft's Own Terms of Service Call Copilot 'Entertainment Purposes Only'
Microsoft's Copilot terms of use explicitly state the product is 'for entertainment purposes only' and warn users not to rely on it for important advice — a striking contradiction for a tool the company has spent billions marketing as an indispensable enterprise co-worker. The clause, which dates to an October 2025 ToS update, went viral across tech media this weekend. Microsoft confirmed it will update the 'legacy language,' but the damage to the enterprise credibility pitch has been done. The Insight: Every AI company has disclaimers like this buried somewhere — Microsoft's just became a news story, and that's a useful reminder to read the ToS before deploying AI in any workflow that actually matters.
Operator's Spotlight Read
Popular American VC Marc Andreessen to Companies Blaming AI for Layoffs: You Are ‘Lying
Marc Andreessen, co-founder of Andreessen Horowitz, has dismissed claims that artificial intelligence is driving mass layoffs. He argues that companies blaming AI are “lying,” pointing instead to cyclical economic factors like post-COVID corrections and interest rate hikes. Andreessen highlights that tech hiring is rebounding, with over 67,000 software engineering openings — the highest in more than three years. He insists AI will boost productivity and create new demand, not eliminate jobs. The Insight: Andreessen’s stance challenges the dominant narrative of AI-driven job loss. By framing layoffs as macroeconomic rather than technological, he positions AI as a growth engine. The tension here is clear: while Andreessen sees elasticity and opportunity, skeptics warn that AI could still compress certain roles, intensify competition, and reshape labor markets in unpredictable ways.
Operator Industry Radar
MaxToki Predicts How Your Cells Age → A new temporal AI model analyzes cellular aging trajectories and suggests molecular interventions to slow age-related disease — a genuine convergence of AI and longevity medicine that goes beyond measuring biological age to recommending action.
China Is Wiring AI Into K-12 Schools — At Scale → ChinaTalk published a detailed look at China's national program integrating AI into primary and secondary education not as a supplemental add-on, but as a government-mandated core curriculum component with real implementation timelines.
Medvi: The Two-Employee AI Startup With $1B+ Revenue — Under the Microscope → Gary Marcus published a sharp analysis of Medvi, the medical AI startup the NYT lauded as a two-person billion-dollar operation, questioning whether the business model and revenue claims hold up under real scrutiny.
Was this email forwarded to you? Don’t miss any updates Subscribe to TechWithAdit for sharp, no - noise tech intelligence stay sharp, – Adit