Anthropic Captures 73% of Enterprise AI Spending as OpenAI Retreats and SaaS Stocks Shed $2 Trillion

Photo: AI Generated / TheMarketContext.com

Three seismic shifts in artificial intelligence converged in a single week — and one podcast managed to connect all of them. On this week’s All-In Podcast, hosts Chamath Palihapitiya, David Sacks, David Friedberg, and Jason Calacanis mapped out what may be the most consequential stretch in the AI industry’s short commercial history. The through-line: the AI race is no longer about who builds the best model. It is about who captures enterprise budgets, who survives the consumer shakeout, and whose software gets replaced entirely.

Anthropic’s Enterprise Takeover

Anthropic is on what the hosts described as a generational run. The company’s annualized revenue has surged past $19 billion as of early March 2026, up from $1 billion in December 2024 — a pace of roughly 10x annual growth sustained for three consecutive years. In February, Anthropic closed a $30 billion Series G round at a $380 billion valuation.

David Sacks noted on the episode that Anthropic’s strategic bet on coding and developer tools has paid off massively, penetrating enterprise IT budgets in a way that consumer-facing AI products have not. Claude Code, launched in May 2025, reached $2.5 billion in annualized revenue within nine months. Business subscriptions to the tool have quadrupled since January 1, 2026, and 4% of all public GitHub commits are now authored by Claude Code — projected to exceed 20% by year-end.

Chamath Palihapitiya drew a sharp distinction between the two AI leaders on the episode: OpenAI dominates consumer subscriptions while Anthropic dominates the API and enterprise market. The data backs him up. Ramp’s March 2026 AI Index shows Anthropic now captures 73% of all first-time enterprise AI spending — a dramatic swing from the 60/40 split that favored OpenAI as recently as early December. Enterprise customers account for roughly 80% of Anthropic’s revenue, and more than 500 customers now spend over $1 million annually. The monetization gap is striking: Anthropic generates approximately $211 per monthly user compared to OpenAI’s roughly $25 per weekly user.

OpenAI Retreats From Consumer Side Quests

Then came the Sora shutdown. On March 24, OpenAI pulled its AI video generation app, the API, and sora.com entirely. The move dissolved Disney’s $1 billion investment and three-year licensing deal — with Disney reportedly learning the night before the public announcement.

The hosts analyzed the collapse in detail. Sora downloads had fallen roughly 75% from their November 2025 peak of 3.3 million to just 1.1 million in February. Lifetime revenue from in-app purchases totaled approximately $2.1 million, per Appfigures estimates. As Sacks observed on the podcast, the shutdown signals OpenAI’s broader pivot away from consumer experiments and toward enterprise productivity as the company prepares for its IPO.

OpenAI has been operating under what insiders describe as a code red since December. CEO of Applications Fidji Simo cut what she called side quests at a recent all-hands meeting. The company also canceled its Instant Checkout shopping feature and is now consolidating its browser, ChatGPT app, and Codex into a single desktop super app. OpenAI’s annualized revenue stands at roughly $25 billion compared to Anthropic’s $19 billion — but Anthropic is accelerating faster, per Axios reporting. NBC News reported that OpenAI has come under intense pressure from its rival.

The SaaS Reckoning

The broader fallout extends well beyond these two companies. The All-In hosts spent significant time on what has become a market-wide repricing of software stocks. Public SaaS companies have shed roughly $2 trillion in market capitalization since the start of 2026 as investors price in the threat of AI agents replacing traditional software workflows.

Chamath described the dynamic as strangulation as a service — the idea that enterprise clients no longer want complex dashboards and per-seat licensing models. They want AI agents that work invisibly in the background. He shared that his own team replaced an entire HR system and rebuilt a company website over a single weekend using AI tools. Friedberg offered his own example: he built a Chrome extension in just days for a domain he had owned for 15 years, calling the process vibe coding.

Companies like Salesforce, ServiceNow, and Workday have faced immediate valuation pressure. The iShares Expanded Tech-Software ETF has plunged nearly 21% year-to-date through late March. As Jason Calacanis noted on the episode, mega-cap tech incumbents — Apple, Meta, Alphabet, Nvidia — continue to soar on their infrastructure moats while the software layer beneath them gets repriced.

From the Podcast to the West Wing

In a twist that underscores how intertwined Silicon Valley and Washington have become on AI policy, two of the four All-In hosts now formally advise the White House. David Sacks confirmed on the episode that his 130-day stint as White House AI and Crypto Czar has ended, and he now co-chairs the President’s Council of Advisors on Science and Technology (PCAST) alongside Michael Kratsios. Friedberg was also appointed to the council.

PCAST — for readers unfamiliar — is a presidential advisory body that makes science and technology policy recommendations directly to the White House. It does not set policy, but its influence on executive priorities is significant. The current roster reads like a who’s who of American tech: Jensen Huang, Mark Zuckerberg, Larry Ellison, Sergey Brin, Lisa Su, Marc Andreessen, and Michael Dell, among others. Sacks called it the most star-studded advisory group of its kind ever assembled. The council will focus on AI, advanced semiconductors, quantum computing, and nuclear power.

Friedberg noted on the episode that China has overtaken the United States in peer-reviewed scientific research output — a trend that makes having industrial tech leaders, rather than purely academic appointees, shaping national science policy especially urgent. Only one of the 13 initial PCAST members is an academic: John Martinis of UC Santa Barbara.

What to Watch

Federal Reserve / Economic Calendar: The March jobs report drops Friday, April 3. February’s weak print of just 92,000 nonfarm payrolls and a 4.4% unemployment rate rattled markets. Another soft reading could shift rate-cut expectations for the second half of 2026.

Earnings: Nike reports Q3 fiscal 2026 results on Tuesday, March 31, after market close. Analysts expect $0.29 EPS on $11.2 billion in revenue — both sharply lower year-over-year — as the company navigates a $1.5 billion tariff headwind and a struggling Converse subsidiary. The stock is down roughly 60% over five years.

Broader Market: Oil prices have surged past $110 per barrel on Brent crude, pushing import prices up 1.3% month-over-month — the largest monthly jump since March 2022. Futures traders now price in a 52% probability of a rate hike by year-end, crossing the 50% threshold for the first time this cycle. The oil-inflation dynamic is complicating the Fed’s path and creating stagflationary pressure across equity markets.