Enterprise AI spending narrows
Multiple reports say enterprise AI budgets are consolidating around a smaller set of vendors, with Anthropic winning a disproportionate share of new enterprise spend while IT services firms treat AI as core revenue. That pattern — buyers choosing integration fit and workflow utility over headline model fame — is visible in firm-level disclosures and market analyses. (businessinsider.com) (economictimes.indiatimes.com) (investing.com)
A year ago, companies were spreading artificial intelligence budgets across dozens of pilots. In early 2026, the money is bunching up around a smaller group of vendors, and Andreessen Horowitz said the market is moving toward an “emerging oligopoly” instead of a wide-open field. (a16z.com) One winner keeps showing up in those checks: Anthropic. Andreessen Horowitz wrote in January that Anthropic had become the leader in enterprise large language model market share, ahead of OpenAI, as businesses picked tools they could plug into daily work. (a16z.com) Anthropic’s own numbers point the same way. In its February funding announcement, the company said Claude Code run-rate revenue had passed $2.5 billion, business subscriptions had quadrupled since the start of 2026, and more than half of Claude Code revenue now came from enterprise use. (anthropic.com) The company also said more than 500 customers now spend at least $1 million a year with Anthropic, and eight of the Fortune 10 are Claude customers. That is not consumer app traffic; that is procurement departments signing large recurring contracts. (anthropic.com) This is why Michael Burry’s swipe at Palantir landed. Business Insider reported on April 8 that Burry argued Anthropic was “eating Palantir’s lunch,” because buyers were favoring software that drops into existing workflows over platforms that still need heavier custom setup. (businessinsider.com) Palantir is not small or collapsing. Palantir’s 2025 annual report said it generated $4.5 billion in revenue, with 54 percent from government customers and 46 percent from commercial customers, which shows why investors are watching whether private-sector artificial intelligence budgets shift toward model vendors and away from broader data platforms. (investors.palantir.com) The same narrowing is visible one layer down, in the companies paid to wire artificial intelligence into old corporate systems. Tata Consultancy Services said on April 9 that its annualized artificial intelligence revenue crossed $2.3 billion in the quarter ended March 31, 2026, while total contract value for the quarter reached $12 billion. (tcs.com) That tells you where the spending is going. Big companies are still paying consultants, but they are paying them to roll out a smaller set of approved tools across call centers, software teams, cloud systems, and back-office operations instead of funding endless experiments. (economictimes.indiatimes.com) The market is still getting bigger at the same time it gets narrower. Gartner said worldwide artificial intelligence spending will hit $2.52 trillion in 2026, up 44 percent from 2025, so consolidation does not mean less money; it means more of the money is flowing to vendors that clear security reviews, fit procurement rules, and solve one job reliably. (gartner.com) That is why headline fame is starting to matter less than boring details like connectors, compliance, and whether a tool works inside Microsoft software, Amazon Web Services, or a bank’s internal codebase. In enterprise artificial intelligence, the winner is increasingly the company that behaves less like a chatbot demo and more like plumbing. (a16z.com)
