Palantir just posted its fastest revenue growth since going public, a wafer-scale chip startup is about to become a $26 billion company, and Anthropic teamed up with Goldman Sachs and Blackstone to build something that looks a lot like McKinsey except it runs on Claude. Big week.
Palantir Grew Revenue 85% Last Quarter. That Is the Fastest It Has Ever Grown as a Public Company.
Palantir reported first-quarter 2026 revenue of $1.633 billion on May 4, an 85% jump year-over-year and the highest growth rate the company has posted since its direct listing in 2020. It beat Wall Street's estimate of $1.54 billion by roughly $90 million, and it was the eleventh consecutive quarter in which the company's revenue growth rate actually accelerated, not just grew.
The American business is the real story. U.S. revenue surged 104% year-over-year to $1.282 billion, crossing the 100% threshold for the first time since the company went public. Net income on a GAAP basis came in at $871 million, a 53% margin. Adjusted free cash flow hit $925 million. The Rule of 40 score, a standard benchmark in enterprise software that adds growth rate to profit margin, reached 145%, a level that CEO Alex Karp said is matched in the tech industry only by NVIDIA, Micron, and SK hynix.
The deal velocity behind those numbers is striking. Palantir closed 206 deals worth at least $1 million each in a single quarter, 72 deals worth $5 million or more, and 47 deals at or above the $10 million threshold. New agreements were signed with GE Aerospace, Airbus, Stellantis, Bain, and Nvidia. Net dollar retention came in at 150%. That is not a company fighting to keep customers. That is a company whose customers keep finding new ways to spend more.
The driver behind all of this is Palantir's Artificial Intelligence Platform, known as AIP. It is not a chatbot or a wrapper on top of someone else's model. AIP is an orchestration layer that connects large language models to a company's live operational data through what Palantir calls the Ontology, a real-time, permissioned, fully auditable map of every decision, asset, and workflow inside an organization. For heavily regulated industries like aerospace, defense, and healthcare, where a generic AI model that cannot explain its outputs or enforce data governance is essentially useless, AIP is filling a gap that no one else has filled at scale.
Management raised full-year 2026 revenue guidance to $7.650 to $7.662 billion, implying 71% growth and landing $380 million above the prior Wall Street consensus. Karp said on CNBC he expects the U.S. business to double again in 2027. The stock has still fallen about 18% this year despite all of this, caught in a broader software selloff. That gap between the fundamentals and the share price is one of the more discussed tensions in markets right now.
Cerebras Is Going Public at $26.6 Billion. It Makes AI Chips the Size of a Dinner Plate and Has a $20 Billion Deal With OpenAI.
Cerebras Systems filed updated IPO terms with the SEC on May 4, setting a price range of $115 to $125 per share for 28 million shares of Class A stock. At the high end of that range, the company would raise $3.5 billion and carry a market cap of $26.6 billion, up roughly 16% from the $23 billion valuation it fetched in its Series H round just three months ago. The stock is set to trade on Nasdaq under the ticker CBRS and is expected to price on May 13. Banks are already fielding around $10 billion in orders for the $3.5 billion on offer, suggesting it will price at or above the announced range.
Cerebras builds the Wafer Scale Engine, a chip fabricated as a single uncut silicon wafer rather than dozens of individual dies packaged together. The result is a chip that is 57 times larger than NVIDIA's H100, with far fewer interconnects between compute cores and dramatically lower latency on inference workloads. The thesis is that for certain AI tasks, particularly high-speed inference on large models, the architecture wins in ways that matter enough to justify the complexity and cost.
Revenue hit $510 million in 2025, up 76% year-over-year. Net income for Q4 2025 alone was $87.9 million, rare for a semiconductor startup at this stage. Remaining performance obligations stand at $24.6 billion. The biggest line item in that backlog is a deal with OpenAI signed in January 2026 committing 750 megawatts of Cerebras compute capacity through 2028, valued at over $20 billion. OpenAI also lent Cerebras $1 billion secured by warrants. Amazon Web Services signed a binding term sheet to deploy Cerebras chips inside its own data centers. AMD participated in the Series H round.
The concentration risk is real. Two customers accounted for 86% of revenue last year. But if you believe inference workloads will continue to scale and that NVIDIA does not have a permanent lock on the market, Cerebras is the most credible public test of that thesis available. If it prices at or above range, it will be the largest tech IPO of 2026 so far.
Read more: https://techcrunch.com/2026/05/04/openais-cozy-partner-cerebras-is-on-track-for-a-blockbuster-ipo/
Anthropic Just Launched a $1.5 Billion Consulting Firm With Goldman Sachs and Blackstone. It Wants to Replace McKinsey for AI Transformation.
Anthropic announced on Monday that it is partnering with Blackstone, Hellman and Friedman, and Goldman Sachs to launch a new AI services company with approximately $1.5 billion in committed capital. Blackstone, Hellman and Friedman, and Anthropic are each investing around $300 million. Goldman Sachs is contributing $150 million. General Atlantic, Leonard Green, Apollo Global Management, Singapore's GIC, and Sequoia Capital are also participating.
The model is not a software licensing business. Instead of selling subscriptions to Claude and letting customers figure out implementation on their own, the new firm will embed Anthropic engineers directly inside mid-sized companies to redesign workflows from the ground up around AI agents. The starting pipeline is the portfolio companies owned by the private equity sponsors backing the deal, which together control hundreds of businesses across healthcare, manufacturing, and financial services. Goldman's Marc Nachmann described the core problem as a talent bottleneck: there are not enough engineers who know how to apply frontier AI tools inside real businesses, and traditional consulting firms are not equipped to fix that gap quickly.
The structure mirrors Palantir's forward-deployment model. Blackstone President Jon Gray said the venture is designed to break down one of the most significant bottlenecks to enterprise AI adoption. The tools being deployed include Claude's core models and Claude Code, the AI coding assistant that has already rattled the legacy software sector this year.
Anthropic was reportedly in talks this week about a new funding round that would value the company at $900 billion, more than doubling its $380 billion February valuation. OpenAI is building a near-identical joint venture structure with TPG and Bain Capital. Both companies are reaching the same conclusion: the real revenue opportunity in enterprise AI is not in the model itself, it is in the implementation layer. The consulting industry generates trillions in annual fees at a ratio of roughly $6 in services for every $1 in software. That is the implicit target.