Anthropic Acquisitions
M&A Intelligence · Company Acquisitions
Anthropic's acquisition strategy is the most disciplined in enterprise AI — six deals in six months, each targeting a single layer of the Claude product stack. Where OpenAI buys breadth, Anthropic buys depth: developer infrastructure, computer-use agents, life-sciences AI, and enterprise implementation. This page tracks every confirmed acquisition, the strategic logic behind each, and the deal that got away.
 |
| Image credit: Wikimedia Commons |
Anthropic made zero acquisitions in its first three years. Between its founding in 2021 and December 2025, the company built everything internally — Claude models, the Claude Code product, the Model Context Protocol, and the enterprise sales motion. That changed in December 2025, and the six deals that followed reveal a strategy that is deliberately different from OpenAI's.
Where OpenAI has completed 15 acquisitions since 2023 across hardware, consumer media, and enterprise services, Anthropic has focused on a narrow set of infrastructure layers: the JavaScript runtime that powers Claude Code, the SDK generation layer that connects agents to APIs, the computer-use agents that let Claude operate software, and the life-sciences platform that opens a new vertical. Each deal maps to a specific technical layer of the Claude stack, and none of them were about acquiring revenue. They were about acquiring capability.
Ranked Acquisitions
1. Coefficient Bio
Deal value: ~$400M (all-stock) · Year: 2026 · Status: Closed
2. Stainless
Deal value: ~$300M+ · Year: 2026 · Status: Closed
3. Bun
Deal value: Undisclosed (reportedly several hundred million) · Year: 2025 · Status: Closed
4. Vercept
Deal value: Undisclosed · Year: 2026 · Status: Closed
5. Fractional AI (via Ode)
Deal value: Undisclosed · Year: 2026 · Status: Closed
6. Casper Studios (via Ode)
Deal value: Undisclosed · Year: 2026 · Status: Closed
The CODEW Lens: Rankings are based on disclosed or reported deal value. The strategic rationale matters more than the ranking — a reportedly ~$300M deal like Stainless may matter more than the ~$400M Coefficient Bio deal if it removes a shared dependency from every competitor's workflow.
Coefficient Bio
Anthropic's acquisition of Coefficient Bio, a stealth biotech AI startup founded barely eight months earlier, is the clearest example of the company's vertical strategy. The all-stock deal, valued at just over $400 million, brought a team of fewer than 10 people — a price per head that works out to more than $40 million each. Coefficient Bio develops AI models for biological research, including AI-driven solutions for biopharmaceutical intelligence and drug development. The acquisition came two days after OpenAI announced its own acquisition of tech talk show TBPN, but the two deals pointed in opposite directions: OpenAI was buying narrative influence, while Anthropic was buying scientific capability.
The deal made sense as the third step in a sequence Anthropic had been building for months. In October 2025, it launched Claude for Life Sciences, connecting the model to research databases like PubMed and UniProt. In January 2026, at the JPMorgan Healthcare Conference, it launched Claude for Healthcare with HIPAA compliance. Coefficient Bio was the capstone — giving Claude the ability to support drug discovery pipelines, not just literature review. The acquisition also came during Anthropic's Series G fundraise, which valued the company at $380 billion post-money.
The CODEW Lens: Coefficient Bio was not a biotech acquisition. It was a vertical entry strategy. Anthropic spent six months building the healthcare narrative — research access, compliance certification, then the drug discovery capability — before making the deal. The acquisition was the last move, not the first.
Stainless
Anthropic's acquisition of Stainless, a New York-based developer tools startup founded by former Stripe engineer Alex Rattray, was the most strategically aggressive deal in the company's history. Stainless automated the creation and maintenance of software development kits — the libraries developers use to interact with APIs — and its software was used by OpenAI, Google, Cloudflare, Replicate, Runway, Meta, Groq, and Cerebras. For three years, Stainless had generated every official Anthropic SDK since the earliest days of the Claude API. It had also done the same for OpenAI. The acquisition reportedly cost Anthropic more than $300 million — roughly twice Stainless's last valuation.
The strategic logic was a supply cut-off dressed as an acquisition. Anthropic told existing Stainless customers they would keep the SDKs they had already generated and could modify and extend them, but the company would wind down all hosted products, including the SDK generator. Going forward, the tool that had quietly powered OpenAI's and Google's developer libraries would only be available to Anthropic. Combined with Anthropic's Model Context Protocol — the open standard for how agents connect to external systems — Stainless completed a three-piece set: Claude as the model, Stainless as the interface, and MCP as the connection.
The CODEW Lens: Stainless was not a developer tools acquisition. It was an infrastructure denial play. Anthropic bought the layer that sits between models and developers — and shut it down for everyone else. OpenAI's own Python SDK README still says "Generated by Stainless based on the OpenAPI specification." That is what Anthropic now owns.
Bun
Anthropic's first acquisition — announced in December 2025 — was Bun, the high-speed JavaScript runtime founded by Jarred Sumner in 2021. Bun was already deeply embedded in the Claude Code infrastructure, and Anthropic disclosed the deal alongside a milestone: Claude Code had reached $1 billion in run-rate revenue just six months after becoming publicly available. Bun gets more than 7 million monthly downloads, has earned over 82,000 stars on GitHub, and has been adopted by companies including Netflix, Spotify, KPMG, L'Oréal, Salesforce, Midjourney, and Lovable. The deal price was not disclosed, though reports suggested it was in the several-hundred-million range.
The strategic rationale was infrastructure scale. Claude Code was growing faster than Anthropic's tooling could support, and Bun gave the company a runtime that was dramatically faster than the leading competition — an all-in-one toolkit combining runtime, package manager, bundler, and test runner. Anthropic committed to keeping Bun open source and MIT-licensed, and said it would continue investing in making it the best JavaScript runtime for all developers. For Claude Code users, the acquisition meant faster performance, improved stability, and new capabilities.
The CODEW Lens: Bun was an infrastructure acquisition disguised as a developer tools deal. Anthropic was not buying a runtime — it was buying the performance layer that makes Claude Code viable at enterprise scale. The $1B run-rate milestone and the acquisition announcement were not coincidental.
Vercept
Anthropic's acquisition of Vercept, a Seattle-based AI computer interface company, was designed to strengthen Claude's computer-use capabilities — the ability to navigate apps, click, type, and complete multi-step tasks on a user's behalf. Vercept was founded in 2024 by former Allen Institute for AI researchers including Oren Etzioni, who served as AI2's founding CEO, and had raised more than $50 million from backers including former Google CEO Eric Schmidt, Google DeepMind chief scientist Jeff Dean, and Cruise founder Kyle Vogt. Its flagship product was Vy, a cross-platform AI agent that enabled users to control their computers with natural language.
The acquisition came just after Anthropic unveiled Claude Sonnet 4.6, its best model yet for computer usage, and the deal was widely seen as a way to accelerate that capability. Not everyone was pleased: Etzioni himself posted on LinkedIn that Vercept was "throwing in the towel and giving their customers 30 days to get off the platform." Vy was scheduled to shut down on March 25. Founding member Matt Deitke had already left Vercept to join Meta the previous summer for a package reportedly worth $250 million over four years. The acquisition was as much about consolidating scarce frontier-agent research talent as it was about technology.
The CODEW Lens: Vercept was a talent acquisition dressed as a product acquisition. Anthropic wanted the researchers who had spent two years building computer-use agents, not the Vy product. The shutdown of Vy within weeks of the deal confirmed it — the team was the asset.
Fractional AI (via Ode)
Anthropic's enterprise strategy took a different structural form in May 2026, when Ode — a $1.5 billion joint venture Anthropic created with Blackstone and Hellman & Friedman — acquired Fractional AI, an applied-AI implementation firm based in San Francisco. Anthropic, Blackstone, and Hellman & Friedman each invested $300 million into the venture, with Goldman Sachs contributing $150 million as a founding investor. The deal was Ode's first acquisition and marked a shift in how Anthropic approaches enterprise services: rather than building a consultancy inside the company, it created a separate entity with private equity partners and used it to buy implementation firms.
Fractional AI was founded in 2024 by three former LiveRamp executives and focused on moving generative AI from pilot to production deployment inside enterprise workflows. It had previously been an eleven-month partner with OpenAI, but that relationship ended when the Ode deal closed. Ode CEO Chris Taylor, who came from Fractional, said at launch that it was "pretty easy to imagine this as a trillion-dollar company someday if we execute well." The acquisition gave Ode a ready-made team of senior engineers — more than half of whom had previously founded startups — and a delivery model built around sending small teams into large companies to identify where AI can help and then build it.
The CODEW Lens: Fractional AI was not an Anthropic acquisition. It was an Ode acquisition. The distinction matters because it shows how Anthropic is separating the model business from the implementation business — using private equity capital to fund enterprise services while keeping the core AI company focused on research and platform.
Casper Studios (via Ode)
Ode's second acquisition, announced in August 2026, was Casper Studios, an AI consultancy that helped companies put Claude to work inside the software their staff already used. The announcement gave no price. Ode framed the two businesses as different ends of the same job: Ode builds custom systems for the hardest problems, while Casper does the volume work — the skills, connectors, and context inside Anthropic products that spread AI across an organisation. Casper CEO Jay Singh said the two firms had "found common cause," and Ode CEO Chris Taylor noted that the teams had already worked together on several engagements before the deal.
Ode highlighted one shared client, Sphera, as evidence of the strategic fit. Ode built a custom internal tool for part of Sphera's business that it says cut bottlenecks from time-intensive operational tasks by 70%. Casper separately automated processes across Sphera's customer support, consulting, and project planning teams. The detail Ode chose to emphasize was interoperability: both firms built in the same Claude Code setup, so tools written by one team could be picked up by the other immediately. Two consultancies that share a toolchain are worth more than two that do not.
The CODEW Lens: Casper was a scale acquisition for Ode, not a capability acquisition. Fractional AI gave Ode the engineers to build hard things; Casper gives it the volume capacity to spread Claude across entire organisations. The two deals together describe a full enterprise services stack — custom builds on one end, broad adoption on the other.
The Deal That Got Away: Decart
In August 2026, reports emerged that Anthropic was in talks to acquire Decart AI, an Israeli AI startup, for approximately $6 billion — which would have been the largest acquisition in Anthropic's history and its first in Israel. The price represented a premium of roughly 50% over Decart's last funding round valuation of $4 billion, and the deal was widely seen as a strategic move to establish a development centre in Israel and to shore up Anthropic's position in video generation and world models. Decart had raised about $450 million since its inception, including $300 million in May 2026 at a $4 billion valuation, with backing from Nvidia.
By September 2026, the deal had collapsed. Anthropic completed due diligence and walked away, according to Bloomberg. The company's largest completed acquisition remains the approximately $300 million it paid for Stainless. The episode is a reminder that acquisition talks at the frontier of AI often fail — and that the gap between a reported deal and a signed deal can be vast. Anthropic and Decart reportedly still may partner in other ways.
The CODEW Lens: The Decart deal would have been 20 times larger than any acquisition Anthropic has actually closed. Its collapse is a reminder that the company's acquisition strategy is disciplined to the point of restraint — it would rather walk away than overpay for a capability it can build or partner for.
Industry Patterns
Anthropic's acquisition strategy is defined by layer discipline. Each deal targets a specific part of the Claude product stack: Bun powers Claude Code's runtime, Stainless connects agents to APIs, Vercept gives Claude the ability to operate software, Coefficient Bio opens life sciences, and Ode's consultancies deploy Claude into enterprise workflows. No deal is random. No deal is about acquiring revenue. Every deal is about acquiring a capability that Anthropic cannot build fast enough on its own.
A second pattern is the contrast with OpenAI. OpenAI has made 15 acquisitions since 2023 across hardware, consumer media, developer tools, healthcare, security, and enterprise services. Anthropic has made six, all in the last six months, all infrastructure-focused. The two strategies reflect different anxieties: OpenAI is building a consumer-facing empire that depends on public perception, while Anthropic is building an enterprise-focused platform that depends on technical moats and developer loyalty. OpenAI buys narrative. Anthropic buys barriers.
A third pattern is the use of joint ventures to fund enterprise expansion. Ode — the $1.5 billion company Anthropic created with Blackstone and Hellman & Friedman — allows Anthropic to build an enterprise services business without spending its own capital or distracting its core team. Ode's acquisitions of Fractional AI and Casper Studios are Anthropic acquisitions only in the loosest sense; they are Ode acquisitions funded by Ode capital, serving Ode customers, and operating under Ode's brand.
The CODEW Lens: Anthropic's pattern is layer-specific acquisition. The company does not buy businesses — it buys layers of the Claude stack. Bun is the runtime layer. Stainless is the API interface layer. Vercept is the computer-use layer. Coefficient Bio is the life-sciences layer. Ode is the enterprise implementation layer. The pattern is the strategy.
Notable Honorable Mentions
Anthropic's acquisition history is deliberately short, but a few additional details matter for context. The company's Series G fundraise — $30 billion at a $380 billion post-money valuation, announced in February 2026 — gave it the balance sheet to make larger deals if it chose to. The Bun acquisition came with a commitment to keep the runtime open source and MIT-licensed, a decision that preserves developer goodwill while giving Anthropic internal control over the toolchain. And the failed Decart deal — reportedly valued at $6 billion — remains the clearest data point for how much Anthropic is willing to pay for an acquisition it actually wants.
The company has also invested in or partnered with several companies that it has not acquired, including its deep commercial ties with Amazon (which has committed up to $8 billion in funding and infrastructure) and its partnership with Nvidia and Microsoft (which invested $15 billion in November 2025). These relationships are not acquisitions, but they are central to Anthropic's infrastructure and go-to-market strategy.
The CODEW Lens: The investments Anthropic did not make are as revealing as the ones it did. Its deepest relationships — Amazon, Nvidia, Microsoft — are partnerships and infrastructure commitments, not acquisitions. Anthropic's strategy is to own the application and developer layers while renting the compute and cloud layers from partners.
FAQ
How many companies has Anthropic acquired?
Anthropic has completed six acquisitions: Bun (December 2025), Vercept (February 2026), Coefficient Bio (April 2026), Stainless (May 2026), Fractional AI (via Ode, May 2026), and Casper Studios (via Ode, August 2026). The company made no acquisitions before December 2025.
Why does Anthropic acquire so much less than OpenAI?
Anthropic's strategy is deliberately narrower. It targets specific infrastructure layers of the Claude stack rather than expanding into consumer media, hardware, or services. Its ~80% enterprise revenue mix means it optimizes for developer trust, technical moats, and enterprise credibility — not public narrative.
What was Anthropic's largest acquisition?
By disclosed value, Coefficient Bio at approximately $400 million. Stainless was reportedly over $300 million. Bun's price was undisclosed but reportedly in the several-hundred-million range. The failed Decart deal would have been $6 billion — roughly 20 times larger than any completed deal.
What makes Anthropic acquisitions different?
Layer specificity. Anthropic does not buy businesses — it buys technical layers of the Claude stack. Each deal maps to a specific layer: runtime, API interface, computer-use agents, life sciences, or enterprise implementation. The deals are small, precise, and infrastructure-focused. There is no portfolio diversification here.
Related Reading
How Tech Acquisitions Work — The full process, step by step.
How Tech Company Valuations Work — Revenue multiples, comps, DCF, and strategic premium.
Glossary of M&A Terms — Common valuation, deal structure, and diligence terminology.
AI Acquisition Trends — What drives Big Tech's AI dealmaking.
Big Tech Acquisitions Hub — The central directory for Big Tech M&A coverage.
Tech M&A Database — Track real deals as they happen.
The CODEW Stat
6 acquisitions · ~$400M largest · 6 months Anthropic completed six acquisitions in six months after three years of zero. Its largest disclosed deal is Coefficient Bio at ~$400M. The failed Decart deal would have been $6B — 20 times larger. The numbers tell you the size. The strategy tells you why each one mattered — runtime infrastructure, API interfaces, computer-use agents, life-sciences verticals, and enterprise implementation. Anthropic does not collect businesses. It buys layers.
Reviewed by Erwin Castro
on
Sunday, July 05, 2026
Rating:
No comments: