A $300M Bet on the SDK Layer
Anthropic has acquired Stainless, a New York-based SDK startup founded by former Stripe engineer Alex Rattray, in a deal reported to be worth more than USD 300 million (approx. RM1,380 million). Stainless quietly powered a large share of the AI industry’s API layer, automatically generating and maintaining SDKs across languages such as Python, TypeScript, Go, Java, and Kotlin. Its technology underpins official clients for Anthropic’s Claude API libraries, and has also been used by OpenAI, Google, Cloudflare, Runway and others to ship polished developer experiences without building in-house tooling. Anthropic confirmed that all hosted Stainless products will be wound down and kept for internal use only, meaning existing customers keep rights to SDKs already generated but lose access to future hosted services. What looks like a routine AI SDK developer tools deal is in fact a decisive move to seize control of critical connective tissue between models and developers.
From Shared Infrastructure to Competitive Moat
Stainless’s core product turned an API specification into production-ready SDKs, CLIs and MCP servers, the connectors that let AI agents call external APIs through a standard interface. For many AI providers, Stainless acted like a contract manufacturer for developer experience: feed it an OpenAPI spec and it produced idiomatic clients and kept them in sync as the API evolved. By bringing this shared “factory” in-house and shutting it to outsiders, Anthropic is not just acquiring technology; it is removing a neutral supplier that OpenAI, Google, Cloudflare and others depended on. Competitors must now either rebuild generation pipelines, migrate to alternative tools, or maintain SDKs manually as their APIs change. None of those paths is catastrophic, but all introduce friction in a layer that had been invisible and reliable, shifting the balance of power in AI infrastructure competition.
The New Battleground: Toolchains, Not Just Models
Anthropic’s Stainless acquisition lands amid a broader shift in AI strategy: frontier labs increasingly see developer tooling and execution layers as durable moats, even as large models themselves trend toward commoditisation. Anthropic has already bought Bun, the JavaScript runtime underpinning Claude Code, and Vercept, which focuses on AI-mediated computer usage. OpenAI, meanwhile, is acquiring Python tooling company Astral. Together, these moves suggest that owning runtimes, package managers, SDK generators and agent connectivity infrastructure is becoming as important as model performance. SDKs are especially “sticky”: whoever ships the cleanest, most up-to-date library tends to win long-tail developer mindshare. By consolidating Stainless inside its own stack and sunsetting it as a shared platform, Anthropic signals that the real contest is over who controls the AI developer toolchain—and, by extension, the enterprise AI workflows and agents that will be built on top of it.
Implications for Developers Using OpenAI and Google Tools
For developers, the immediate impact is subtle but significant. OpenAI’s Python, Node, Java, Go and Ruby clients, as well as many Google and Cloudflare SDKs, were generated by Stainless. Those existing libraries will continue to function, and customers retain full rights to the code they already have. The real pain point comes as APIs evolve. Without Stainless’s hosted generator, rival labs must allocate new engineering capacity to keep multi-language SDKs and MCP connectors updated, or switch to alternative generators and absorb migration costs. That overhead could slow the rollout of new features, introduce fragmentation across language clients, and widen experience gaps between ecosystems. Over time, if Anthropic can iterate faster on Claude API libraries and MCP-based agent integrations, its tooling may feel more cohesive and modern, subtly pulling developers—and their enterprise workloads—toward Anthropic’s corner of the AI infrastructure stack.