Version 5

Anthropic overtook OpenAI in U.S. business AI spend — 34.4% vs. 32.3% per Ramp's May 2026 AI Index [1] — driven by Claude Code reaching $2.5B in annualized revenue [1]. In mid-May, Anthropic acquired SDK and MCP tooling firm Stainless (~$300M) [2][3] and confirmed Andrej Karpathy to lead Claude pretraining research [10][11][12]. OpenAI formally launched DeployCo on May 11 — a majority-owned standalone business backed by $4B+ from 19 global investors led by TPG, acquiring Tomoro's ~150 Forward Deployed Engineers, and enlisting McKinsey, Bain & Company, and Capgemini as partners [14]. Google's A2A protocol has crossed 150 adopting organizations in enterprise production use [19]. Unverified social posts now claim Anthropic's share has surged further to ~73% [25], though no formal index has confirmed this.

Enterprise AI contracts embed into procurement cycles and compliance workflows, creating compounding switching costs for any vendor that lands first [1]. OpenAI's DeployCo — with TPG backing, a named FDE acquisition, and major consultancy partnerships — upgrades the competition from a model capability contest to a battle for enterprise relationship depth. The Stainless acquisition has simultaneously generated a new 'AI SDK supply chain risk' discourse, signaling that developer toolchain dependencies are now recognized as competitive vulnerabilities across the industry [7].

Through most of 2025, OpenAI held a commanding lead in enterprise AI adoption — Anthropic's share of U.S. business AI spend sat at roughly 8% on Ramp's corporate credit card index in May 2025 [1]. By May 2026, that figure had jumped to 34.4%, crossing OpenAI — now at 32.3% — for the first time. The primary engine of that reversal was Claude Code, Anthropic's autonomous coding agent, which accumulated $2.5 billion in annualized revenue as early as February 2026 [1]. Anthropic's total annualized revenue reached $30 billion by April 2026, surpassing OpenAI's approximately $24 billion [1]. The structural reason this shift may prove durable is enterprise contract stickiness: unlike consumer subscriptions, enterprise AI deployments embed into procurement budgets, internal workflows, and compliance approvals, making switching require a new procurement cycle rather than deleting an app [1].

On May 18–19, Anthropic made two high-signal moves in rapid succession. First, it acquired Stainless — the company behind all official Anthropic SDKs since the Claude API launched, and the SDK and MCP tooling provider for OpenAI, Google, Meta, Stripe, and hundreds of other companies — for a reported approximately $300 million [2][3][4]. The competitive framing has hardened since initial reports, with multiple outlets describing Anthropic 'cutting off' competitors' SDK access and 'forcing OpenAI and Google to rebuild' [5][6]. In parallel, a distinct 'AI SDK supply chain risk' analytical lens has emerged in the developer community, with advisors flagging that enterprises and AI providers now have infrastructure dependencies on a competitor-owned toolchain [7]. Community reaction has been sharp — characterizing it as 'the most savage acquisition of 2026' [8] and 'a move most people ignored' despite its infrastructure significance [9]. Whether access will actually be restricted remains unconfirmed. Second, multiple major outlets confirmed that Andrej Karpathy — OpenAI co-founder and former head of Tesla's AI team — has joined Anthropic to lead Claude pretraining research and form a new Claude-focused research team [10][11][12]. The combination of the Stainless acquisition and the Karpathy hire is widely read as coordinated positioning at both the infrastructure and foundational model layers [13].

OpenAI's enterprise response has taken two concurrent forms. On May 11, it formally launched the OpenAI Deployment Company (DeployCo), a majority-owned standalone business backed by more than $4 billion from 19 global investors and consultancies led by TPG [14]. DeployCo is simultaneously acquiring Tomoro, bringing approximately 150 Forward Deployed Engineers at launch, and has enlisted McKinsey, Bain & Company, and Capgemini as partners — giving it claimed reach across thousands of businesses worldwide [14]. DeployCo FDEs are designed to work inside client organizations to design, build, and deploy production AI systems connected to customers' data and business processes [14]. This makes the 'Palantir-style consulting arm' framing concrete: DeployCo is a formally capitalized, separately structured entity with acquired human talent, major consulting partnerships, and its own investment base. In parallel, on May 18, OpenAI announced a partnership with Dell to bring Codex into hybrid and on-premises enterprise environments via the Dell AI Factory [15][16][17], connecting the coding agent to internal codebases and workflows that enterprises are unwilling to expose to cloud APIs — a deployment model analysts characterize as a defensive moat for large regulated enterprises [18].

Google has emerged as a visible third competitor with a structurally different strategy. Its A2A (Agent-to-Agent) protocol has crossed 150 adopting organizations and reached enterprise production use within its first year [19][20], and has attracted attention as a potential disruptor of major ERP ecosystems including SAP [21]. The broader agent protocol landscape now features competing standards — A2A, MCP, ACP, and ANP — with active developer-community analysis of their respective trade-offs [22][23][24]. The Stainless acquisition directly affects Google as a former Stainless client, adding a new dimension to its competitive calculus with Anthropic [5][6]. The three companies' enterprise strategies diverge sharply: Anthropic is acquiring the developer toolchain and pre-training talent; OpenAI is building a formally capitalized consulting subsidiary and partnering with on-premises hardware incumbents; Google is betting on open interoperability protocols and existing cloud relationships.