Anthropic Could Pass OpenAI This Week — And Claude Code Is Why

Anthropic is in talks to raise between $30 billion and $50 billion at a valuation as high as $950 billion. If the round closes anywhere near that number, Anthropic passes OpenAI — currently valued at $852 billion — and becomes the most valuable private company in history. The single number that anchors the story: $950 billion. The single product that is driving the bid: Claude Code.

Why this matters. Twelve months ago Anthropic was a $61.5 billion company. The reported valuation puts the gain at roughly 1,445% in one year. The math isn't speculative: Anthropic's projected annualized recurring revenue is near $45 billion, and corporate paid-seat adoption has, by Fortune's reporting, already passed OpenAI. The Claude Code release this month makes the run defensible — not the model alone, but a developer surface that lets one engineer run many autonomous agents in parallel.

The numbers. Google has pledged up to $40 billion in investment. Amazon up to $25 billion. Anthropic now serves enterprises that buy seats by the thousand — Claude Code is what they buy. The product just leveled up: Agent View landed in Claude Code v2.1.139 as a research preview, putting every active session — local, background, and remote — into a single CLI dashboard. The /goal command lets a developer hand Claude an outcome rather than a step list. The /bg command sends a session into the background while you keep working. Combined with Opus 4.7's native 1 million token context window, a single human can supervise five or ten autonomous coding agents at once.

The backstory. The Claude Code launch in early 2025 was the bet. It looked overpriced at the time — a $200-per-month subscription on top of API consumption. Twelve months later the seats are gone. The Fortune reporting this week names what happened: OpenAI kept its product strategy aimed at consumers and the AGI narrative, and Anthropic quietly built the tool businesses with deep pockets actually use. That gap is what the new valuation is pricing.

The countervailing angle. OpenAI did not let yesterday pass without a counter-punch. On May 14, Codex landed inside the ChatGPT mobile app on both iOS and Android — free across all plans, including Free and Go. The pitch: your phone becomes a live control panel for a Codex agent running on your laptop or remote dev box. Remote SSH is generally available the same day. The trap: the desktop side is macOS only at launch. Windows is coming, but no date. And the cyber lane belongs to Anthropic for now — OpenAI launched Daybreak, a direct response to Anthropic's Mythos cyber model, the same week.

What to watch. Three things over the next ninety days. First: whether the Anthropic round actually closes at $950B — sources have called it "talks," not signed. Second: Codex on Windows. The day the desktop app ships for Windows is the day Codex becomes a default install for enterprise. Third: the customer attrition number on OpenAI's enterprise side — if Fortune's read is right, the seat losses show up at the September renewal cycle.

The line to screenshot: The coding agent war is no longer one product fight. It is two product strategies — and the market just told you which one prints the higher valuation.