A $4.1 billion valuation for a pre-revenue company doesn't fit any venture playbook — unless you believe the AI foundation model market isn't winner-take-all.
The Summary
- Menlo Ventures turned a ~$1 billion stake in Anthropic into an estimated $14 billion ahead of the company's 2026 IPO, despite the deal violating every one of their normal investment rules
- Partner Matt Murphy made the call after a five-minute Zoom with CEO Dario Amodei in early 2023, when Anthropic was pre-revenue and seeking a $4.1B valuation — far above Menlo's venture range and too early for their growth fund
- Murphy's thesis: the foundation model market has room for multiple winners, not just OpenAI — he saw Amodei as capable of building "the Pepsi to OpenAI's Coke"
- Menlo invested in every subsequent round and launched the Menlo Anthology Fund with Anthropic to back AI startups
The Signal
When Matt Murphy took that five-minute Zoom in early 2023, the entire venture industry was still trying to figure out if OpenAI's $29 billion valuation represented a real market or a temporary euphoria. Anthropic wanted $4.1 billion for a company with zero revenue and a model that hadn't launched publicly. For a firm like Menlo, this checked none of the boxes. Too expensive. Too early. Too risky.
Murphy invested anyway because he made a contrarian bet that matters more today than it did then: foundation models are infrastructure, not a winner-take-all platform. If you believe AI agents, vertical AI products, and specialized models all need different underlying engines, then the market can sustain multiple foundation model companies. Anthropic wasn't trying to beat OpenAI at its own game. It was building for enterprises that wanted safety, reliability, and a vendor that wouldn't suddenly pivot to consumer or get acquired by a cloud provider.
"I was blown away by his energy, his description of what they had built and why they thought that there was room in the market for another player."
The "heartburn" inside Menlo wasn't about Amodei or the technology. It was about violating discipline. Venture firms have rules for a reason: to avoid getting seduced by narrative when the numbers don't work. But Murphy saw something the spreadsheet couldn't capture. Amodei had been OpenAI's VP of Research before leaving to start Anthropic. He wasn't a challenger — he was a co-architect of the entire field who disagreed about where it should go next. That's not disruption. That's a fork in the protocol.
What makes this story signal and not just a victory lap: Menlo didn't lead the round. They knew it was outside their mandate but couldn't afford to miss it. So they participated, kept writing checks in every subsequent round, and co-launched the Menlo Anthology Fund to get closer to the Anthropic ecosystem. That fund structure is new. It's not just "invest in the company." It's "invest in the company and then build a dedicated vehicle to back everything in its orbit." This is how venture adapts when the companies get too big to fit the old models but too important to ignore.
Key dynamics at play:
- Foundation models require different venture math: $4B pre-revenue isn't insane when the TAM is "every company that writes software"
- The Pepsi-to-Coke thesis matters: markets with network effects still have room for a strong number two if the product is differentiated enough
- Anthropic's IPO timing (late 2026) signals confidence that public markets are ready to price AI infrastructure, not just hype
The Implication
If Anthropic's IPO succeeds, it proves that foundation models are venture-backable at scale and that the market will pay for choice in AI infrastructure. For founders, this means the next wave of massive outcomes won't just be OpenAI clones — they'll be companies that serve different customer bases with different risk profiles. Enterprises don't want to bet everything on one model provider any more than they wanted to bet everything on one cloud in 2010.
For LPs and other VCs watching this, the lesson is harder: sometimes the best deals break your rules. Menlo's $14 billion return came from ignoring their own playbook because Murphy recognized a foundational shift. That's not a strategy you can systematize, but it is a skill you can't afford to lose.