HappyRobot lands $150M Series C at $1.2B for freight AI agents

Enterprise AI has spent two years mostly producing chatbots that draft suggestions for a human to sign off on. HappyRobot, a Madrid startup founded in 2022, is betting on the harder thing, agents that make the call, take the freight rate, book the dock slot, and investors just bid its valuation to $1.2 billion. Fortune reported the company closed a $150 million Series C co-led by Prysm Capital and Eurazeo, with Bankinter, Kfund, Koch Disruptive Technologies, Orange and T Capital also in the round. It follows a $44 million Series B less than a year earlier and a $15.6 million Series A led by a16z in December 2024.

The pitch is that HappyRobot's voice and text agents actually run supply-chain workflows for more than 150 enterprise customers, including DHL, Uber, Kuehne + Nagel, Naturgy and Repsol, and the company is now pushing into telecom, energy, utilities, airlines and financial services. CEO Pablo Palafox says revenue is up more than fivefold since the Series B and net dollar retention has topped 150%, with one large U.S. supply-chain customer reportedly expanding its contract by a factor of ten within a year. Take those specifics as reported, not settled; they are the founder's numbers, not audited disclosures.

Why this matters is less the valuation and more where the money is going. Freight brokers make thousands of price calls a day, and if a software agent can carry those negotiations reliably, the labor economics of an entire back office moves. But the same reason the category is exciting is the reason it is dangerous. As a16z general partner Anish Acharya put it, "If the model once in a while hallucinates the price of a million dollars, that could be a big problem." That is exactly the failure mode that separates a demo from a load-bearing production system, and it is the thing enterprise buyers will be measuring quietly regardless of the headline round size.

What the reporting doesn't give you is how HappyRobot is measuring or guaranteeing accuracy on those live negotiations, what the per-call error rate looks like, or which pieces of the DHL and Uber deployments are fully autonomous versus supervised. Prysm's Kerry Wei says "They're not trying to be a cool AI startup," and the phrasing is telling. The moat, if there is one, is the unglamorous work of embedding into logistics teams' actual phone systems and dispatch software. If HappyRobot can hold that ground while the hyperscalers ship general-purpose enterprise agents next year, the vertical-first bet pays; if the generalists ship a good-enough voice layer, the moat is thinner than the valuation implies.