Mubadala led a $250M Series C into Moove, which runs autonomous fleets for Waymo in Phoenix and Miami with London announced. The structural point: autonomy developers want software margins, fleet owners hold depreciating assets, depots, maintenance labour and insurance exposure — and those capital profiles historically separate. Also where the jobs in autonomy actually are, and where the thesis breaks.
Robotaxi coverage is almost entirely about the autonomy stack — whose software drives best, who has the most miles, which city allows what. Moove's new raise points at the part of the business nobody writes about and everybody eventually has to solve: somebody has to buy the vehicles, park them, charge them, clean them, insure them and fix them, and that somebody has a completely different balance sheet than a software company.
Moove raised $250 million at a $2.1 billion valuation, led by Mubadala Investment Company and co-led by Woven Capital — Toyota's growth fund — and Ion Pacific, with participation reported from BlueCrest Capital Management and Sona among others, alongside existing backers including BlackRock, MUFG, Franklin Templeton and Uber. The company runs autonomous fleets for Waymo, with operations active in Phoenix and Miami and London announced as its first international autonomous market.
The useful comparison is not to Uber. It is to aviation and telecoms, where the operator of the asset and the provider of the service separated decades ago — not for strategic reasons but because the capital profiles were incompatible.
An autonomy developer wants software margins and rapid iteration. Its costs are engineering, compute and validation; its asset is code that improves without being replaced. A fleet owner has the opposite shape: depreciating physical assets, real estate for depots, maintenance labour, utilisation risk, and insurance exposure that scales with miles driven rather than with revenue.
Bundling those into one company means a software multiple applied to what is substantially a leasing business. That is precisely the tension that makes vertically integrated robotaxi economics so hard to read from outside. When a single entity reports on both, you cannot tell whether the unit economics are improving because the driving got cheaper or because the fleet got younger.
Separating them makes each legible. It also lets each be financed appropriately — equity for the software, asset-backed debt for the vehicles — which is a mundane point with large consequences, because fleets financed as software are expensive fleets.
The use of funds tells you what this layer actually is. Moove says it will hire around 350 people for the AV business, growing that workforce from roughly 150 to roughly 500 by the end of the year, and build automated depots it calls Nests — robotics-enabled facilities where self-driving vehicles are charged, cleaned, serviced and inspected around the clock with minimal human intervention.
That is the more honest description of the labour transition in autonomy than either the boosters or the doom-mongers offer. Driverless does not mean personless. It means the people moved from the front seat to the depot. The jobs change character substantially — fewer, more technical, concentrated in facilities rather than distributed across a city — and it would be wrong to present that as neutral. But the claim that autonomous fleets are unstaffed is simply false, and the Nests model is what the staffing actually looks like.
It is worth noting where these figures come from. The vehicle count of roughly 42,000 across 29 cities in 13 countries, and approximately $420 million in annual recurring revenue, are Moove's own numbers relayed by reporters, not audited disclosures. The hiring plan is a stated intention. Treat all of it as company-reported.
The separation might be temporary, and the honest version of this story says so.
The first risk is insourcing. Waymo could bring fleet operations in-house once route density justifies the fixed cost — exactly as it has already done with some depot functions. The logic that makes outsourcing sensible at low density inverts at high density: below a threshold, a specialist operator spreads depot capital across multiple customers; above it, the customer has enough volume to justify its own. A fleet-operations layer that is valuable during the scaling phase can become redundant precisely when the market it serves succeeds.
The second is captive-supplier risk. A specialist operator's leverage depends entirely on serving multiple autonomy developers. One that derives most of its AV revenue from a single partner is not an infrastructure layer; it is an outsourced department with equity investors. The distinction shows up in pricing power, contract duration, and what happens at renewal. Watch whether Moove's AV customer list broadens beyond Waymo — that, more than city count, determines whether this is a durable position.
The third is that the pivot is real and recent. Moove was founded in 2020 by Ladi Delano and Jide Odunsi to finance vehicles for ride-hailing drivers in Africa. Becoming the operations layer for autonomous mobility is a genuine change of business, not an extension of one, and should be described as such. The existing fleet-financing operation gives it depot experience and vehicle-lifecycle knowledge that a pure startup would lack — which is the case for the pivot — but the AV business is roughly 150 people today.
For a fleet-operations layer to hold durable margin, three conditions have to hold together. Depot automation has to produce a real cost advantage over an autonomy developer doing it in-house, which means the robotics in the Nests have to work rather than merely exist. The customer base has to diversify enough that no single partner can dictate terms. And utilisation has to be high enough that the asset-heavy balance sheet earns its cost of capital, which depends on demand the operator does not control.
None of those is guaranteed, and the $250 million is a bet that they resolve favourably before insourcing pressure arrives. But the structural observation stands regardless of whether this particular company wins: the robotaxi business has two halves with incompatible financial characteristics, and the industry is now pricing them separately. That is a more informative development than another autonomy benchmark.
It varies by market and increasingly by design. In Moove's arrangement, the fleet-operations partner manages the vehicles and depot infrastructure for the autonomy developer. The broader trend is toward separating the two, because a company optimising for software margins and a company holding depreciating vehicles want very different balance sheets and very different financing.
DrafterDaily covers where value and risk actually sit — unit economics, capital structure, and the parts of a story the product announcements leave out.
Read more Business coverageThe headline was the 94% drawdown from a $3.44 billion peak. The more useful arithmetic is the cash sitting inside the thing being bought — and the fact that everyone except the angels was paid in another private company's paper.
The biography is the least interesting part. What the board actually did was unbundle the CEO job — keeping Tim Cook on with a defined policymaker-engagement remit — and that separation says more about Apple's next decade than the appointment does.
Every outlet reported the size of Meta's teen-safety settlement and the screen-time limits. Almost nobody read the payment structure — where roughly $5.3bn is released only if two competitors act.