The Robotaxi Business Is an Energy Business

Waymo is running roughly half a million paid rides a week and has set a public target of one million by the end of 2026. In February it raised $16 billion at a $126 billion valuation. On 8 July it announced fully driverless operations for four more cities: Las Vegas, Denver, San Diego and Tampa. That rollout is staged rather than simultaneous, beginning with Las Vegas and opening first to Waymo's own employees before the public.

The technology question is not fully resolved, but it is no longer the most interesting one.

The interesting question is whether a fleet of several thousand electric vehicles can be kept earning revenue for enough hours a day to justify what it cost to build. That is not an autonomy problem — it is an energy and operations problem, and it is being solved by an entirely different set of companies than the ones in the headlines.

Consider what actually limits a robotaxi fleet's revenue.

It's not whether the vehicle can drive. It can. The limit is how many hours per day it spends carrying paying passengers rather than sitting at a depot with a cable attached. Every hour plugged in is an hour of unrecovered capital. Every kilowatt hour bought at a peak tariff rather than an off-peak one comes straight out of the contribution margin. Every depot that cannot get sufficient grid capacity from the local utility caps the fleet size in that city regardless of how good the software is.

It is the same arithmetic that governs an airline, a bus operator or a car sharing business: utilisation, energy cost, dwell time, maintenance throughput.

None of it is novel. What is novel is that the companies confronting it were built as artificial intelligence research organisations, and are now discovering that their binding constraint is a transformer capacity conversation with a municipal utility.

The clearest evidence is who is doing the work.

On 15 June 2026, Element Fleet Management announced a multi-year partnership to provide end-to-end fleet operations for Waymo, beginning in San Diego and expanding to further markets. The published scope covers vehicle lifecycle management, charging infrastructure and energy management, maintenance coordination and operational fleet optimisation. Waymo retains the rider relationship in its own app and retains responsibility for validating the Waymo Driver. Everything between the depot gate and the curb now belongs to a partner. Element's chief commercial officer, David Madrigal, described what his firm contributes as “operational intelligence, disciplined execution.”

Element is not the first of these partners. It is roughly the seventh.

Transdev has been running Waymo fleets since 2019. Moove, the fintech-turned-fleet-operator, covers Phoenix and Miami. Avomo, the autonomous-vehicle arm of the Spanish fleet manager Moove Cars, runs the Waymo on Uber service in Austin and Atlanta; despite the similar name, it is a separate company from Moove. Avis Budget Group signed as fleet operations partner for Dallas in July 2025, taking on infrastructure, vehicle readiness, maintenance and depot operations. Lyft's Flexdrive subsidiary took depot operations when Nashville opened in April 2026. Hertz has established a new unit, Oro Mobility, whose early mandates include charging, maintenance, cleaning and depot staffing for Uber's robotaxi fleet in the San Francisco area.

Rental car groups, transit operators, European shared mobility businesses. Not one of them writes perception software.

Avis chief executive Brian Choi has been explicit that the Waymo agreement marks his company's shift from renting cars to supplying operational infrastructure to the wider mobility sector. That is a rental business repositioning itself as an operator of autonomous fleets, and it tells you where the value is expected to accrue.

An industry does not produce seven specialist operators for a problem that is easy.

Follow the financing and the same conclusion arrives faster.

On 24 June 2026, Terawatt Infrastructure, which counts Waymo among its customers, announced borrowing of up to $300 million: a five-year facility for up to $150 million with an option to add a further $150 million incrementally, arranged through a syndicate of global banks with RBC as lead arranger. It is the company's first commercial bank facility. Terawatt operates purpose-built charging sites for autonomous and commercial electric fleets and controls the full stack beneath them, including real estate, power management software and charging uptime. Chief executive Neha Palmer framed the raise as supporting robotaxis becoming “an attractive new asset class.”

Read that as a statement about lending rather than about robotaxis. Banks are now underwriting robotaxi charging infrastructure on project finance terms. The collateral is electrical.

The batteries have begun a second career as well. On 4 June 2026, Waymo announced a programme with B2U Storage Solutions to move retired packs from its all-electric fleet into grid-scale storage rather than straight to recycling, with first deployments in Texas and California and a stated ambition of hundreds of megawatts. Waymo has not disclosed how the economics are split, and B2U's chief executive is the one discussing monetisation, so treat the financial case as unproven. The strategic direction is not in doubt.

Zoox built the same foundations before it had earned a single fare. Panasonic Energy began supplying battery cells in early 2026 under a multi-year agreement, and its Hayward factory is cleared to build up to 10,000 vehicles a year. For most of 2026 Zoox charged nothing in any market, running free rides while it awaited a federal exemption for its steering-wheel-free vehicle; NHTSA granted that exemption on 30 July 2026, clearing the way to charge fares beginning in Las Vegas, though California still requires separate state permits. The cells were bought and the depots built well before that fare arrived.

Companies do not make multi-year supply commitments for a cost line they consider incidental.

When utilisation disappoints, the instinct is often to hire more autonomy talent, or to bring in a conventional chief operating officer with a generic mandate. Neither reliably helps. As with any scaling business, the right next hire depends on which constraint is actually binding.

If depot capacity is limiting fleet growth in a given city, the constraint is infrastructure and utility relations. The company needs a leader who has planned electrical capacity against a multi-year vehicle roadmap, negotiated connection agreements, and defended a depot capital plan to a lender or a board. This person is rarely found in software. They are found in charge point operators, utilities, bus fleet electrification programmes and industrial site development.

If the vehicles are available but not earning, the constraint is operational optimisation. The mandate is charging schedules against predicted demand, tariff arbitrage, rebalancing across a city, and getting cars back into service quickly after cleaning and inspection. This is a live discipline in European car sharing, where operators have been running exactly this optimization for years on thin margins.

If unit economics remain opaque, the constraint is financial. Someone must make visible what a vehicle actually costs per revenue hour once energy, depot labour, insurance, depreciation and idle time are properly allocated. Robotaxi businesses consume capital long before contract value arrives, and the honest picture is frequently uncomfortable.

If expansion keeps stalling at the city boundary, the constraint is regulatory and municipal. Permits, kerb access, incident reporting and utility interconnection are decided locally and slowly. The relevant experience is public affairs in regulated infrastructure, not automotive.

The correct sequence is not universal. What is universal is that naming the constraint precisely produces a better mandate than naming a title.

The leadership requirement also depends on which kind of company is asking.

Autonomy developers that own their fleets carry the full operating burden and need genuine depth in energy, depot and maintenance leadership, often earlier than they expect.

Autonomy developers operating through partners need a different profile entirely: commercial leaders who can select, contract with and hold accountable a network of fleet operators across many cities, with real operational literacy behind the vendor management.

Fleet operators now competing for autonomous mandates need executives who can win and retain those contracts while running physical sites at margin, which is a service business, not a technology one.

Charging infrastructure operators serving these fleets are effectively real estate and energy businesses with a software layer, and are increasingly hiring against project finance expectations rather than startup ones.

The phrase “autonomous mobility” now covers four quite different businesses. Their leadership needs cannot be reduced to a single hiring sequence.

The instinct is to recruit from other robotaxi companies. There are not many, they are well funded, and their people are expensive.

The better pool is adjacent and largely overlooked. European shared mobility operators have been solving fleet utilisation and charging optimisation at commercial scale, without abundant capital, for the better part of a decade. Our own work with Necture, the Vienna-based electric fleet management platform, is one example of what that expertise looks like from the inside.

Public transport electrification is another underused source. Bus operators have been through depot capacity planning, tariff negotiation and scheduled charging at fleet scale under public scrutiny, which is close to the robotaxi problem and further along.

Transferable experience has limits, of course. A charge point operator executive may understand energy but underestimate the pace of a venture-backed environment. A shared mobility leader may know utilisation but have never managed a lender relationship. The task is to be precise about which parts of the experience genuinely transfer and which must be present from day one.

What matters is evidence rather than adjacency. Has this person raised fleet utilisation against a measurable baseline? Have they negotiated a tariff that materially changed cost per kilometre? Have they built a depot network to a plan and defended it financially?

Bringing in an operations leader is only useful if the company will let them lead.

An autonomy-first organisation tends to treat operations as downstream execution, a function that receives decisions rather than shapes them. If depot siting, vehicle procurement and city launch sequencing are all settled before the operations leader is consulted, the hire will not deliver, and will most likely leave within eighteen months.

The deeper adjustment is cultural. Research organisations reward novelty. Operating businesses reward reliability and repeatability. Those value systems do not merge automatically, and the executive team has to decide deliberately that the second one now matters as much as the first.

The robotaxi companies that win will not necessarily be the ones with the best driving models. Several stacks already work well enough for commercial service. The differentiator is whether the vehicles are earning.

The useful question for any board in this sector is not which executive to hire next. It is this: what will prevent us from putting our next thousand vehicles into revenue service? Grid capacity? Depot throughput? Energy cost? Permits? Capital structure?

Once that is answered honestly, the leadership mandate writes itself.

At Key Search, we work with founders, investors and boards across robotics, autonomous mobility, EV technology and climate tech to translate growth constraints into clear leadership mandates, and to find the executives who can actually resolve them.

For the core driving stack, yes. But most companies at commercial scale already have that capability. The binding constraint is more often fleet operations, energy management and depot infrastructure, and those roles are frequently filled too late because they are treated as support functions rather than as leadership positions.

Predominantly in shared mobility, public transport electrification, charge point operators, utilities and commercial fleet management. These sectors have been running electric fleets at high utilisation under real cost pressure for years, which is precisely the problem robotaxi companies are now encountering at scale.

Both models are in use. Waymo works with several partners across different cities while retaining the rider relationship and responsibility for its driving system. The decision affects hiring directly: owning depots requires deep operational leadership in house, while partnering requires strong commercial and vendor governance capability instead.

Robotaxi fleets are electric fleets. Charging cost, grid capacity, tariff structure and battery lifecycle are core economics, not sustainability reporting. The executives with the most relevant scar tissue on those questions have generally built their careers in climate tech and e-mobility.

Before the constraint becomes visible in the numbers. If city launches are slipping on depot readiness or utility timelines, the capability was needed several quarters earlier. Grid connection and site development timelines are measured in quarters, not weeks, which makes late hiring expensive in a way that is difficult to recover.