The American automobile was born as a machine of freedom — four wheels, an open road, a driver in command. Now the industry’s biggest players and its scrappiest upstarts are converging on a radically different vision: a vehicle with no steering wheel, no pedals, and in many cases, only two seats. It’s a bet that the future of urban transportation looks less like a personal car and more like a glorified pod, summoned by app, shared by strangers in sequence, and owned by a fleet operator who cares about one metric above all others: cost per mile.
Tesla’s Cybercab, unveiled last year to considerable fanfare and persistent skepticism, is the most prominent example. But it’s far from alone. Lucid Motors, Zoox (owned by Amazon), China’s Baidu, and a constellation of smaller companies are all converging on purpose-built autonomous vehicles designed from scratch for a world without human drivers. The common thread isn’t just the absence of a steering wheel. It’s the aggressive pursuit of a smaller, cheaper, more efficient form factor — one that jettisons the assumptions embedded in a century of car design.
As The Verge recently detailed, the logic behind the two-seater robotaxi is brutally simple. The average ride-hail trip in the United States carries 1.1 to 1.4 passengers. Building a five-seat vehicle to serve that demand is like dispatching a school bus to pick up a single child. Every pound of unnecessary structure, every square foot of unused cabin space, every extra tire represents wasted energy, wasted materials, and wasted money. Strip the vehicle down to two seats, remove the entire driver interface, and you get something dramatically lighter, cheaper to manufacture, and less expensive to operate.
That operating cost is everything.
Waymo and Cruise have spent years retrofitting existing cars — Jaguar I-PACEs, Chevy Bolts — with sensor arrays and computing hardware bolted onto platforms designed for human drivers. The approach works. Waymo’s vehicles currently provide over 200,000 paid rides per week across San Francisco, Los Angeles, Phoenix, and Austin. But retrofitting is inherently inefficient. You’re paying for a dashboard nobody looks at, a steering column nobody touches, mirrors nobody checks. The sensor mounts are compromises, perched on roofs and fenders in configurations the original engineers never intended. And the vehicles themselves carry the full weight and cost of a consumer car, including profit margins for the original manufacturer.
Purpose-built robotaxis promise to break that cycle. Tesla CEO Elon Musk has said the Cybercab will cost under $30,000 to produce, a figure that analysts have treated with varying degrees of credulity given Tesla’s history of optimistic cost projections. But even at a higher price point, a two-seat autonomous vehicle with no steering mechanism, no pedal assembly, no instrument cluster, and a simpler HVAC system would carry a meaningfully lower bill of materials than a comparable four- or five-seat EV.
Lucid, which has struggled to translate its luxury sedan technology into volume sales, is now openly pursuing the robotaxi market. CEO Peter Rawlinson has discussed a purpose-built autonomous vehicle that would capitalize on Lucid’s industry-leading powertrain efficiency. The company’s current sedan achieves over 500 miles of range on a single charge, a figure that suggests a smaller, lighter robotaxi platform could deliver extraordinary efficiency — and therefore extraordinarily low per-mile energy costs. For a fleet operator running vehicles 12 to 16 hours a day, energy cost per mile is a line item that compounds relentlessly.
Then there’s Zoox. Amazon’s subsidiary has taken perhaps the most radical approach, building a bidirectional vehicle with no discernible front or back, four seats arranged in facing pairs, and a top speed of 75 miles per hour. Zoox began testing on public roads in Las Vegas and Foster City, California, and represents a vision closer to a horizontal elevator than a traditional car. The design eliminates the need for U-turns — the vehicle simply drives in the other direction. It’s clever. Whether it’s commercially viable at scale remains an open question. Amazon has reportedly invested more than $2 billion in Zoox since acquiring it in 2020, and revenue remains negligible.
The Chinese market, as usual, is moving faster than most Western observers appreciate. Baidu’s Apollo Go service operates what it claims is the world’s largest autonomous ride-hailing fleet, with vehicles active in multiple Chinese cities. The company has introduced a purpose-built robotaxi, the RT6, with a reported manufacturing cost of roughly $37,000 — a fraction of what Waymo’s Jaguar-based vehicles cost to produce and equip. The RT6 features a detachable steering wheel, allowing it to function in both autonomous and human-driven modes during the transition period. It’s a pragmatic hedge, and one that Western manufacturers have largely avoided.
So why two seats specifically? The answer comes down to a collision between statistics and physics.
Ride-hail data consistently shows that the vast majority of trips involve a single passenger. Even when accounting for shared rides and occasional groups, the average occupancy hovers well below two. A two-seat vehicle captures the overwhelming majority of demand while dramatically reducing vehicle size, weight, and cost. The aerodynamic benefits alone are significant. A smaller frontal area means less drag, which means less energy consumed per mile, which means either more range or a smaller (cheaper, lighter) battery — or both.
As The Verge noted, the drag coefficient improvements from a smaller vehicle compound across millions of miles driven. For a fleet operator, shaving even a few percentage points off energy consumption per mile translates into meaningful savings at scale. And scale is the entire point. The economics of autonomous ride-hailing only work if utilization rates are high and per-mile costs are low enough to undercut both human-driven ride-hail services and, eventually, private car ownership.
This is the core financial thesis. Not selling cars to consumers. Selling miles to passengers.
The implications for the traditional auto industry are profound. If autonomous ride-hailing reaches cost parity with private car ownership — and proponents argue it will reach levels well below parity — then demand for personally owned vehicles in dense urban areas could decline significantly. Morgan Stanley analyst Adam Jonas has estimated that a fully autonomous ride-hail service could eventually offer rides at $0.25 to $0.50 per mile, compared to the roughly $0.70 per mile that AAA estimates for private car ownership when accounting for depreciation, insurance, fuel, and maintenance. At those prices, owning a car in a city starts to look like owning a horse in 1920: technically possible, emotionally appealing to some, but economically irrational for most.
But the gap between prototype and profitable fleet remains vast.
Waymo, despite its operational lead, is not yet profitable. The company doesn’t disclose detailed financials, but Alphabet has invested more than $5 billion in the unit. Cruise, majority-owned by General Motors, suspended operations entirely in late 2023 after a pedestrian was dragged by one of its vehicles in San Francisco, and has only recently begun a slow and cautious restart. The regulatory environment remains fragmented, with different cities and states imposing different rules, and public trust oscillating between curiosity and alarm with every headline about an autonomous vehicle behaving erratically.
Tesla’s approach adds another layer of complexity. Unlike Waymo and Zoox, which use lidar, radar, and cameras in combination, Tesla has committed to a vision-only system — cameras and neural networks, no lidar. Musk has repeatedly argued that lidar is an expensive crutch, and that a sufficiently advanced vision system can match or exceed human perception at a fraction of the hardware cost. The Cybercab is designed around this philosophy. If Musk is right, Tesla’s per-vehicle sensor cost will be dramatically lower than its competitors’, giving it a structural advantage in fleet economics. If he’s wrong, the consequences could be measured in lives.
The timeline remains contentious. Musk originally promised full self-driving capability by 2020. Then 2021. Then “next year,” repeatedly. The Cybercab is currently slated for production in 2026, a date that even Tesla bulls treat as aspirational. Meanwhile, Waymo is already operating commercially, expanding to new cities, and accumulating real-world data at a pace that no competitor can currently match. The question isn’t whether purpose-built robotaxis will eventually arrive. It’s whether the companies betting on them can survive the years of capital destruction required to get there.
There’s also the question of what passengers actually want. A two-seat pod optimized for cost efficiency is, by definition, a spartan experience. No trunk for groceries. Limited space for luggage. No room for a third friend who decides to join at the last minute. The vehicle becomes a pure transportation utility — a commodity measured in minutes and cents, not comfort or status. For short urban trips, this may be perfectly acceptable. For longer rides, airport runs, or any trip involving cargo, the limitations become apparent quickly.
Some companies are hedging. Zoox’s four-seat configuration offers more flexibility. Waymo’s current fleet of Jaguar I-PACEs seats four passengers comfortably and includes a trunk. And several Chinese manufacturers are developing robotaxi platforms in multiple configurations — two-seat, four-seat, and van-style — to serve different use cases within the same fleet. The idea is that an AI-powered dispatch system would match vehicle type to trip characteristics, sending a two-seater for a solo commuter and a larger vehicle for a family heading to the airport.
The manufacturing question looms large. Building a new vehicle from scratch is enormously capital-intensive, and the history of automotive startups is littered with companies that designed brilliant vehicles but couldn’t produce them at scale. Tesla itself nearly went bankrupt during the Model 3 ramp. Lucid has struggled to scale production of its Air sedan despite strong reviews. Rivian burned through billions before reaching modest volume. A purpose-built robotaxi, with its novel architecture and specialized components, presents manufacturing challenges that are different from but no less daunting than those faced by conventional EVs.
And yet the money keeps flowing.
Alphabet continues to fund Waymo. Amazon continues to fund Zoox. Tesla is allocating significant engineering resources to the Cybercab. Saudi Arabia’s Public Investment Fund, which backs Lucid, has shown willingness to sustain losses in pursuit of long-term strategic positioning. Chinese state-backed capital supports Baidu and several smaller autonomous vehicle companies. The collective bet is enormous — tens of billions of dollars deployed against the thesis that autonomous mobility-as-a-service will become one of the largest markets in the global economy.
The competitive dynamics are also shifting in unexpected ways. Nvidia, which supplies the computing hardware that powers most autonomous driving systems, has become a kingmaker of sorts — its Drive Thor platform is designed specifically for autonomous vehicles and is being adopted by multiple robotaxi developers. The company’s market position gives it unusual visibility into the strategies of competing AV companies, and CEO Jensen Huang has been vocal about his belief that autonomous vehicles represent a multi-trillion-dollar opportunity. Mobileye, the Intel subsidiary that supplies driver-assistance technology to dozens of automakers, is pursuing its own vision of autonomous ride-hailing, offering a turnkey system that traditional automakers could deploy in their own vehicles.
The insurance implications are staggering. Today, auto insurance is priced primarily on human driver risk — age, driving history, location, vehicle type. In a world of autonomous fleets, insurance shifts from a consumer product to a commercial one, priced on the safety record of the software rather than the behavior of the passenger. If autonomous vehicles prove significantly safer than human drivers — and early data from Waymo suggests they are, with lower rates of injury-causing crashes per mile driven — then insurance costs per mile could drop substantially. That’s another line item that tilts the economics toward fleet autonomy and away from private ownership.
Not everyone is convinced the two-seat form factor will dominate. Some transportation researchers argue that the efficiency gains from smaller vehicles will be offset by the need for more vehicles to serve the same demand, increasing congestion and complicating fleet management. Others point out that consumer preferences are stubborn — people like space, even when they don’t need it — and that a slightly larger vehicle might command a sufficient price premium to justify its higher cost. The market will sort this out, but it may take a decade or more of real-world operation before the optimal configurations become clear.
What is clear is that the automobile is being unbundled. For a century, a car was a single product that served as commuter vehicle, grocery hauler, road trip machine, status symbol, and personal space. The robotaxi thesis says that’s inefficient. Different trips have different requirements, and a fleet of specialized vehicles — small pods for solo commutes, larger vehicles for groups, vans for cargo — can serve those requirements more efficiently than a single general-purpose car sitting in a garage 95% of the time.
Whether that thesis survives contact with American consumer psychology is the trillion-dollar question. Americans have a deep, almost spiritual attachment to their cars. The vehicle in the driveway represents independence, identity, readiness. Asking people to give that up in exchange for a cheaper, more efficient alternative is asking them to change not just their transportation habits but their relationship with mobility itself.
The two-seat robotaxi is, in that sense, the sharpest possible expression of the bet. It’s a vehicle that does exactly one thing — move one or two people from point A to point B — and does it as cheaply as physics and engineering allow. No pretense of versatility. No concession to ego. Just motion, metered and monetized.
It might be the future. It might be a footnote. But right now, some of the world’s richest companies and most ambitious entrepreneurs are building it as fast as they can.
WebProNews is an iEntry Publication