Unlocking the full value of autonomous vehicle ownership — access the moment you need the car, earnings in every window you don't.

Autonomy is becoming something people own — Tensor, Lucid, Ford, and Nuro are all building toward personally-owned self-driving cars. A car that drives itself can earn while its owner doesn't need it, and even Tesla is opening its robotaxi network to vehicles it doesn't own. Idle time is becoming earning time.

But the full value of that ownership is still locked. Every major player optimizes for fleet revenue, utilization, or vehicle sales — no one optimizes for the owner. The problem isn't technology. It's alignment.

RoamingOS stands for the owner. We guarantee access the moment you need the car, then earn the most it can in every window you don't. Loyalty is to the owner alone: access is the hard constraint, earning only the slack beneath it.

Population baselineSF averageSegmentationcategory archetypesCommuteNight-shiftFreelanceWeekendClassificationwhich type?Digital twinfits their travelComplete poolingPer-owner data accruesIndividual calibration
Calibrated Access
How far it may roam, and when it must return
The car earns only as far as a calibrated forecast of its owner's need lets it return in time — access first, earning second. Illustrative simulation on mobility data.
Always
Owner first
The owner's access is the hard constraint — never traded for earnings
Calibrated
How we hold it
Honest about its uncertainty · the car returns early whenever we're unsure
The Surplus
Earning
The car earns only within the radius that still guarantees a timely return
AV
Owner · free
Owner · needs car
Ride request

How we get there

01
Stage 1 · Today's world
Maximize the passenger-AV owner's benefit
02
Stage 2 · Today's world
Maximize every AV owner's benefit
03
Stage 3 · A new world
The trusted operating layer for privately owned autonomous capital.

Why this is ours to build

Our advantages are not three things but one chain — each link earns the right to the next, and only the player holding all three reaches the end.

What we build now
The calibration engine — pretrained globally, calibrated to each owner's need.

The technical precondition for an owner to trust us with their car — and the scarce thing the whole model runs on.

What it qualifies us for
Neutrality — we don't build cars, so every brand can trust us between them.

Owning no fleet of our own, we can sit between an owner and any brand without ever competing — a structural neutrality no carmaker can copy.

North star · not yet built
Aggregation — cross-brand idle capacity as one operating system.

Enough supply and demand across every brand becomes one network that compounds and resists replication — the full operating system, and where the durable value lives, not yet built.

Many can predict. Many could aggregate. But only a participant that is at once neutral — holding no fleet of its own — and capable of inferring individual availability has the standing to assemble the network. Today that intersection is empty but for us.

Research

Predicting when an individual needs their car is uncharted — the literature does not exist — so we validate the engine directly in simulation on real human mobility data, before any vehicle moves.

Publications informing this work
Large-Scale Multi-GPU Based Parallel Traffic Simulation for Accelerated Traffic Assignment and Propagation · Transportation Research Part C: Emerging Technologies, 2024
Designing a Time-Driven Simulation Framework for Large-Scale Traffic Networks · ACM SIGSIM Conference on Principles of Advanced Discrete Simulation (PADS), 2024
Simulation-Based Optimization for Vertiport Location Selection: A Surrogate Model with Machine Learning Method · Transportation Research Record, 2024
Robust Reinforcement Learning Strategies with Evolving Curriculum for Efficient Bus Operations in Smart Cities · Smart Cities, 2024

Team

Built at the intersection of transportation science, machine intelligence, and systems engineering — with research roots in the JTL Urban Mobility Lab at MIT, California PATH and the MLDM Lab at UC Berkeley, and the Urban Freight Lab at UW.

MITUC BerkeleyUniversity of Washington

We are based in San Francisco.