Segment disclosure beats the keynote. Nvidia's FY2020 annual report, filed February 20, 2020 and reached through EdgarBeast against the SEC 10-K, describes DRIVE AV for autonomous driving and DRIVE IX for in-vehicle AI assistance, alongside a scalable simulation solution. The framing is deliberate: these are products other companies buy.
That is the arms-dealer model, and for a money desk it is a more durable position than operating a robotaxi fleet. Nvidia does not have to win the autonomy race outright; it has to be the compute layer that whichever carmaker wins is built on. The bet is on breadth of design wins, not on a single deployed product.
The trade-off is visibility. Because autonomy revenue rides inside Nvidia's automotive line rather than a standalone robotaxi segment, an outside analyst cannot cleanly isolate how much of the business is self-driving compute versus infotainment and cockpit silicon. The filing tells you the products exist; it does not hand you a clean autonomy revenue number.
The capital story is the attraction. A platform vendor funds R&D and sampling, then collects as customers ship; it does not carry the capex of a vehicle fleet, the liability of operating one, or the regulatory exposure of removing a driver. As of this filing, that is a structurally cheaper way to be levered to autonomy than building cars.
What to track going forward: design-win announcements that turn into shipped volume, and any sign that automotive becomes large enough to warrant its own segment detail. Until then, the automotive line is the proxy, and its growth rate is the signal that the platform thesis is converting to revenue.
The honest limit: a 10-K that lists DRIVE AV as a product does not prove automakers will standardize on it, or that autonomy will arrive on the timeline buyers assume. It proves Nvidia has chosen to sell autonomy as infrastructure — a bet that the field's winners will all need someone else's compute.