The segment finally moves from contracts toward chips. Nvidia's FY2023 annual report, filed February 24, 2023 and reached via EdgarBeast against the SEC 10-K, says it started production of the DRIVE Orin autonomous-vehicle SOC and introduced next-generation DRIVE Thor. After years of development-agreement framing, that is a meaningful change in the story.
Production of Orin matters because it is the point where Nvidia's autonomy business begins to look like a volume silicon business rather than an engineering-services business. Development agreements get you design wins; production chips get you per-vehicle revenue. This filing marks the start of that transition.
Introducing Thor in the same breath is the platform-vendor playbook: announce the successor while the current product ramps, locking automakers onto a roadmap. For a money desk, a credible next-generation part reduces the risk that a design win is a one-off and signals Nvidia intends to compound its position in the vehicle.
The capital logic stays attractive. Nvidia is shipping the compute layer that automakers building autonomy will pay for per vehicle, without carrying fleet capex or operating liability. As Orin moves into production, the automotive line gets a clearer path from pipeline indicator to scaled, recurring silicon revenue.
What to track from here: how quickly Orin production volume grows inside the automotive line, and whether Thor design wins accumulate before Orin has even scaled. The healthy pattern is overlapping generations — Orin shipping while Thor books future programs.
The honest limit: “started production” is not the same as high volume, and the filing does not isolate autonomy silicon revenue from the rest of automotive. What FY2023 establishes is direction — the autonomy business has crossed from development engagements into shipping product, which is the harder, more valuable side of the line.