✦ The Future BetsNarrow moat

Nvidia (NVDA) — the future bets

Nvidia speculates the way a monopolist should — robots, robotaxis, ministries and cryostats, every one of them just CUDA wearing a new body.

Nvidia's present is the data center — three quarters of a hundred billion dollars a quarter of it — and the market prices that present exhaustively. The future bets are where Jensen Huang points the same playbook at markets that barely exist yet: sell the chip, give away the software, own the ecosystem before the industry knows it is one.

Edge Computing revenue by quarter ($B)$4.95BQ1 FY26$5.65BQ2 FY26$5.81BQ4 FY26$6.37BQ1 FY27$7.20BQ2 FY27NVIDIA Q1 and Q2 FY2027 CFO commentaries; robots, cars, PCs and workstations
The segment that holds robots and cars grew 27% to $7.2 billion, still about a twelfth of the data center.

Four are live. Physical AI: the Jetson Thor robot computer went on general sale in August 20251, the Isaac GR00T robot foundation models were opened to the field, and the adopter list — Amazon Robotics, Boston Dynamics, Agility, FANUC — reads like a census of the robot industry2. Automotive: $2.3 billion of fiscal-2026 revenue, up 39%3, with a contracted path to 100,000 Level-4 vehicles for Uber's network beginning in 20274. Sovereign AI: nation-states as a brand-new customer class, led by Saudi Arabia's HUMAIN — AI factories of up to 500 megawatts and several hundred thousand GPUs over five years5. And quantum, where Nvidia builds no qubits at all but wired seventeen quantum makers and nine national labs into its GPUs through NVQLink6.

Notice the pattern: none of these asks Nvidia to win a new game. Each one extends CUDA, the networking, and the systems business into a new body — a robot, a car, a ministry, a cryostat. That is the right way for a monopolist to speculate, and the honest arithmetic is equally instructive: all four together round to a low-single-digit share of revenue against the data center's flood. They are options, and the market — fixated on the AI-capex question — charges almost nothing for them. The one that already reports is the one to grade: automotive has guided contracts, named customers, and a growth rate. Watch that line, and watch whether robotics graduates from developer kits to fleets; the day either becomes a ten-billion-dollar business, this page stops being about options.

Moat trajectory: Widening

Widening from a small base: Jetson Thor shipped, the Uber contract turned autonomy talk into a 100,000-vehicle order book, sovereign deals stack up by the hundred-megawatt, and NVQLink signed seventeen quantum makers at launch. None is large yet; all four moved from slideware toward purchase orders inside eighteen months.

The number that tests this moat
Reported
Edge Computing revenue, latest quarter
$7.2B in Q2 FY2027, +27%

Nvidia's new Edge Computing platform holds the devices for physical AI: workstations, PCs, robotics and cars. Growth outrunning PC demand, which memory prices are slowing, would show robotics and automotive starting to count.

Source: NVIDIA Q2 FY2027 CFO commentary ↗
✦ Future bets — beyond today's moat
References
  1. ReportedJetson Thor went on general sale in August 2025.
    NVIDIA Newsroom — Blackwell-powered Jetson Thor generally available (Aug 2025): 7.5x the AI compute and 3.5x the energy efficiency of AGX Orin; 2,070 FP4 teraflops, 128GB — Aug 2025 · publ. Aug 25, 2025 · source ↗
  2. ReportedIsaac GR00T models opened; adopters incl. Amazon Robotics, Boston Dynamics, FANUC.
    NVIDIA Newsroom — new physical-AI models released as global partners unveil next-generation robots; Isaac GR00T open models and reference designs; adopters incl. Amazon Robotics, Boston Dynamics, Agility, FANUC, Hitachi — 2025–2026 · publ. 2026 · source ↗
  3. ReportedAutomotive revenue: $2.3B in fiscal 2026, +39%.
    NVIDIA Corporation, Form 10-K (FY2026) — Fiscal year ended late Jan 2026 · publ. Filed Feb 2026 · source ↗
  4. ReportedUber partnership: toward 100,000 Level-4 vehicles beginning 2027.
    NVIDIA Newsroom — NVIDIA and Uber partnership (Oct 2025): scaling Uber's global autonomous fleet toward 100,000 Level-4 vehicles beginning 2027, with OEMs incl. Stellantis, Lucid and Mercedes-Benz on NVIDIA DRIVE — Announced Oct 2025 · publ. Oct 2025 · source ↗
  5. ReportedHUMAIN: up to 500MW of AI factories, several hundred thousand GPUs over five years.
    NVIDIA Newsroom — HUMAIN (Saudi PIF) and NVIDIA strategic partnership (May 2025): AI factories of up to 500MW, initial 18,000 GB300 Grace Blackwell systems, several hundred thousand NVIDIA GPUs over five years — Announced May 2025 · publ. May 13, 2025 · source ↗
  6. ReportedNVQLink launched with 17 quantum builders and 9 national labs.
    NVIDIA Newsroom — NVQLink introduced, connecting quantum processors and GPU computing, with 17 quantum builders and 9 scientific/national laboratories on board — Announced late 2025 · publ. Oct 2025 · source ↗
Sources
Generated September 18, 2026