Billionaire Ken Griffin Buys 2 AI Stocks Chasing a $1 Trillion Market Opportunity in Robotaxis (Hint: Not Tesla)

Robotaxis represent an estimated ~$1 trillion U.S. opportunity; Nvidia and Amazon are highlighted as likely winners. Nvidia supplies GPUs and a software stack (Omniverse, Cosmos, Alpamayo, Hyperion) for autonomous driving, with management forecasting 'hundreds of billions' in robotaxi revenue over the next decade and Street estimates ~38% annual EPS growth over the next three years; the stock trades near ~35x forward earnings. Amazon-owned Zoox has provided >350,000 autonomous rides, is testing in Austin and Miami, has applied to operate up to 2,500 commercial robotaxis (NHTSA decision expected in early April), and Morgan Stanley projects Zoox could capture ~12% of autonomous rides by 2032; Amazon's EPS is modeled to grow ~19% annually and trades near ~29x forward earnings.

Analysis

The primary structural advantage at play is not a single product but a vertically integrated stack that couples high-performance compute with simulation-trained models and telemetry loops — that combination creates data-driven switching costs for fleet builders. If engineers standardize on a full-stack vendor, sensor and component suppliers will face winner-take-most dynamics: a small set of compute/software platforms could capture a disproportionate share of recurring software and OTA revenue even if hardware margins compress.

For fleet operators, the second-order lever is monetization beyond fare-per-mile: local commerce, targeted in-ride advertising, and route-level logistics arbitrage can add meaningful EBITDA per ride without changing vehicle economics. A conservative scenario where ancillary revenue adds $0.50–$1.50 per trip could convert an otherwise marginal route into a sustainably profitable corridor within 3–7 years; incumbents with cloud/ads stacks have an asymmetric advantage here.

Near-term catalysts live on the supply and regulatory side. Constrained GPU supply cycles, node transitions at foundries, and high warranty/insurance costs can compress volumes and delay commercialization on a 6–18 month cadence; by contrast, scientific/regulatory milestones (liability frameworks, certification regimes) determine whether fleets scale over multi-year horizons. The biggest structural risk is vertical vertically-integrated OEMs or hyperscalers designing custom accelerators and bypassing third-party stacks, which would hollow out implied software economics.

The market is currently conjoining two distinct value pools — immediate datacenter AI demand and long-term fleet services — into a single multiple. That conflation overstates short-term optionality from vehicle fleets while understating durable cashflow from enterprise AI, creating asymmetric trade opportunities that favor exposure to compute/software IP and underweight commodity hardware paths.

Key Decisions for Investors

• Long NVDA (12–18 months): buy a calendar of 9–12 month call spreads to express convexity to datacenter + software moat (e.g., buy 12-month 25% ITM calls, sell 25% higher strike 12-month calls). Risk = premium paid; reward = >2x payoff if NVDA captures outsized enterprise AI spend and licensing flows from fleets.
• Pair trade — Long AMZN / Short INTC (12 months): overweight AMZN equity for cloud/ads optionality + fleet adjacencies, financed by a short in INTEL to express the narrowing margin on commodity silicon. Target: 1–2% portfolio size, stop-loss 12% each leg. Expected asymmetric reward if software/IP wins and commodity fabs lag.
• Relative-value — Long NVDA / Short TSLA (12 months): NVDA benefits from platform licensing and recurring software, TSLA exposed to execution risk on its vertically integrated stack. Size as a 0.5–1.0% net exposure, risk-managed with options (buy puts on TSLA vs call spread on NVDA); intended payoff if market re-rates platform premiums.
• Event hedge — buy insurance into regulatory risk (6–12 months): purchase out-of-the-money puts on a small basket of mobility operators (e.g., short-dated UBER puts or index hedge) to protect against sudden regulation/ liability shock that compresses ride volumes. Cost should be sized at <0.5% portfolio to cap tail exposure.