How will the massive investments in AI generate returns? Goldman Sachs provides a specific answer: autonomous driving.
Currently, tech giants have invested nearly $700 billion in AI data centers, approximately ten times the level seen in 2020. Whether this capital can yield sufficient returns is a central question on Wall Street.
Analyst Mark Delaney from Goldman Sachs recently released an updated report focusing on "AI's Impact on Profit Pools," specifically evaluating the incremental profits that AI-powered transportation could create. The report argues that the commercialization of autonomous driving is accelerating and represents a significant source of return on AI investments.
**Robotaxis: Projected Nearly Threefold Increase Within a Year** Delaney had issued initial projections around a year ago, in June 2025. The latest update shows a substantial upward revision.
Forecasts for the US robotaxi market: * **2030:** Revised upwards from $7 billion to $19 billion. * **2035:** Further increasing to $48 billion.
On a global scale, Delaney projects the robotaxi market will reach $415 billion by 2035. He estimates that vertically integrated operators could achieve gross margins between 30% and 50%, translating to approximately $150 billion in gross profit.
The core rationale for this upward revision is the rapid expansion of autonomous vehicle deployment scales in the US, China, and Europe. Delaney notes that "these deployments rely both on proprietary technology from companies like Waymo, Tesla, and Pony.ai, and benefit from commercial Physical AI tools provided by companies such as NVIDIA (e.g., the Alpamayo platform)."
Signals of market demand are already emerging. Data indicates that Waymo has captured 30% of the market share in San Francisco just 20 months after commencing commercial operations. Concurrently, Google search data shows a sustained rise in consumer interest in robotaxis.
**Autonomous Trucks: An Even Larger Profit Pool** Beyond robotaxis, Delaney identifies autonomous trucks as another heavyweight source of profit.
Projected data: * **US Class 8 Autonomous Truck Market:** $16 billion by 2030, growing to $105 billion by 2035. * **Global Market:** Approximately $560 billion by 2035. * **Global Gross Profit:** Exceeding $135 billion by 2035, with a cumulative total of around $300 billion over the next decade.
Costs are also improving rapidly. Delaney anticipates that the per-mile cost for autonomous trucks will decline significantly by the end of this decade and stabilize around $2 per mile by the mid-next decade.
"Overall, we expect the global gross profit pool for autonomous trucks to grow from near zero in 2025 to approximately $135 billion by 2035, as cost structures improve and fleet scales expand," Delaney wrote in the report.
Several companies are already positioned in the North American market, including Aurora, Kodiak, Waabi, and Plus, all planning rapid fleet expansion in the coming years.
**Impact: Disruption Looms for $440 Billion Existing Market** The expansion of autonomous driving isn't just about creating new markets; it also signifies the displacement of existing industries.
Citing US Bureau of Labor Statistics data, Delaney estimates that the economic activity potentially impacted in the US amounts to roughly $440 billion, encompassing: * Wage income for taxi, ride-hail, shuttle, and freight drivers. * The portion of ride-hailing platform revenue allocated to drivers. * Potential declines in vehicle sales if personal travel demand is fully met by autonomous ride-hailing services.
The report directly addresses the ride-hailing industry: "Our base case forecast is that by 2030, autonomy will capture 5% of total US and Canadian ride-hailing bookings; under a bear case scenario, this proportion could rise to 16%."
Waymo's 30% market share in San Francisco serves as a concrete reference point.
Delaney concludes that the large-scale emergence of autonomous vehicles on roads in the US and other developed countries is an inevitable trend. Fleets are already operational, and true scaling is set to begin next year.
Comments