Waymo’s Custom Silicon: Vertical Integration as Competitive Moat
TL;DR: Waymo built a custom 5nm ASIC chip delivering 1,000+ TOPS to process sensor data in its Ojai robotaxi, signaling deeper vertical integration as the company races toward profitability across Los Angeles, Phoenix, and San Francisco.
The Hardware Play Behind Waymo’s Expansion
Waymo disclosed this week that it designed custom silicon specifically for its sixth-generation self-driving system, marking a significant shift toward in-house semiconductor development. The 5nm ASIC chip handles raw data preprocessing from the Ojai’s 13 high-fidelity cameras before it reaches the autonomous decision-making core.
The chip’s 1,000+ TOPS performance places it alongside Nvidia’s DRIVE AGX Thor, a benchmark automotive processor. This parity suggests Waymo achieved competitive compute density while optimizing for robotaxi-specific workloads rather than relying on general-purpose automotive silicon.
The operational implication is straightforward: custom silicon reduces latency, power consumption, and per-unit costs—three variables directly tied to operating margins in a low-margin transportation business model.
Why Vertical Integration Matters for AV Economics
Waymo has long positioned cost-per-mile as the bottleneck preventing profitability. The Ojai generation promised cheaper manufacturing and maintenance; the custom chip addresses the compute cost side of that equation.
By designing silicon in-house, Waymo reduces dependency on third-party processors, accelerates optimization cycles, and potentially secures manufacturing advantages through exclusive partnerships. The company listed AMD, Micron, Nvidia, Samsung, SanDisk, Socionext, and TSMC as collaborators—a supply chain typically reserved for tier-1 semiconductor vendors.
This mirrors Tesla’s approach with its Dojo infrastructure and Apple’s move to custom chips, signaling that autonomous systems companies view compute as defensible intellectual property.
Geopolitical Undertones in AV Supply Chains
In parallel reporting, the Idaho National Laboratory is evaluating whether Chinese lidar sensors pose security risks for U.S. autonomous vehicles, funded by industry players who remain publicly silent about the effort.
Rivian, GM, Ford, Kodiak, Lucid, Nuro, and Uber claimed ignorance when contacted—a response that suggests either compartmentalized procurement or deliberate information asymmetry. This dynamic underscores growing tensions over AV supply chain sovereignty.
Background: Waymo’s Fleet Acceleration
Waymo, the autonomous driving unit spun from Alphabet in 2018, has accelerated geographic expansion throughout 2026. The company recently opened its Ojai robotaxi to all riders in Los Angeles, Phoenix, and San Francisco, consolidating its position as the leading robotaxi operator in North America by service area coverage.
The Ojai represents the sixth generation of Waymo’s self-driving platform and incorporates learnings from five previous iterations deployed across 100,000+ miles of testing. Each generation has reduced sensor count, simplified mechanical design, and lowered manufacturing complexity.
Waymo’s economics hinge on reaching operational profitability per ride, a threshold requiring sub-$0.50 compute costs per mile. Custom silicon is one lever in that equation; expanded service areas provide demand to absorb manufacturing fixed costs.
The company remains privately held but operates under Alphabet’s balance sheet, giving it capital access that competitors like Aurora and Cruise cannot match. This financial buffer enables the kind of long-term vertical integration plays that constrain competitors.
Background: The Custom Chip Ecosystem
Custom silicon for autonomous systems is not novel—Tesla, Mobileye, and Nvidia have pursued variations for years. What distinguishes Waymo’s announcement is the scale of partnerships and explicit focus on preprocessing efficiency rather than end-to-end inference.
The 5nm process node, managed by TSMC, places the chip at the frontier of commercial semiconductor manufacturing. TSMC’s exclusivity agreements with major clients (Apple, Nvidia, AMD) typically prevent small-volume automotive plays; Waymo’s presence suggests either exceptional volume commitments or deep negotiating leverage through Alphabet.
AMD and Nvidia’s involvement as partners—rather than competitors—signals a layered compute architecture where custom silicon handles camera data fusion and Nvidia’s processors manage higher-level planning algorithms. This hybrid approach balances optimization with ecosystem flexibility.
Also’s Series D: A Separate but Related Signal
Rivian’s autonomous mobility spinout Also raised $150 million in Series D funding led by Prysm Capital, bringing total capitalization to $455 million since its March 2025 launch. The rapid progression from seed to Series D in 18 months mirrors Waymo’s timeline but from a lower capital base.
Also’s mission evolution—mentioned but not detailed in public reporting—likely reflects market reality: standalone AV software companies struggle without OEM integration or hardware control. Waymo’s vertical integration strategy implicitly validates this constraint.
