The Waymo 5nm ASIC chip is not just a piece of silicon: it is the clearest sign yet of how deeply the robotaxi company is building inward. Waymo this week disclosed the custom processor at the centre of its sixth-generation self-driving system, and the technical details tell a story about a company reorganising itself around the economics of scale.
Inside the Waymo 5nm ASIC Chip
The chip was described in a company blog post co-authored by VP of Engineering Satish Jeyachandran and Compute Lead Daniel Rosenband. Their job for the processor: to ‘process, fuse, and run advanced neural networks on raw sensor data in real time,’ before that data ever reaches the core inference engine of the self-driving stack.
Think of it as a highly specialised gatekeeper. The Ojai robotaxi carries 13 high-fidelity cameras. Every frame from every lens arrives as raw, unprocessed information. The Waymo 5nm ASIC chip handles temporal denoising for low-light perception and sensor fusion at the front end, passing a leaner, more structured data stream to the system’s CPUs, GPUs, and other accelerators in what Waymo calls a balanced, heterogeneous compute architecture.
The chip delivers more than 1,000 TOPS (trillions of operations per second), placing it in roughly the same performance bracket as Nvidia’s DRIVE AGX Thor automotive processor. Waymo claims the result is ‘unmatched efficiency and performance’ for high-density urban environments.
Fabricated on TSMC’s N5A automotive node, which entered volume production for automotive OEMs in 2025, the processor is one of the earliest large-scale commercial deployments of that node. Rosenband is scheduled to present a keynote titled ‘Compute in Motion: Challenges of Autonomous Driving’ at Hot Chips 2026 at Stanford on 24 August 2026, where further technical specifications are expected to be disclosed.
Waymo was careful to signal that this is not a single bet. The company described the ASIC as ‘just one of several exciting custom components we’re developing,’ pointing to a broader programme of in-house silicon.
The Numbers Behind the Ambition
The chip’s debut lands against a backdrop of fleet metrics that illustrate why cost discipline matters. Waymo’s sixth-generation Driver reduced total sensor count by 42% compared with prior generations. As of March 2026, the company had logged 220.6 million rider-only miles and was completing roughly 500,000 weekly paid trips. Its fleet stood at approximately 3,500 vehicles as of July 2026.
Those are large numbers for a service that does not yet turn a profit. The custom chip, alongside a leaner sensor suite and a lower-cost vehicle platform, is part of Waymo’s argument that unit economics can eventually work. Partners listed in the disclosure include AMD, Micron, Nvidia, Samsung, Sandisk, Socionext, and TSMC, suggesting the company is building deep supply-chain relationships even as it develops proprietary components.
Also Crosses $1 Billion, and the Lidar Probe Worth Watching
Elsewhere in the sector, Also, the startup that spun out of Rivian in March 2025, closed a $150 million Series D round led by Prysm Capital, with participation from Eclipse, Greenoaks, and MVP Ventures, according to an announcement published via Business Wire on 19 August 2026. The raise pushed Also’s valuation past $1 billion and brings its total funding to $455 million since launch.
The funding history tells its own story. Also raised a $105 million Series B in March 2025 when it launched as a Rivian spinout, led by Eclipse. A $200 million Series C followed in March 2026, led by Greenoaks, at a $1 billion valuation. The Series D maintains that valuation floor while the company’s scope has expanded from pedal-assist bikes and cargo quads to autonomous delivery vehicles, backed by a multiyear commercial agreement with DoorDash.
Rivian founder and CEO RJ Scaringe co-founded Also and remains chairman of both companies, per Tech Funding News. Prysm Capital co-founder Jay Park framed his firm’s commitment plainly: ‘We backed Rivian early because we saw the potential behind wonderfully designed, vertically integrated electric trucks, vans, and SUVs.’
A separate story worth tracking: the Idaho National Laboratory, a US Department of Energy facility, is conducting a funded review of whether Chinese-made lidar sensors could be hacked or remotely disabled if deployed at scale on American vehicles. The lab declined to discuss the project with reporters in both June and July 2026, and China’s leading lidar suppliers Hesai and RoboSense did not respond to requests for comment. Chinese-made lidar is currently legal in US passenger vehicles and is installed at three major American airports. The findings, whenever they surface, could shape procurement decisions across the autonomous vehicle industry for years.
Waymo’s chip keynote at Stanford on 24 August may be the next moment to watch: if Rosenband releases detailed efficiency or power-consumption figures, the gap between Waymo’s in-house silicon strategy and the off-the-shelf approaches taken by rivals will become considerably easier to measure.
