Skip to content
gentic.news — AI News Intelligence Platform
Connecting to the Living Graph…

Listen to today's AI briefing

Daily podcast — 5 min, AI-narrated summary of top stories

A sleek white Waymo robotaxi with sensor dome parked on a sunlit city street, surrounded by modern buildings and…
Big TechBreakthroughScore: 85

Waymo Builds Custom 1,000-TOPS Chip, Cuts Nvidia Dependence

Waymo built a custom 1,000-TOPS robotaxi chip on TSMC 5nm, cutting Nvidia and AMD reliance. The move targets cost reduction as Zeekr-based fleet scales.

·4h ago·3 min read··16 views·AI-Generated·Report error
Share:
Source: the-decoder.comvia the_decoderSingle Source
Did Waymo build its own chip to replace Nvidia in its robotaxis?

Waymo built a custom chip for its robotaxis that delivers over 1,000 TOPS and is manufactured by TSMC on a 5-nanometer process. The Alphabet subsidiary says the chip processes sensor data faster and runs AI models, cutting reliance on Nvidia and AMD. It ships in the new Zeekr-built robotaxi.

TL;DR

Waymo developed custom robotaxi chip · Delivers over 1,000 TOPS on 5nm TSMC · Ends reliance on Nvidia and AMD silicon

Waymo has built a custom robotaxi chip delivering over 1,000 TOPS, manufactured by TSMC on a 5-nanometer process, Bloomberg reports. The Alphabet subsidiary is ending its reliance on Nvidia and AMD silicon.

Key facts

  • Chip delivers over 1,000 TOPS
  • Manufactured by TSMC on 5-nanometer process
  • Waymo previously relied on Nvidia and AMD
  • Chip ships in new Zeekr-built robotaxi
  • Alphabet subsidiary developed chip in-house

Waymo has developed a custom chip for its self-driving taxis that is already running in the company's latest generation of vehicles, Bloomberg reports. The Alphabet subsidiary says the chip processes sensor data faster and runs the AI models that help its robotaxis understand and react to their surroundings. Until now, the company relied on chips from Nvidia and AMD.

The chip delivers more than 1,000 TOPS (trillions of operations per second) and is manufactured by TSMC on a 5-nanometer process, according to the company. That puts it on par with Nvidia's current autonomous driving systems. The move is meant to bring costs down.

Why in-house silicon wins on economics

The shift is a direct attack on Nvidia's pricing power in the automotive segment. Waymo is installing the chip in its new robotaxi, which it's building with Chinese manufacturer Zeekr. Vertical integration lets Waymo control its bill of materials, a critical lever when scaling a fleet that must amortize hardware costs across thousands of vehicles. Nvidia has not commented on the design win loss, and the company did not disclose the per-unit cost of its custom silicon.

The broader de-Nvidia-ification trend

Waymo joins a growing list of hyperscalers and OEMs designing custom accelerators. Google, Waymo's parent, has long built its own TPUs for data center workloads. Tesla has shipped its own FSD computer for years. The pattern is consistent: at scale, the Nvidia tax becomes unacceptable. The difference here is the target is not training but the edge inference workload inside a moving vehicle, where power efficiency and latency matter as much as raw TOPS.

Waymo's 5nm process choice is notable for an edge chip — many automotive parts use older, cheaper nodes. The decision signals performance was prioritized over cost at design time, with the expectation that volume production on Zeekr's platform will drive unit economics down.

What remains unclear is whether the custom chip handles all inference tasks or is paired with a co-processor for redundancy. Waymo has not disclosed the full system architecture, nor has it said whether the chip will be offered to other autonomous vehicle developers. The company also did not disclose the chip's power draw or whether it supports over-the-air updates to its neural networks.

Key Takeaways

  • Waymo built a custom 1,000-TOPS robotaxi chip on TSMC 5nm, cutting Nvidia and AMD reliance.
  • The move targets cost reduction as Zeekr-based fleet scales.

What to watch

Waymo just raised $5.6 billion to spread robotaxis …

Watch for Waymo's fleet expansion numbers on the Zeekr platform and any disclosure of per-mile hardware costs in its next investor update. Also track whether Nvidia announces a pricing response for its automotive DRIVE platform, and whether other robotaxi operators — Cruise, Zoox, or Baidu — follow with their own custom silicon.


Source: the-decoder.com


Sources cited in this article

  1. Bloomberg
Source: gentic.news · · author= · citation.json

AI-assisted reporting. Generated by gentic.news from 1 verified source, fact-checked against the Living Graph of 4,300+ entities. Edited by Ala SMITH.

Following this story?

Get a weekly digest with AI predictions, trends, and analysis — free.

AI Analysis

This is the strongest signal yet that the automotive AI chip market is bifurcating. Nvidia's DRIVE platform has dominated robotaxi development for years, but Waymo's move validates the hyperscaler playbook: when you control the full stack — sensors, compute, software, and the vehicle — the marginal cost of a custom chip is justified by removing the supplier margin. Tesla proved the model works at scale; Waymo is now following with a different sensor architecture. The 1,000 TOPS figure is telling. It matches roughly the performance of Nvidia's automotive-grade Orin/Thor-class parts, suggesting Waymo is not trying to leapfrog on raw compute but rather match it at a price point it controls. The 5nm process choice is aggressive for automotive, where long product lifecycles and thermal constraints often push designers to older nodes. That Waymo chose 5nm suggests they are confident in volume and thermal management within the Zeekr platform. The more interesting strategic read: Waymo is preparing for a world where it may sell its autonomy stack — including hardware — to other OEMs. Building its own chip gives it a differentiated product to license, not just a cost-saving measure. If Waymo starts offering the chip externally, it becomes a direct competitor to Nvidia in the AV market, not just a customer that walked away.
Compare side-by-side
Nvidia vs Waymo

Mentioned in this article

Enjoyed this article?
Share:

AI Toolslive

Five one-click lenses on this article. Cached for 24h.

Pick a tool above to generate an instant lens on this article.

Related Articles

From the lab

The framework underneath this story

Every article on this site sits on top of one engine and one framework — both built by the lab.

More in Big Tech

View all