· Event impact

Broadcom and Google Highlighted for Co-Developed AI Tensor Processing Units (TPUs)

Type: supply_chainConfidence: 0.8Verified: keep
This highlights a significant alternative to the merchant silicon market (dominated by Nvidia), where hyperscalers partner with custom silicon designers like Broadcom to create bespoke, cost-efficient AI hardware.

Transmission path

This highlights a significant alternative to the merchant silicon market (dominated by Nvidia), where hyperscalers partner with custom silicon designers like Broadcom to create bespoke, cost-efficient AI hardware.

Market mechanism

This highlights a significant alternative to the merchant silicon market (dominated by Nvidia), where hyperscalers partner with custom silicon designers like Broadcom to create bespoke, cost-efficient AI hardware.

Extended read

While much of the market's attention is on Nvidia's GPU dominance, a recent report brings focus to the successful custom silicon partnership between Google and Broadcom. The two companies co-develop Google's Tensor Processing Units (TPUs), which are the backbone of Google's internal AI efforts and a key offering on Google Cloud. This vertical integration strategy offers Google a potential long-term cost advantage over cloud rivals that rely more heavily on merchant silicon. For Broadcom, the partnership provides a large, durable revenue stream in the high-end custom AI chip market. The scale of the operation is underscored by Google Cloud's reported $460 billion backlog, a significant portion of which is likely tied to AI workloads running on this custom hardware.

Exposed assets

AVGO · GOOGL · NVDA

Countries: USA

Continuous event tracking, options routing, and portfolio overlap for this event and the assumptions it moves live inside Market Ontology. Start a trial →