Brind.
Part of TSMC is manufacturing AI chips for AMD and Nvidia, while the companies explore advanced scaling solutions.

How could Intel's foundry expansion affect Google's AI chip supply chain?

Already happening What happens if… A possibility worth considering, not a forecast.

Google secures multi-million unit custom AI chip orders from Intel Intel's strategic pivot into foundry services has resulted in Google securing substantial commitments for its custom AI processors. Specifically, Google is adopting Intel's Embedded Multi-die Interconnect Bridge (EMIB-T) packaging for its inference-class chips, such as the ninth-generation Tensor Processing Unit, codenamed "Humufish." This commitment includes an order reportedly covering more than three million units for 2028. This move is part of a broader strategic partnership, with Intel deepening its ties with Alphabet via Google Cloud, allowing Google to diversify its supply chain and leverage Intel's advanced manufacturing capabilities.

Reported by 10 independent outlets Written Thursday
Effect
Strong positive
How direct
2 steps, all reported
When
Over the long term
The story
No new developments lately

How it reaches Google

Reported by news outlets

Tap any step to see the evidence behind it.

The facts so far

As reported. Each one links to where it comes from.

  • Google's 2028 TPU order from Intel reportedly covers more than three million units.techtimes.com
  • Google is adopting Intel's EMIB-T packaging for its ninth-generation Tensor Processing Unit, codenamed "Humufish."techtimes.com
  • Intel's EMIB claims a roughly 50 percent cost reduction versus CoWoS-L or CoWoS-R at equivalent package sizes.techtimes.com
  • Intel is deepening its ties with Alphabet via a strategic partnership through Google Cloud.yahoo.com

Why it matters

For Google, securing large-scale, multiyear contracts for custom AI processors is critical to maintaining its competitive edge in the AI market. By adopting Intel's EMIB, Google is gaining a viable alternative to TSMC's CoWoS, which currently faces severe capacity constraints, with lead times stretching 52 to 78 weeks for non-Nvidia customers.

This diversification is strategically important because Google relies heavily on custom silicon, such as its Tensor Processing Units, for its inference-class workloads. The ability to secure high-volume, cost-effective supply from Intel allows Google to manage its production needs and scale its AI infrastructure without being solely dependent on one foundry's capacity.

What we don't know yet

  • Will Intel's 18A foundry process maintain the necessary yield levels to fulfill the multi-million unit orders?
  • How will the adoption of EMIB-T packaging affect the overall performance of Google's inference-class chips compared to CoWoS?

Is this still moving?

No new developments lately Reached 3 outlets in its first 24 hours
Reports
19
Developments
13
Repetition
68%

What would change this answer

TSMC resolves its CoWoS capacity constraints and reduces lead timesGoogle's need for alternative packaging solutions would diminish, potentially reducing the urgency of the EMIB commitment.
Intel fails to maintain the 98 percent yield milestone for EMIB-T at mass production volumeThe commercial viability of the EMIB-T platform would be jeopardized, forcing Google to rely more heavily on existing supply chains.

— Other considerations

Beyond this explainer

Explore connected ideas

Each opens a map of what the web connects to it, with Brind's news on it.

The ideas come from how millions of web pages place them near each other. The questions are Brind's AI reasoning, written only from this explainer and Brind's own reporting.

Who else could feel it

Other possible ripples from the same event.

Keep going

Brind's analysis is written by AI from the reporting linked above and can be wrong. This explainer is a possibility worth considering, not a forecast: steps marked as Brind's reasoning are not stated in the reporting. It is not investment advice; do your own research.