Intel's 18A-P Pulls in Apple's Next M Chips While EMIB Reportedly Wins Google TPUv8e As Customer Confidence Amps Up

Intel's foundry business is gaining momentum, with reports that Apple will use Intel's 18A-P process for its next-generation M-series chips while Google has selected Intel's EMIB packaging technology for its TPU AI accelerators.

The dual wins represent a significant vote of confidence for Intel's turnaround strategy under CEO Pat Gelsinger, who has staked the company's future on rebuilding its manufacturing capabilities. Apple's decision is particularly noteworthy — the company has relied exclusively on TSMC for its custom silicon since transitioning away from Intel's x86 architecture for Macs. Bringing Apple back as a foundry customer would be a symbolic and financial milestone.

The 18A-P process is Intel's most advanced manufacturing node, competing directly with TSMC's 2nm technology. If it delivers on performance and yield promises, it could help alleviate the capacity constraints that have plagued the chip industry as AI demand surges.

Google's selection of EMIB (Embedded Multi-die Interconnect Bridge) for its TPU production addresses a different but related bottleneck. As AI chips grow larger and more complex, packaging technology — the way multiple chiplets are connected within a single package — has become as important as the manufacturing process itself. EMIB allows Google to build larger, more powerful AI accelerators by combining multiple chiplets in a single package.

The AI boom has exposed severe supply constraints at TSMC, which has struggled to meet demand from every major tech company simultaneously. Intel's foundry, if it can deliver competitive yields, offers an alternative that could reshape the semiconductor supply chain.

**What This Means For You:** If you follow tech stocks or the semiconductor industry, Intel's foundry wins are a signal that the company's turnaround has real momentum. If you're a consumer, more foundry capacity means more chips and potentially better availability for the devices and AI services you use. The TSMC bottleneck has been a real constraint on the pace of AI development — Intel stepping up as a viable alternative could accelerate the entire industry's trajectory.