With Taiwan Semiconductor Manufacturing Company (TSMC) proposing its Foundry 2.0 model, some industry experts compare it with Intel's IDM 2.0. However, the challenges Intel faces today also reflect the inconvenient truth: everyone sees the benefit of market competition by having Intel compete with TSMC in providing foundry services, but are they truly competing?
The Semiwiki forum has supported Intel's leadership role in the American semiconductor industry despite TSMC taking 60% of the share in the foundry market. Yet in a recent note on TSMC's Foundry 2.0, it saw a different picture than the one described by Intel CEO Pat Gelsinger.
Although the "5 nodes in 4 years" progress was impressive and allowed Intel to provide a reliable alternative to customers and grab "non-TSMC" orders, the author questioned, "Will Intel take business away from TSMC on a competitive basis?" It's one thing to take orders away from Samsung Foundry; it's an entirely different matter if it's taken from TSMC. He quoted Nvidia CEO Jensen Huang, saying that TSMC and the foundry business are the true heroes of the semiconductor industry.
TSMC has achieved a landslide victory in the N3 process, and even Intel has become its customer. TSMC proposed Foundry 2.0 not as a strategic detour, but to include packaging, testing, and masks as a comprehensive product service and maintain customers' engagement.
Foundry 2.0 is clearly a frontal attack on Intel's IDM2.0. With the N2 process progressing smoothly, on track for mass production in 2025 while N2P features a further 5% performance at the same power or 5% to 10% power benefit at the same speed on top of N2, Intel's 18A is losing its advantage as a break-out node because Intel's CPUs based on 18A process are not expected to ramp up to high volume until 2026.
"N2P will support both smartphone and HPC applications, and volume production is scheduled for the second half of 2026," said C. C. Wei confidently in the earnings conference. Semiwiki now hinges hope on Intel's 14A to become the next break-out node.
TSMC's current technological prowess and scale of production capacity still seem to be leading the pack, and its overwhelming dominance is not to be underestimated. Gelsinger's original goal in promoting Intel's IDM 2.0 was to spin off its wafer foundry business and reduce America's reliance on Asian foundries.
Now, three years later, Intel continues to make progress in advanced manufacturing processes, but revenue growth is nowhere in sight. Can Intel's foundry business support customers' demand for advanced process node foundry services in the face of huge losses and keen competition?
It's a question that will continue to be a lesson for the US government regardless of which administration is in office as it continues to count on Intel to revitalize US semiconductor manufacturing.
