Intel and UMC Collaborate on 12nm Process to Challenge TSMC

In a revealing report, Liang-rong Chen spotlighted a crucial shift in Intel and UMC’s business strategy: their co-development of a 12nm platform in partnership. At first glance, the alliance seems like a tactical hedge amid rising US-China tech tensions. But according to an industry source quoted in Chen’s report, the real objective is more direct:

“12nm is the most advanced node that doesn’t require EUV lithography, making it a sweet spot in terms of price-performance. Many customers stick with 12nm simply because it’s more cost-efficient.”

The same source noted:

“While UMC publicly positions the Intel collaboration as a hedge against China, deep down, it’s about grabbing market share from TSMC.”

This strategic pivot forms the foundation of UMC’s re-entry into the performance process game. And it marks a sharp departure from the company’s 2017 stance, when it exited advanced node development beyond 12nm. Now, in 2025, UMC is doubling down on this node, channeling most of its critical R&D into the Arizona-based project with Intel.

Read the full report by Liang here: Intel’s Gamble: Bringing in UMC to Tap into Taiwan’s Secret Sauce

A Strategic Alliance with Broad Reach

The UMC–Intel partnership was officially announced on January 25, 2024, and involves developing a 12nm FinFET process at Intel’s Ocotillo campus in Chandler, Arizona, specifically in Fabs 12, 22, and 32. UMC engineers are embedded in Intel’s ecosystem, collaborating on the process development, PDKs, and EDA tools, effectively creating a TSMC-compatible, “T-like” process flow while navigating IP boundaries.

This 12nm platform also serves a geopolitical purpose. Although not subject to EUV-related restrictions, it still offers a legally compliant and commercially attractive alternative to engaging with Chinese foundries like SMIC. SMIC continues to face strict export controls from the US, limiting access to advanced tools and EDA software necessary for sub-14nm development.

Moreover, Taiwan’s Ministry of Economic Affairs has added both Huawei and SMIC to its export control list, requiring approval before local firms can export high-end tech to these entities. These restrictions make UMC’s U.S.-based 12nm collaboration particularly attractive for customers navigating rising compliance and geopolitical risk.

In May 2025, UMC CFO Chi-Tung Liu confirmed that the majority of the company’s ongoing R&D is now focused on maturing the 12nm node, which is in the customer onboarding phase. UMC committed NT$15.6 billion (US$490 million) in 2024 to R&D, a large portion of which supports this joint venture.

At the same time, UMC has kept capital expenditure lean. It plans to spend US$1.8 billion in 2025, down from nearly $2.9 billion in 2024, demonstrating a disciplined capital allocation strategy aimed at maximizing ROI rather than node leadership.

UMC is also strengthening its global footprint. Its Fab 12i Phase III in Singapore is scheduled for volume production ramp in late 2026, with a target capacity exceeding 1 million wafers annually. This facility will provide backup capacity to the Arizona line, allowing UMC to offer customers geographically diversified options.

According to TrendForce, UMC holds 4.7% of global foundry revenue as of Q2 2025. That puts it just ahead of GlobalFoundries (4.2%) but far behind TSMC (67.6%). The Intel partnership is a bid to challenge TSMC in a niche but lucrative space where TSMC has few direct threats.

GlobalFoundries, the other major 12nm player, doesn’t offer TSMC-compatible processes. Its only major 12nm customer remains AMD, leaving room for UMC to onboard clients seeking greater design compatibility.

In parallel, Taiwan has moved to restrict exports to Huawei and SMIC, aligning itself more tightly with U.S. policy. This development reinforces the attractiveness of the Intel-UMC platform for clients that want mature-node performance without the geopolitical risk.

The partnership’s first tape-outs are expected in late 2026. Success here could trigger a domino effect, with other cautious clients committing to volume. Intel has reportedly placed no cap on the capacity UMC can utilize at the Arizona site, allowing scalability if demand spikes.