US semiconductor reshoring efforts collide with the AI chip boom

The U.S. is pushing to rebuild domestic semiconductor capacity through incentives, tariff structures, and strategic partnerships, while leading foundries expand their U.S. manufacturing footprints. At the same time, AI is amplifying demand for logic and memory at a pace that is lifting industry revenues to record highs and amplifying the consequences of any bottleneck or allocation shift across the supply chain.  

US-Taiwan deal and Arizona expansion accelerate reshoring

After lengthy negotiations, the semiconductor relationship between the U.S. and Taiwan has reached a new era of formalized cooperation. On January 15, 2026, the two signed a comprehensive trade and investment agreement aimed at strengthening U.S. semiconductor reshoring efforts while easing tariffs on Taiwan’s most vital industry.  

At the center of the deal is a commitment to $250 billion in direct investment from Taiwanese semiconductor and technology firms into the U.S., paired with an additional $250 billion in credit to support future investment. Notably, the $100 billion already committed by TSMC in 2025 for its Arizona expansion is included in these figures.  

In reciprocation, the deal includes a 15% cap on broad Taiwanese exports, lowering the duty from 20%. However, chipmakers that expand production in the U.S. will see lower tariffs on semiconductors and manufacturing equipment, with some allowed to be imported duty-free. The deal also includes provisions for 0% tariffs on generic pharmaceuticals, aircraft components, and “unavailable natural resources,” according to the U.S. Commerce Department.  

For advocates of U.S. semiconductor reshoring, this deal is a significant breakthrough. Pairing industrial-scale investment with clearly aligned trade policy will create an environment in which capital deployment comes with lower risk.  

A Reuters report frames the move as a critical geopolitical maneuver that reiterates the importance of semiconductors as the heart of U.S.-Taiwan relations. Rather than the punitive trade actions that have often been the go-to strategy of the second Trump administration, this latest deal leans on incentives and policy stability to shift high-value manufacturing westward.

Nowhere is this more visible than in Arizona, where TSMC’s Phoenix expansion project is accelerating. The $100 billion investment made by the chipmaker last year was initially seen as a hedge against U.S. tariffs. Now, it is the central pillar of TSMC’s global manufacturing strategy and takes on a more important role in broader political relations.  

New regional reports suggest TSMC is preparing for capacity additions beyond 3nm production, pointing to future readiness for 2nm-class logic and advanced packaging lines.  

For OEMs and EMS providers, this deal is one to note. As momentum for reshoring builds alongside AI-driven demand surges, allocation cycles will tighten. Moreover, sizable price hikes will continue to put pressure on tight margins.  

Forward-thinking sourcing strategies that include lifecycle planning and franchise diversification will be critical in managing BOM exposure to geopolitical and capacity-related risks. Sourceability can help customers map vulnerabilities in their supply chain, identify alternate parts across design generations, and secure much-needed continuity through our authorized and diversified supplier network.  

AI pushes semiconductor revenues past $1T in 2026

Omida’s latest forecast signals that global semiconductor revenues will surpass the $1 trillion mark in 2026, a feat that underscores how deeply artificial intelligence is reshaping demand fundamentals across the industry. According to the release, the milestone is the result of a concentrated surge in memory and logic IC revenue growth fueled by data center expansion to support AI workloads.  

Omida projects the total semiconductor market will expand some 30.7% year-over-year in 2026. That growth would outpace 2025’s already strong mark of 20.3% en route to driving revenues past the trillion-dollar threshold for the first time.  

Memory and logic segments are a catalyzing force for the industry’s growth. DRAM and NAND memory ICs, which form the backbone of AI training and inference infrastructure, are expected to see “unprecedented” expansion as data centers scale and as higher memory pricing persists. These markets are complemented by broad gains in logic ICs, which are equally critical to satisfying hyperscale and enterprise AI compute demand.  

According to Omida, computing and data storage is projected to lead all segments in 2026, eyeing a 41.4% year-over-year growth rate to exceed $500 billion. Hyperscale capex is behind much of that expansion. Estimates suggest that the top four hyperscalers (Amazon, Microsoft, Google, and Meta) will invest roughly $500 billion this year alone with additional room for growth.  

Of note, without the contributions from memory and logic ICs, the industry’s overall revenue growth would collapse from 31% to 8%. This startling reality illustrates just how disproportionate AI’s influence has become within the chip sphere.  

For semiconductor firms and supply chain leaders, Omida’s 2026 projections carry a few important takeaways.  

• Allocation pressure hasn’t peaked. With memory and logic segments commanding an outsized share of revenue growth, fabs are likely to continue reallocating capacity toward AI-related products. As a result, availability for adjacent categories will tighten in the months ahead.  ‍
• Prices and pricing volatility will increase. The premium pricing environment for memory chips, especially those tailored to AI workloads, will persist as demand is expected to outstrip supply well into 2026.  ‍
• Hyperscaler capex dynamics are shaping market cycles. As large AI infrastructure spenders influence capacity planning and procurement decisions, OEMs and EMS providers must take note. Strategic sourcing plans will be essential to avoid costly delays or spot buys.  

Sourceability’s deep market insights and more than a decade of strategic sourcing expertise help customers anticipate demand shocks like the one we’re seeing in the memory market today. Our market intelligence tools and teams of experts can help you strengthen your diversified sourcing strategy and mitigate allocation risk as AI continues to redefine long-held dynamics in the semiconductor industry.