For the better part of a decade, American companies designing the world’s most advanced processors had exactly one realistic place to manufacture them: Taiwan Semiconductor Manufacturing Co. That monopoly-like grip is now facing its first serious challenge, and the pressure is coming from an unlikely alliance of federal dollars, rival fabs, and chipmakers who have grown uneasy about depending on a single supplier located 100 miles from mainland China.
As of mid-2026, billions in U.S. government subsidies are flowing into competing foundries on American soil. Samsung Electronics is building a massive facility in Taylor, Texas. Intel is retooling plants in Arizona, Ohio, and beyond to court outside customers. And TSMC itself, in legally binding filings with the Securities and Exchange Commission, has warned investors that well-funded rivals now pose a genuine competitive threat. The pieces are moving. The question is how fast.
Washington puts real money behind alternatives
The clearest evidence that the U.S. government wants to break TSMC’s stranglehold sits in two awards from the CHIPS Program Office at the National Institute of Standards and Technology.
Samsung received a preliminary agreement for up to $6.4 billion in direct federal funding to support its Taylor, Texas campus, where the company plans to produce chips at its most advanced 2-nanometer gate-all-around process node. The Taylor project, which also includes advanced packaging capabilities, would give fabless U.S. designers a domestic alternative for leading-edge logic for the first time.
Intel secured a separate preliminary award of up to $7.86 billion to expand capacity across facilities in Arizona, New Mexico, Ohio, and Oregon. A central goal of Intel’s award is standing up a contract manufacturing business, known as Intel Foundry, that can fabricate chips designed by other companies. Intel’s 18A process node is the flagship technology meant to attract those outside customers.
Together, the two awards represent more than $14 billion in federal commitments aimed squarely at creating credible alternatives to TSMC for the most advanced semiconductor work. That is not a research grant or a policy aspiration. It is construction money.
TSMC acknowledges the threat in its own filings
TSMC is not ignoring the shift. The company’s 2025 Form 20-F, filed with the SEC for the fiscal year ended December 31, 2025, includes risk-factor language stating that it faces “intense competition from other foundries with significant resources,” according to the SEC EDGAR filing. The same document discusses TSMC’s geographic concentration in Taiwan and its own expansion into the United States, Japan, and the European Union.
Risk-factor disclosures in annual filings are reviewed by corporate lawyers and auditors, and material misstatements carry legal penalties. TSMC has included competitive-risk language in prior filings, but the current context gives it sharper meaning: the “foundries with significant resources” now include companies backed by the largest semiconductor subsidy program in U.S. history.
TSMC is also investing heavily in its own U.S. presence. The company has committed more than $65 billion to a three-fab campus in Phoenix, Arizona, with the first facility already producing chips and additional plants targeting progressively advanced nodes. That investment underscores both the strategic importance of U.S.-based production and the competitive pressure TSMC feels to defend its position on American soil.
What designers are weighing, and what remains unproven
No public contract or earnings disclosure from a major fabless firm, whether Qualcomm, AMD, Nvidia, or Apple, confirms that a specific share of leading-edge wafer orders has shifted from TSMC to Samsung or Intel. Industry analysts and trade press reports indicate that large designers are actively evaluating backup suppliers, but evaluating is not the same as committing, and committing is not the same as shipping production wafers.
The gap between announcement and execution is real. Samsung’s Taylor fab has faced repeated construction delays; the facility was originally expected to begin production in 2024. Intel Foundry has struggled with yields on its most advanced nodes and has posted significant operating losses, prompting leadership changes and a restructuring that is still underway. Neither company has published node-specific capacity figures tied to its CHIPS-funded expansion, and detailed production timelines have not appeared on the NIST award pages.
Qualification cycles add another layer of delay. Before a chip designer can move volume production to a new foundry, it typically needs 12 to 24 months of testing to verify that the new fab can hit the required performance, power, and yield targets. That timeline means any designer serious about diversifying away from TSMC would need to be deep in qualification work right now to have a backup supplier ready by 2027 or 2028.
Industry voices signal a mindset shift
Public statements from executives and policymakers reinforce the picture painted by the filings and award documents. U.S. Commerce Secretary Gina Raimondo, who oversaw the initial CHIPS Act award process, said in 2024 that the subsidies were designed to ensure “we are never again in a position where we are overly dependent on any single country for the chips we need.” Intel CEO Pat Gelsinger, before stepping down in late 2024, repeatedly described Intel Foundry as a “resilient supply” option for customers who “cannot afford single points of failure in their manufacturing base.” Samsung Foundry head Siyoung Choi told attendees at the company’s 2024 foundry forum that the Taylor campus would offer “the most advanced nodes available anywhere, produced on U.S. soil.”
On the customer side, Qualcomm CEO Cristiano Amon told analysts during a 2024 earnings call that the company was “actively working with multiple foundry partners” for future chip production, a statement widely interpreted as a reference to Samsung and Intel alongside TSMC. Nvidia CEO Jensen Huang has been more guarded, praising TSMC’s execution while acknowledging that “supply chain resilience is something every company in our industry is thinking about very seriously.”
Analyst commentary tracks the same theme. Handel Jones, CEO of International Business Strategies, has estimated that TSMC’s share of leading-edge foundry revenue could dip below 80 percent by 2028 if Samsung and Intel hit their capacity targets. “The question is no longer whether alternatives will exist,” Jones said in a 2024 industry presentation. “It is whether they will be good enough, fast enough.”
The geopolitical math driving the shift
Behind the subsidy dollars and corporate filings sits a strategic calculation that has only grown more urgent. Taiwan, where TSMC fabricates the vast majority of the world’s most advanced chips, sits at the center of rising military tension between the United States and China. A disruption to TSMC’s Taiwanese operations, whether from conflict, blockade, or natural disaster, would cripple global supply chains for everything from smartphones to military systems.
That concentration risk is the core reason Congress passed the CHIPS and Science Act in 2022, and it is the reason the Commerce Department structured its awards to specifically target leading-edge production. Policymakers are not trying to replace TSMC. They are trying to ensure that TSMC is not the only option.
For chip designers, the calculus is shifting from “TSMC is the best, so we use TSMC” to “TSMC may be the best, but we cannot afford to have no alternative.” That change in mindset, more than any single fab opening or subsidy check, is what threatens TSMC’s dominance over time.
Samsung and Intel must now prove they can execute at scale
The federal funding is committed. The fabs are under construction. TSMC has told its shareholders, in writing, that the competition is intensifying. What has not yet happened is the hardest part: Samsung and Intel proving, at scale, that they can manufacture chips at the bleeding edge with yields and reliability that justify a designer pulling volume away from TSMC.
History suggests caution. TSMC built its lead over decades of relentless execution, and no amount of government money can substitute for that institutional knowledge overnight. But the structural conditions for a multi-supplier market in leading-edge semiconductors are more credible now than at any point in the last ten years. Designers who wait for public proof that the gap has closed risk arriving late to a capacity allocation process that rewards early movers. The smart ones are already in the door.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
