Trump chip tariff is back in the headlines. The Trump administration is weighing a second round of semiconductor tariffs that would eliminate the data center, startup, and consumer-device carve-outs that have protected the AI infrastructure buildout since January, and would extend duties to servers, laptops, and gaming consoles built with chips already taxed under Phase 1. According to a Politico report cited by Tom’s Hardware, eight people familiar with the deliberations described a fundamental shift away from Phase 1’s narrow exemptions toward a broader regime that could reshape the economics of AI data center construction in the United States.
The timing of the news landed awkwardly. On the same day the talks were disclosed, SK Hynix broke ground on a more than $4 billion advanced-packaging facility at Purdue Research Park in West Lafayette, Indiana, the South Korean chipmaker’s first US high-bandwidth memory production hub. The ceremony was attended by Indiana Governor Mike Braun and Senator Todd Young. The juxtaposition captured the central tension in the administration’s semiconductor agenda: courting foreign investment with one hand while weighing tariffs that could cost those same investors billions more per year with the other.
Trump chip tariff — What Phase 1 Actually Protected
Presidential Proclamation 11002, signed on January 14, 2026, imposed a 25% tariff on a specific set of advanced AI accelerators, including Nvidia’s H200 and AMD’s MI325X, and explicitly labeled the action “Phase 1.” The same document carved out six exemption categories covering chips used in US data centers, research and development, startups, repairs, non-data-center consumer and industrial applications, and public-sector uses. Those categories together account for the overwhelming majority of chip procurement by American technology companies. Removing them would represent a structural reversal of the policy’s original design.
Four of the Politico sources, independently confirmed by Tom’s Hardware, said Commerce officials have indicated in private talks that the Phase 1 exemptions may not carry into Phase 2. Three of those sources said recent industry lobbying sessions moved against the industry’s position, suggesting the administration has not found the sector’s arguments persuasive. One person involved in the talks put the timeline for meaningful domestic manufacturing build-out at more than five years, longer than any phase-in period the administration has allowed on previous tariff rounds.
Lutnick’s Investment-Linked Quota Formula
Commerce Secretary Howard Lutnick is the primary architect of the mechanism at the center of the emerging framework. Under a structure Lutnick favors, tariff-free chip import allowances would be tied proportionally to each company’s committed investment in US semiconductor manufacturing capacity. The more US fabrication capacity a company commits to establishing, the more chips it could import without paying full tariffs.
At their most aggressive, Lutnick and other administration officials have repeatedly said South Korean and Taiwanese companies that decline to invest in the United States could face 100% chip tariffs. The 100% rate is a stated negotiating position rather than an enacted policy, but it frames the upper bound of what the administration is willing to threaten. The White House defended the approach, telling Politico that “Reshoring semiconductor manufacturing is a top priority for President Trump,” while adding that unannounced tariff plans should be treated as speculation.
Why the Formula Cannot Cover Hyperscaler Demand
The investment-linked quota structure has a quantifiable problem that industry representatives have raised directly with Lutnick and Bureau of Industry and Security undersecretary Jeffrey Kessler. The duty-free import volumes the formula would generate are structurally insufficient to cover what American cloud companies are procuring during the current record AI spending run.
Taiwan Semiconductor Manufacturing Company controls 73% of the foundry market and essentially 100% of the leading-edge logic chip production that powers Nvidia’s AI accelerators and the custom silicon chips designed by Amazon, Google, and Microsoft for their own data centers. Taiwan produces more than 90% of the world’s most advanced chips. TSMC has committed $265 billion to Arizona, the largest foreign direct investment in US history, yet projects only about 30% of its most advanced fabrication capacity will be located there at full build-out. Under the existing Taiwan trade agreement, TSMC can import 2.5 times its current US manufacturing capacity duty-free while plants are under construction, tightening to 1.5 times once facilities are operational. With only 30% of TSMC’s most advanced capacity projected to be in Arizona, the quota math leaves the majority of what hyperscalers need subject to potential tariff exposure.
Servers, Laptops, and Consoles Pulled Into Scope
The proposed Phase 2 framework would extend duties beyond standalone chips to the finished products built with them. Servers, laptops, and gaming consoles assembled using Phase 1-taxed accelerators and logic chips would face new tariff exposure under the deliberations described to Politico. That expansion would mark a significant departure from the original proclamation, which targeted discrete semiconductor imports rather than downstream hardware categories.
For hyperscalers building new data centers, the downstream extension matters as much as the chip-level tariff itself. A 25% duty on imported server racks would translate directly into higher capital expenditure per megawatt of compute capacity, and the largest cloud providers have already disclosed record 2026 capex run rates tied to AI training and inference expansion.
Industry Reaction and Stakes for the AI Buildout
Jonathan McHale, digital policy chief at the Computer and Communications Industry Association, whose membership includes Amazon, Google, and Meta, offered a pointed characterization of the stakes. He compared the current AI data center buildout to past infrastructure booms and warned that tariff escalation could distort investment decisions in ways that slow rather than accelerate domestic capacity. Industry representatives argued in private talks that a quota formula keyed to current domestic US capacity simply cannot cover the volumes that hyperscalers are purchasing right now.
For Trump chip tariff Phase 2 watchers, the practical question is not whether exemptions are politically defensible but whether the proposed investment-linked quota math can absorb the volumes American cloud buyers are committing to today. With TSMC’s most advanced output still concentrated in Taiwan and hyperscaler procurement accelerating, the gap between domestic capacity and procurement demand is widening rather than narrowing, and Phase 2’s structural choices will determine whether that gap is bridged by tariffs, by investment commitments, or by a combination that the industry argues is mathematically impossible under current timelines.
Source: TechTimes https://www.techtimes.com/articles/trump-chip-tariff-phase-2-data-center-carve-out