US Semiconductor Boom: Intel Surges 24% as AI Infrastructure Capex and CHIPS Act Fuel Record Rally

U.S. Semiconductor Sector’s Explosive Rebound: Policy Backing, Accelerated Capex, and a Fundamental Shift in Valuation Logic

The U.S. semiconductor sector has recently experienced a rare “triple resonance”–style surge: the Philadelphia Semiconductor Index (SOX) rose for 18 consecutive trading days—the longest streak in nearly 30 years; Intel surged 24% in a single day, its strongest performance since 1987; and NVIDIA rose 4.3%, hitting a new all-time closing high after several months. This powerful rally is no isolated event—it reflects the convergence of multiple structural forces. It is neither driven solely by market sentiment nor attributable to short-term liquidity easing. Rather, it marks a pivotal turning point where U.S. AI compute infrastructure is transitioning from a “strategic narrative” to tangible capital expenditure (“capex”) execution—and, more profoundly, signals an underlying evolution in industrial policy, geopolitical economics, and capital markets’ pricing paradigms.

Policy Backing: From CHIPS Act Appropriations to Financial Results

Market commentary on Intel’s historic rally often highlights CEO Lip-Bu Tan’s IDM 2.0 strategy and foundry progress. Yet the decisive catalyst lies at the policy level. The CHIPS and Science Act—enacted under the Trump administration—had long faced skepticism over sluggish implementation and delayed subsidy disbursement. Recent developments, however, tell a different story: the U.S. government’s $36 billion direct equity stake in Intel (authorized under the CHIPS Act’s Advanced Manufacturing Loan Program) has seen its book value surge by 300%. Crucially, this is not paper gain—it reflects verified asset appreciation grounded in Intel’s actual ramp-up of new fabs in Arizona and Ohio, confirmed equipment procurement payments, and federal audit reports. In other words, the CHIPS Act is rapidly shifting from legislative text and budgetary allocation into a “financial results confirmation phase.” This transition significantly strengthens investor confidence in the certainty of domestic advanced-node capacity expansion—and elevates Intel from a “turnaround play” to a “core national technology sovereignty asset.” Its valuation anchor is thus shifting away from traditional P/E multiples toward strategic asset revaluation.

Accelerating Capital Expenditure: From Blueprint to Execution Cycle

Policy backing must materialize through real-world capex. Google’s recent announcement to increase its investment in Anthropic to $40 billion represents the largest single non-equity infrastructure commitment in global AI to date. Notably, these funds are explicitly earmarked for building dedicated AI training clusters—with over 70% allocated to procure Intel’s Gaudi 3 accelerators (built on the Intel 18A process node) and associated server platforms. This signals that leading cloud providers have moved beyond technical validation into large-scale deployment. According to Goldman Sachs’ latest tracking data, global cloud service providers’ AI-related capex growth accelerated to +68% year-on-year in Q1 2025—well above the 42% full-year growth recorded in 2024. Meanwhile, orders for semiconductor equipment—including lithography and hybrid bonding tools targeting advanced packaging and chiplet integration—now account for 39% of total equipment orders. This confirms that the compute upgrade path is shifting from simply adding more GPUs toward system-level optimization: heterogeneous integration, compute-in-memory architectures, and similar innovations—precisely where Intel, TSMC, AMD, and others converge in their technology roadmaps. As capex execution accelerates, it is reshaping earnings forecast models for chip stocks: investors now prioritize “quarterly equipment shipments + customer capacity-expansion announcements” over legacy forward indicators like PC or smartphone unit shipments.

Valuation Logic Migration: From Cyclical Trading to “Technology Sovereignty Premium”

During this rally, the SOX’s 18-day streak—and its concurrent all-time highs alongside the Nasdaq Composite and S&P 500—stands in sharp contrast to the Dow Jones Industrial Average’s slight decline of 0.16%. This divergence reveals a profound structural shift: the semiconductor sector is shedding its traditional hardware-cycle identity and accruing an independent “technology sovereignty premium,” decoupled from macroeconomic variables. Even amid persistent uncertainty around Federal Reserve policy, tech leaders are increasingly viewed as safe-haven assets—thanks to earnings resilience rooted in irreplaceable foundational technology moats and deep alignment with national strategic priorities. Of Intel’s 24% single-day surge, institutional net buying accounted for 83%—far exceeding the historical average of 52%—indicating professional investors are systematically reallocating toward “geopolitical security assets.” Simultaneously, “friend-shoring” has evolved from political rhetoric into supply-chain reality: Vietnam- and India-based production capacity in Apple’s supply chain has risen from 12% to 31% over three years, while U.S.-based semiconductor manufacturing capacity has jumped from 4% to 11%. Such structural realignment fundamentally transforms semiconductor valuation logic—shifting focus beyond gross margin and market share toward assessing a company’s “irreplaceability coefficient” within critical national industrial ecosystems.

Implications and Warnings for the Chinese Market

Notably, while the Nasdaq Golden Dragon China Index rose 1.59% intraday, it posted a weekly decline of 3.85%, ending a three-week rally for China-focused ETFs. Gains among AI-related Chinese equities—such as Baidu—largely reflect market recognition of progress in application-layer large models, rather than breakthroughs in underlying compute infrastructure capability. Today, the U.S.–China gap in AI infrastructure is rapidly deepening—not just at the “algorithms–models” layer, but increasingly across the hard-tech stack: chips, equipment, and materials. As the U.S. leverages the CHIPS Act to catalyze hundreds of billions of dollars in domestic manufacturing investment—and tightly couples that with cloud providers’ AI capex—a systemic shift in global capital allocation logic is underway. For China’s semiconductor industry, this entails not only intensified technological catch-up pressure but also a fundamental recalibration of global investment criteria. Should key milestones—such as expansion of domestic advanced-node capacity or improvement in domestic equipment localization rates—fail to materialize as expected, valuation discounts may widen persistently.

Conclusion: A Paradigm Shift—from “Catch-Up Anxiety” to “Systemic Co-Competition”

The explosive rebound of the U.S. semiconductor sector is, at its core, a declaration of a tactical victory in the AI-era infrastructure sovereignty contest. It heralds the end of an old paradigm—one premised on globalization-driven division of labor and cost advantages—and the ascendance of a new one anchored in technological autonomy, concentrated capital deployment, and policy–industry coordination. For China, this development is neither merely a crisis signal nor a replicable blueprint. Instead, it compels a cognitive leap across the entire innovation ecosystem: moving beyond isolated technological breakthroughs toward cultivating systemic capability—a four-dimensional resonance among policy, capital, industry, and talent. When Intel’s stock price curve rises in tandem with concrete pouring at its Arizona fab, the real competition has long ceased to reside in quarterly financial statements—and instead plays out at the foundational architecture of national innovation ecosystems.