CHIPS Act & US Reshoring Stocks 2026: From Groundbreaking Ceremonies to Operational Fabs
June 20, 2026 · 13 min read
The $52.7 billion CHIPS and Science Act is no longer just a policy announcement — it is a construction site. TSMC Arizona Fab 21 Phase 1 is operational and producing Apple M-series and NVIDIA Blackwell chips. Micron has secured $6.14B in grants for domestic HBM4 memory. Samsung is expanding in Texas. Intel is fighting for its foundry future with the 18A node. Here is the complete investor guide to who benefits, who risks, and what to buy.
CHIPS Act at a Glance 2026 — Key Metrics
Total CHIPS Act Funding
$52.7B
Manufacturing incentives, R&D, workforce
Awards Finalized
~$33.7B
Across 20+ recipients as of mid-2026
TSMC Arizona Status
Phase 1 Operational
4nm/3nm — Apple M4, NVDA Blackwell
Intel Ohio Status
Delayed to ~2030
$20B investment, $8.5B CHIPS grant
Micron Grant
$6.14B
Plus $7.5B loans — HBM4 for AI by 2028
Samsung Texas Fab
$6.4B Grant
4nm expansion, Taylor TX site
Construction Jobs
~50,000+
Direct construction jobs created by fab buildout
US Chip Market Share
37% (1990) → 12% (2022)
The decline that triggered the CHIPS Act
Why the CHIPS Act Exists — The Strategic Case
The United States invented the semiconductor. Bell Labs, Intel, Texas Instruments — the transistor, the integrated circuit, DRAM — all American innovations. Yet by 2022, the US produced only 12% of global semiconductor output, down from 37% in 1990. The CHIPS Act is Washington's answer to a three-decade strategic retreat.
The Taiwan Concentration Problem
Taiwan Semiconductor Manufacturing Company (TSMC) and its Taiwan-based competitors produce more than 90% of the world's most advanced logic chips — those at 7nm and below. The chips inside every Apple iPhone, every NVIDIA AI GPU, every AMD data center processor come from a 35,000-square-kilometer island that sits 160 kilometers from mainland China.
A Taiwan Strait conflict or blockade — even a brief one — would halt advanced chip supply to every US technology company simultaneously
There is no near-term substitute: building a leading-edge fab takes 3–5 years from groundbreaking to production, and the equipment lead times alone can stretch 18 months
The US government concluded that geographic concentration of this magnitude in a geopolitically contested region constitutes a national security vulnerability — not merely a supply chain risk
The COVID Supply Chain Lesson
The 2020–2022 global semiconductor shortage — triggered by COVID-19 demand shifts — caused an estimated $500B in lost automotive production worldwide. The US auto industry alone saw plants idled for months because of a shortage of $1 microcontrollers. This was a mature-node shortage, not even a leading-edge one. The CHIPS Act is partly a response: ensure that the US has domestic production across the full process node spectrum.
The AI Arms Race Dimension
The AI compute buildout has added a new urgency to domestic chip production that did not exist when the CHIPS Act was being drafted in 2021. Every NVIDIA GPU, every AI training cluster, every inference chip is a national asset in what policymakers describe as an AI arms race with China. The US government has imposed sweeping export controls on advanced AI chips to China — but those controls only matter if the US can actually produce those chips domestically at scale.
CHIPS Act Funding Breakdown — Where the $52.7B Goes
$39B — Manufacturing Incentives
Direct grants and loans to semiconductor manufacturers building or expanding fabs in the US. This is the largest bucket — and the one TSMC, Intel, Micron, Samsung, and GlobalFoundries are drawing from.
$11B — R&D (NSTC)
National Semiconductor Technology Center — a public-private research hub tasked with keeping US at the frontier of chip R&D. Separate from the fab grants; meant to ensure the US doesn't just build fabs but also invents the next generation of chip technology.
$2B — Legacy Chips for Defense & Auto
Dedicated funding for mature process node production — the 28nm to 65nm chips used in military systems, automobiles, industrial equipment, and IoT. The 2021 auto shortage exposed the strategic risk of offshore mature-node dependence.
$500M — International Partnerships
Coordination with allied nations (Japan, Netherlands, South Korea, India) on semiconductor supply chain resilience. Includes funding for joint R&D and workforce development programs.
Key Guardrail — No China Expansion for 10 Years
All CHIPS Act grant recipients are prohibited from expanding advanced chip manufacturing capacity in China for 10 years. This is the most significant geopolitical condition attached to the funding — it forces a structural decoupling of US-funded semiconductor capacity from the Chinese market.
Major Fab Investments — Status & Timeline
Here is the current status of the five most important CHIPS Act fab projects as of June 2026:
TSMC ArizonaTSM
Phase 1: Operational4nm / 3nm
Investment: $65B totalCHIPS Grant: $6.6B
Fab 21 Phase 1 went live Q1 2025 — first Apple M-series and NVIDIA Blackwell B200 chips produced outside Taiwan. Phase 2 (2nm) equipment move-in began Q3 2026. Phase 3 (1.6nm A16) announced for late 2020s.
Intel OhioINTC
Delayed — ~20302nm (Intel 20A/18A)
Investment: $20BCHIPS Grant: $8.5B
"One8" campus in New Albany OH was originally slated for 2025. Delayed roughly 5 years due to capital constraints and workforce availability. Intel 18A process targeting risk production in 2026 from Arizona Ocotillo campus.
Clay, NY fab is the largest semiconductor investment in US history when complete. Boise, ID expansion adds HBM4 capacity. HBM4 targeting AI accelerator supply from US soil by 2028.
Samsung Austin TXSSNLF
4nm Expansion4nm
Investment: $17B+CHIPS Grant: $6.4B
Taylor, TX fab expanding to 4nm production. Samsung's US presence largely serves automotive and mobile customers. Less advanced-node exposure than TSMC or Intel but diversifies US fab geography.
GlobalFoundries Malta NYGFS
Mature Node Expansion22nm–65nm (mature)
Investment: $12B+CHIPS Grant: $1.5B
Mature nodes for defense, automotive, and IoT. Sole-source supplier to the US Department of Defense for several chip types. Less glamorous than leading-edge but arguably more strategically critical for near-term national security.
Stock Comparison — CHIPS Act Beneficiaries
How do the major CHIPS Act semiconductor plays compare as investments? Here is a side-by-side view:
Equipment & Materials Plays — Win Regardless of Which Fab Wins
The most consistent way to invest in the CHIPS Act buildout is through semiconductor equipment and materials companies. These businesses supply the tools that fill every fab — regardless of whether TSMC, Intel, or Samsung ultimately captures the most market share. Think of them as the "picks and shovels" of the semiconductor gold rush.
Smaller cap — pure-play domestic fab construction beneficiary
Applied Materials alone estimates approximately $4B in incremental US revenue through 2028 from the domestic fab construction cycle. KLA and Lam Research have similarly sized exposure. These companies benefit from both the CHIPS Act fab buildout and the ongoing global semiconductor capacity expansion driven by AI demand.
Construction & Infrastructure Plays — Building the Fabs
Semiconductor fabs are among the most complex and expensive construction projects in the world. A leading-edge fab costs $15–30B and requires years of specialized construction. The companies managing and executing this construction represent a less-obvious but real investment angle.
JJacobs Engineering
Prime contractor and program manager on multiple CHIPS Act fabs. Multi-year, multi-billion dollar contracts for engineering, procurement, and construction management.
ACMAECOM
Civil and infrastructure engineering for semiconductor campuses. Water treatment, site prep, utility infrastructure — the unglamorous but essential work.
PWRQuanta Services
Electrical infrastructure and specialty construction. Fabs require enormous power delivery systems — Quanta is a leading contractor for this specialized work.
These companies benefit from multi-year construction contracts that provide revenue visibility. The fab buildout cycle extends into the early 2030s — the timeline for Intel Ohio alone represents a decade of construction activity.
Intel Deep Dive — The Comeback Story
No company has more riding on the CHIPS Act than Intel. Once the undisputed leader in semiconductor manufacturing, Intel ceded its process leadership to TSMC over a decade of execution stumbles. The CHIPS Act represents Intel's chance — and the US government's bet — on a domestic advanced foundry.
IDM 2.0 Strategy
Intel's "IDM 2.0" (Integrated Device Manufacturer 2.0) is a restructuring of the company's identity. Rather than being solely a chip designer that manufactures its own products, Intel is building Intel Foundry Services (IFS) — a merchant foundry that will manufacture chips for external customers. This is Intel directly competing with TSMC for Apple, NVIDIA, AMD, and other fabless chip designers.
Intel 18A (1.8nm-class) process: if successful, this would be the most advanced process node in volume production anywhere in the world. The "18A" name reflects angstrom-scale transistor dimensions — a marketing rebranding of nanometers that signals sub-2nm capability
Microsoft as early IFS customer: Microsoft has reportedly signed on as an Intel Foundry Services customer for 18A, providing the critical external customer validation that the process is competitive
"Gate-all-around" transistors: Intel is among the first to deploy this new transistor architecture in volume production — the same architecture Samsung is using for its SF3 process
Backside power delivery: Intel 18A includes power delivery from the backside of the wafer, a technical innovation that improves power efficiency and performance density
Ohio: Delayed but Not Dead
Intel's New Albany, Ohio "Silicon Heartland" campus — originally promised for 2025 — has been pushed to approximately 2030. This delay reflects both Intel's capital constraints and an honest assessment of workforce availability in central Ohio. The US simply does not have 50,000 semiconductor manufacturing engineers sitting idle. Building the workforce is a prerequisite to building the fab.
The $8.5B CHIPS grant to Intel for Ohio is the largest single CHIPS Act award. The government's willingness to maintain that commitment despite delays signals that Washington views domestic Intel foundry capacity as strategically essential — not just economically desirable.
Why the CHIPS Grant Matters for Intel's Survival
Intel carries a significant debt load and is investing heavily in both product development (client and data center chips) and foundry infrastructure simultaneously. The CHIPS grants meaningfully reduce the capital burden of the Ohio and Arizona fab investments. Without them, Intel would face a much harder capital allocation decision between foundry investment and its core chip design business.
TSMC Arizona vs. Taiwan — The Cost-Security Tradeoff
TSMC Arizona Fab 21 Phase 1 is now operational and producing Apple and NVIDIA chips — a genuine milestone. But producing chips in Arizona costs more than producing them in Taiwan, and understanding why matters for assessing the long-term economics.
Arizona Cost Premium
US production costs estimated 30–50% higher than equivalent Taiwan production
Higher labor costs, higher construction costs, higher regulatory compliance costs
Energy costs higher in Arizona than Taiwan (subsidized industrial power)
Supply chain for specialty materials and chemicals less developed in US
Partially offset by CHIPS Act grants and 25% investment tax credit
Strategic Value
Supply chain security: US-produced chips cannot be cut off by Taiwan Strait disruption
US government and defense customers require domestic sourcing for sensitive applications
Apple and NVIDIA gain geographic diversification of their most critical supply
Longer-term: as Arizona supply chain matures, cost premium should compress
TSMC's $65B Arizona commitment signals multi-decade strategic intent, not a one-off
The Talent Challenge
TSMC imported hundreds of Taiwanese engineers to staff Fab 21 Phase 1 — a pragmatic solution to a structural problem. The US does not currently have enough semiconductor process engineers, technicians, and operators to staff the fabs being built. TSMC's long-term plan for Phases 2 and 3 depends on training American workers — a 5–10 year workforce development challenge that universities and community colleges are only beginning to address.
Beyond Leading-Edge — Mature Node & Defense Chips
The semiconductor conversation fixates on leading-edge nodes (3nm, 2nm, 18A) because those chips power iPhones and AI accelerators. But the 2021 automotive shortage was caused by a shortage of 28nm to 65nm chips — mature nodes that are less glamorous but arguably more strategically important for the US economy near-term.
GlobalFoundries — The Defense and Auto Champion
GlobalFoundries (GFS) operates the only US-based advanced mature-node fab at scale. Its Malta, New York facility is a sole-source supplier for several US Department of Defense applications — meaning there is no substitute if GFS production were disrupted. The $1.5B CHIPS Act grant supports expansion of this capacity.
SkyWater Technology — The Pure-Play Defense Foundry
SkyWater Technology (SKYT) operates a CHIPS-funded fab in Minnesota focused on government and defense applications. For national security applications requiring US-soil production, SkyWater is often the only option. Small cap, niche, but uniquely positioned for defense spending tailwinds.
The Auto Industry Lesson
The 2021 auto shortage — which caused an estimated $210B in lost revenue for the US auto industry — happened because automakers had offshored their chip supply to Asian foundries and had no backup supply. The CHIPS Act's $2B mature-node allocation is a direct policy response: ensure that the chips that go into US-built cars can be manufactured in the US.
Key Risks — What Could Go Wrong
Cost OverrunsHigh Risk
Leading-edge fab construction is notoriously prone to cost escalation. TSMC's Arizona costs have already run higher than initial estimates. Intel's Ohio delays reflect partially cost-related capital allocation decisions. Every $1B cost overrun reduces the economics of domestic production.
Talent Gap — 115,000 Workers NeededHigh Risk
The US semiconductor industry needs approximately 115,000 additional skilled workers by 2030 to staff the fabs being built. Universities are expanding semiconductor programs but the pipeline takes years. The near-term solution — importing talent — creates its own regulatory and logistical friction.
Timeline SlippageMedium Risk
Intel Ohio is the most visible example: delayed from 2025 to approximately 2030. Fab construction timelines consistently slip. Investors should assume that any announced timeline has a 12–24 month buffer priced in.
China Retaliation RiskMedium Risk
China is the world's largest semiconductor market. CHIPS Act guardrails prevent US-funded companies from expanding China capacity, but China could retaliate with export restrictions on materials (gallium, germanium) or market access restrictions for US chip companies.
US export controls on advanced chips and equipment to China continue to tighten. If restrictions expand to include additional equipment categories, it could accelerate China's push for semiconductor self-sufficiency — potentially creating a long-term competitor rather than a dependent customer.
Intel Foundry Execution RiskHigh Risk
Intel Foundry Services has not yet won a major external customer at volume. If 18A yields or performance disappoint, TSMC continues to dominate the foundry market and Intel's enormous capital investment in the foundry strategy does not pay off.
Bull Case — Why This Reshoring Cycle Is Different
Strategic necessity overrides economics: the US government has decided that domestic semiconductor manufacturing is a national security requirement, not a market optimization problem. This means policy support is durable across political cycles — the CHIPS Act had bipartisan support and has survived administration changes
AI demand is pulling capacity forward: the AI compute buildout is creating semiconductor demand that was not modeled when the CHIPS Act was written. TSMC Arizona Phase 1 chips are going directly into NVIDIA's most advanced AI GPUs — the AI arms race accelerates the economic case for US production
TSMC operational now reduces execution risk: Fab 21 Phase 1 running is qualitatively different from a groundbreaking ceremony. The hardest part — proving you can actually produce leading-edge chips outside Taiwan — is done
Equipment cycle is multi-year and locked in: Applied Materials, Lam Research, and KLA have order books that extend 3–5 years. Even if a fab is delayed, equipment orders have already been placed. The equipment cycle does not care about construction timelines
Workforce programs are scaling: community colleges in Arizona, New York, and Ohio have launched semiconductor technician programs specifically in response to fab announcements. The workforce gap is real but is being addressed with serious institutional investment
Bear Case — Why the Economics Are Hard
US production costs 30–50% higher than Asia: without policy support, the economics of US semiconductor manufacturing do not close. The CHIPS Act grants are one-time — once deployed, companies must sustain operations at structurally higher costs than Asian competitors
Intel's foundry struggles are real: Intel Foundry Services has not won major external customers at volume. If 18A does not achieve competitive yield and performance, the entire US bet on a domestic advanced foundry outside of TSMC fails — and Intel's debt load becomes a serious concern
10-year timeline before meaningful US market share shift: even optimistically, US semiconductor manufacturing market share does not materially shift until the early 2030s when Ohio and other fabs come online. Investors are pricing in outcomes that are 5–8 years away
China market risk: US chip companies (NVDA, AMAT, INTC, MU) depend on China for 20–30% of revenue. Escalating trade restrictions cut both ways — reducing China's access to US chips also reduces US chip company revenue from China
Geopolitical resolution risk: if US-China tensions de-escalate — trade deal, diplomatic breakthrough — the urgency of domestic chip production reduces. The stocks most dependent on geopolitical risk premium would re-rate downward
Bottom Line Verdict & Investment Approach
The CHIPS Act is the most significant US industrial policy since the space race. The question for investors is not whether reshoring is happening — it clearly is — but how to position around a complex, long-duration, high-stakes industrial transition.
Recommended Framework
Tier 1 — Lowest RiskAMAT, LRCX, KLAC
Equipment companies benefit from every fab being built — by TSMC, Intel, Micron, and Samsung — regardless of who wins the foundry competition. Revenue visibility from multi-year order backlogs. Benefit from both CHIPS Act buildout and global AI-driven semiconductor capacity expansion.
Tier 2 — Medium RiskTSM, MU
TSMC (TSM ADR) has proven it can operate in Arizona and is the highest-quality semiconductor business in the world. The US production premium is real but manageable. Micron (MU) has locked-in government funding, a strategic domestic memory position, and AI HBM tailwind. Both carry sector cyclicality risk.
Intel is a turnaround bet on 18A and IFS — binary outcome potential. GlobalFoundries and SkyWater are defense/auto pure-plays with less glamour but genuine strategic moats. Higher individual company risk, but differentiated from the broader semiconductor cycle.
Tier 4 — Construction CycleJ, PWR, AECOM
Multi-year fab construction contracts provide revenue visibility into the early 2030s. Less semiconductor-specific risk. Benefit from the broader US infrastructure buildout (data centers, power grid, EV manufacturing) as well.
The single most important insight: equipment plays (AMAT, LRCX, KLAC) are the lower-risk way to invest in the CHIPS Act buildout. They get paid to fill the fabs regardless of which company's chips end up winning the AI compute competition. For investors who want direct exposure to the reshoring thesis without taking a view on Intel's foundry execution or TSMC's US cost structure, equipment is the cleanest expression of the trade.
Explore CHIPS Act Stocks
Dig into our analysis tools for the key CHIPS Act plays: