Venture funds are not, historically, designed to have geopolitical theses. The typical VC framework treats market structure and regulatory environment as external variables to manage around — you invest in companies with durable competitive advantages, and you hope the macro environment doesn't go entirely sideways. Deep tech investing in semiconductors no longer has that luxury. The policy environment is now a first-order input into investment decisions, not a background condition.
The 2022 CHIPS and Science Act committed $52.7 billion to US semiconductor manufacturing and research. The concurrent BIS export control rules restricting advanced semiconductor equipment and chip sales to China represented the most significant technology decoupling policy in decades. The EU's European Chips Act pledged EUR 43 billion to build European semiconductor manufacturing capacity. Japan, South Korea, and Taiwan all announced domestic semiconductor investment programs with government incentive structures that would have been unthinkable five years earlier.
This is not normal. The semiconductor industry spent four decades optimizing around comparative advantage — designing in the US, fabricating in Taiwan and South Korea, packaging in Southeast Asia, selling everywhere. That model delivered extraordinary economic efficiency. It also created strategic dependencies that every major government has decided are unacceptable to maintain.
What the Supply Chain Concentration Actually Looks Like
TSMC's share of leading-edge semiconductor manufacturing (sub-5nm) is roughly 90%. A single fab complex in Hsinchu, Taiwan, accounts for a disproportionate fraction of global advanced logic production. Taiwan Semiconductor Manufacturing Company processes chips for Apple, NVIDIA, Qualcomm, AMD, and effectively every major fabless chip designer in the world. There is no near-term alternative. TSMC's process technology lead, accumulated over decades of relentless investment and learning curve advancement, is not something that can be replicated quickly regardless of capital availability.
The supply chain concentration in semiconductor materials and equipment is, if anything, more acute. ASML has a near-monopoly on EUV lithography tools — without which leading-edge nodes cannot be manufactured. ASML's equipment contains components sourced from over 5,000 suppliers across multiple countries. It cannot be redesigned around export restrictions in any short timeframe. Applied Materials, Lam Research, and KLA together supply the majority of deposition, etch, and inspection equipment for advanced fabs globally. All are US companies subject to export control regulation.
The implication for investment: companies that supply into the domestic semiconductor manufacturing buildout — in materials, equipment, process chemistry, or specialized tooling — have addressable markets that are growing at rates driven by policy commitment, not just commercial demand cycles.
The CHIPS Act as Customer Creation
The CHIPS Act's manufacturing incentives are not evenly distributed across semiconductor categories. The bulk of investment targets leading-edge logic and DRAM manufacturing — TSMC's Arizona fabs, Samsung's Texas expansion, Micron's domestic DRAM investment. But the research and workforce development portions fund a broader ecosystem, including the National Semiconductor Technology Center (NSTC) and the National Advanced Packaging Manufacturing Program (NAPMP).
For early-stage deep tech companies, the NSTC is particularly relevant. It's designed to create a public-private research facility for pre-competitive semiconductor process development — essentially a national lab for semiconductor research that reduces the commercialization barrier for novel process technologies. Companies developing next-generation gate dielectrics, 2D materials integration, or advanced packaging approaches now have a potential institutional customer for process development work that didn't exist three years ago.
We've seen this change the deal flow dynamics in our pipeline directly. Companies that would previously have needed to secure direct contracts with TSMC or Intel Foundry for process qualification are now approaching the NSTC and similar programs as an intermediate commercialization step. That creates a viable revenue bridge between research and fab adoption that changes the funding math significantly for seed and Series A companies.
Export Controls and the Innovation Acceleration
The BIS export controls have had an effect that the policy designers probably anticipated but rarely discuss openly: they've accelerated Chinese domestic semiconductor R&D investment at a scale that rivals the CHIPS Act itself. China's domestic chip fund (the "Big Fund") has committed over $50 billion to domestic semiconductor development. Hundreds of Chinese semiconductor startups have received government backing since 2022.
From a competitive standpoint, this matters for portfolio companies in certain categories. Equipment companies, in particular, face a Chinese domestic market that is now being supplied by government-funded competitors who are not subject to the same financial discipline as privately funded companies. For materials companies, the picture is more nuanced — China's materials industry is strong in some areas (rare earth processing, silicon carbide) and weak in others (advanced photoresists, specialty gases for leading-edge processes).
The investment implication we've reached: portfolio companies should be designed to compete in the technology performance dimension, not the price dimension. Government-backed competition on cost doesn't threaten companies whose products are qualified into leading-edge fab processes for technical reasons — you don't replace a qualified gate oxide precursor with a cheaper alternative because the qualification process itself is a multi-year commitment. You compete on performance and reliability, and you maintain that position through continuous technical advancement.
What Changes for LP Conversations
The geopolitical context has made institutional LP conversations about deep tech significantly more receptive than they were five years ago. Pension funds and sovereign wealth funds now have explicit mandates to consider national security implications of their investment programs. University endowments are reviewing their technology exposure. Family offices with roots in manufacturing industries understand supply chain risk viscerally.
This doesn't mean every LP meeting goes well. Explaining why a 12-year fund life is appropriate for deep tech investing requires the same patience it always did. But the opening premise — that semiconductor and quantum technology investments have strategic importance that justifies patient capital and specialized expertise — no longer requires extensive argument. The front pages have done that work.
We're not building a geopolitically motivated portfolio. We're building a technically motivated portfolio that happens to be operating in an environment where national interests and commercial interests are unusually well aligned. For the companies we back, that alignment means government customers, research partnerships, and policy tailwinds that weren't available to the previous generation of deep tech founders. It makes the category more attractive, not less.
Building semiconductor technology with strategic relevance to domestic supply chain goals? Talk to the Coexin team.
