Despite huge demand for AI chips, TSMC's conservative CapEx strategy, driven by fear of a demand downturn, is creating a critical silicon supply shortage. This is causing AI companies to forego immediate revenue.

In the AI gold rush, don't bet on the "miners" like Google and Meta, who are spending billions on a new, high-risk game. Instead, invest in the "pickaxe makers"—the essential toll bridges like TSMC and ASML that every AI company must pass through, ensuring your investment has a higher probability of success.

TSMC's "pure-play foundry" model, where it only manufactures chips and doesn't design its own, builds deep trust. Customers like Apple and NVIDIA can share sensitive designs without fear of competition, unlike with rivals Intel and Samsung who have their own chip products.

For two decades, traditional venture capital firms largely abandoned capital-intensive semiconductor startups for SaaS models. This created a vacuum filled by corporate VCs (Samsung, ARM) and strategic investors, shaping the current concentrated landscape and creating new opportunities as AI reignites the sector.

NVIDIA's financing and demand guarantees for its chips are not just to spur sales, which are already high. The strategic goal is to reduce customer concentration by helping smaller players and startups build compute capacity, ensuring NVIDIA isn't solely reliant on a few hyperscalers for revenue.

NVIDIA's primary business risk isn't competition, but extreme customer concentration. Its top 4-5 customers represent ~80% of revenue. Each has a multi-billion dollar incentive to develop their own chips to reclaim NVIDIA's high gross margins, a threat most businesses don't face.

Conventional venture capital wisdom of 'winner-take-all' may not apply to AI applications. The market is expanding so rapidly that it can sustain multiple, fast-growing, highly valuable companies, each capturing a significant niche. For VCs, this means huge returns don't necessarily require backing a monopoly.

The vast majority of spending and market capitalization in AI today is in the infrastructure layer—compute (NVIDIA), foundation models (OpenAI), and data services. The entire application layer's revenue combined is a rounding error in comparison, highlighting a massive, though likely temporary, imbalance in where value is currently being captured.

While energy is a concern, the highly consolidated semiconductor supply chain, with TSMC controlling 90% of advanced nodes and relying on a single EUV machine supplier (ASML), creates a more immediate and inelastic bottleneck for AI hardware expansion than energy production.