Intel (INTC): The $100 billion moonshot for western chip independence

While investors chase Nvidia’s AI dominance and TSMC’s manufacturing perfection, one company is making the most audacious industrial bet in modern technology: spending over $100 billion to build something neither competitor can offer — leading-edge semiconductor manufacturing on Western soil. Intel isn’t just trying to catch up. It’s attempting to convert a fading CPU monopolist into the foundry the free world needs, backed by tens of billions in government subsidies and the uncomfortable reality that the vast majority of the world’s most advanced chips are made on an island 100 miles from mainland China.

The numbers tell two stories simultaneously. On 2023 figures, Intel’s foundry division lost roughly $1.60 for every dollar of revenue — and management expects 2024 to be the peak loss year. Free cash flow has been running in the negative low-teens billions annually. It cut 15,000 jobs. And yet — governments are writing billion-dollar cheques to keep this bet alive, AWS signed on as a customer, and Intel was the first chipmaker to take delivery of next-generation lithography tools that competitors won’t receive until later in the decade. This is either the most important semiconductor investment of the decade or the most expensive value trap in tech history. The next 18 months will tell us which.

---

---

🔔 Don’t miss out!
Add winvestacrisps@substack.com to your email list so our updates never land in spam.

---

🌍 The geopolitical case: Why governments are bankrolling Intel

Strip away the financials for a moment and ask a simpler question: what happens if Taiwan faces a blockade?

TSMC manufactures the vast majority of the world’s most advanced logic chips — by some estimates, around 90% of leading-edge capacity. A disruption — military, natural, or political — would cripple everything from iPhones to fighter jets to data centres overnight. Intel is the only Western-headquartered company with a credible roadmap to leading-edge manufacturing. That single fact explains why the U.S. government committed $280 billion through the CHIPS and Science Act (covering both semiconductor incentives and broader R&D funding), why Europe pledged €43 billion in semiconductor subsidies, and why Intel specifically has secured up to $7.9 billion in direct U.S. CHIPS Act grants, roughly $11 billion in federal loans, a 25% investment tax credit, and around €10–11 billion in German subsidies for its Magdeburg fab — tens of billions in combined government support that materially socialises the buildout risk.

This isn’t corporate welfare — it’s national security spending routed through a publicly traded company. Intel is building new fabrication facilities in Arizona, Ohio, New Mexico, and Germany, effectively socialising a significant portion of its buildout risk. For investors, this creates a geopolitical option embedded in the stock price: if U.S.-China tensions escalate or Taiwan faces military pressure, Intel’s domestic manufacturing becomes strategically priceless regardless of its cost structure.

No other company in the Western world can offer this. GlobalFoundries abandoned leading-edge nodes years ago. Samsung’s most advanced fabs sit in South Korea — itself within range of North Korean artillery. Intel’s bet is that semiconductor sovereignty isn’t optional, and that the company willing to build the fabs will be rewarded with decades of demand, whether or not it matches TSMC transistor-for-transistor.

---

Want to add INTC to your portfolio? Download the Winvesta app and start investing in Intel and 1,000+ US stocks — directly from India, in minutes.

🏭 The foundry gamble: $7 billion in losses, and that’s the plan

Intel Foundry Services (IFS) is the centrepiece of this transformation — and the source of its most alarming headline numbers.

The brutal math so far

In 2023, Intel Foundry posted roughly $7 billion in operating losses on low-single-digit billions of revenue — meaning the unit lost approximately $1.60 for every dollar it earned. Management has signalled that 2024 will be the peak loss year before a gradual path toward breakeven around 2027. The segment operates at a capital intensity of 60–70% of revenue versus 20–25% for fabless chip designers who outsource manufacturing to TSMC. Intel is absorbing startup costs, depreciation on brand-new fabrication lines, and the learning curve of courting external customers who’ve spent decades optimising designs for a competitor’s process.

Why this might be rational

Foundries are the ultimate fixed-cost business. Once the fabs are built and yields improve, incremental revenue falls almost entirely to the bottom line. TSMC’s operating margins exceed 40% precisely because its factories run near capacity. Intel is in the capital-intensive trough—spending significant sums before revenue arrives. The tens of billions in combined government grants, loans, and tax credits effectively cover a meaningful share of the buildout, meaning Intel’s actual at-risk capital is lower than the headline numbers suggest.

The customer pipeline is thin but real

Key announced customers include Amazon Web Services (which is designing custom chips for Intel’s process), the U.S. Department of Defence, and select automotive chipmakers. The elephant in the room: Intel has yet to win tier-one customers like Apple, Nvidia, Qualcomm, or AMD. Winning them requires not just competitive technology but trust — and Intel must rebuild that after years of delays. Most fabless companies have optimised designs around TSMC’s process design kits; switching costs are enormous.

Intel’s process roadmap targets five nodes in four years, with Intel 18A (comparable to TSMC’s 3nm) expected in the second half of 2025. If 18A ships on time with competitive yields, the customer pipeline could accelerate rapidly. If it doesn’t, the $100 billion programme risks becoming a stranded asset.

For investors, IFS is a call option: worthless if Intel can’t execute, transformative if it can.