Fabless
by Daniel Nenni and Paul McLellan · 2013
The industry-insider account of how splitting chip design from chip manufacturing remade semiconductors -- and made Moore's Law survivable.
Worth reading? Fabless explains a decision that reshaped the entire chip industry: separating the company that designs a chip from the company that manufactures it. Compared to Chip War's sweeping geopolitical narrative, this is a narrower, more technical business history, written by two longtime industry insiders (one of them ran SemiWiki) rather than journalists. It reads more like an industry trade report than a mass-market narrative, and several chapters are contributed by executives from companies telling their own corporate history, so expect some self-promotion mixed in with the analysis. Read it if you want to understand why "fabless" became the dominant model and how TSMC's foundry business grew out of that split. Skip it if you want a current picture of the industry -- treat this as history, not a 2026 snapshot.
| Full Title | Fabless: The Transformation of the Semiconductor Industry |
|---|---|
| Author | Daniel Nenni and Paul McLellan |
| Published | 2013 |
| Category | Business & Money |
The Verdict
Nenni and McLellan aren’t journalists explaining the chip industry from outside, they built careers inside it, and it shows: this is the trade history of why “fabless” became the default model, not a breezy narrative for general readers.
you want the business-model story behind TSMC and the foundry system, told by people who worked inside it
you want a current, up-to-date picture of the industry -- this is a 2013 snapshot, and names like Nvidia's AI dominance and Apple Silicon postdate it
Book Summary
Before the 1980s, a chip company designed and manufactured its own chips in-house. Splitting those two functions, design (fabless) and manufacturing (foundry), let smaller companies compete on design without the multi-billion-dollar cost of building a fab, and let manufacturers like TSMC specialize purely in production.
The foundry model is what made Moore's Law affordable to sustain. As fabs got more expensive to build with each generation, only a few companies could keep investing in bleeding-edge manufacturing, while dozens of design-only firms could still innovate without that capital burden.
Electronic design automation (EDA) software was the quiet enabler of the fabless split. Without tools that let designers hand off a chip blueprint a foundry could manufacture without further collaboration, the fabless model wouldn't have worked at the scale it did.
Top 7 Lessons from Fabless
- Splitting chip design from chip manufacturing let smaller companies compete without funding a fab.
- TSMC's foundry model exists because someone had to specialize purely in manufacturing other companies' designs.
- Fab costs rise so steeply each generation that only a handful of companies can still afford to build them.
- EDA software was the unglamorous enabler that made the design/manufacturing handoff actually work.
- The fabless split let venture-funded startups enter chip design without needing hundreds of millions in fab capital.
- Corporate case studies inside the book show the transition wasn't uniform -- some companies split reluctantly, others were built fabless from day one.
- The industry's move to fabless mirrors a broader pattern: specialization beats vertical integration once a process gets expensive enough.
Frequently Asked Questions
Is Fabless worth reading?
If you want the inside-industry account of why chip design and chip manufacturing split into separate companies, yes. It's more technical trade history than mass-market narrative.
What is the fabless business model?
A fabless company designs chips but doesn't manufacture them -- it pays a foundry (like TSMC) to fabricate the actual silicon, letting the design company avoid the enormous capital cost of building and running a fab.
Is Fabless still relevant given how old it is?
The core business-model history holds up, but it's a 2013 snapshot -- it predates the AI chip boom and won't cover recent players like Nvidia's current dominance or Apple Silicon.
Ready to read it?
Get Fabless on Amazon
