Former Intel executive on how Intel achieved dominance in the computer industry after pivoting from memory chips to CPUs and then lost it as smartphones emerged
Michael Bruck / LinkedIn : Tweets: @chiakokhua , @_msw_ , @jonmasters , @architect_pc , and @fpiednoel Tweets: RetiredEngineer / @chiakokhua : Interesting article by a former Intel executive, advocating for Intel to go fab-less. If you asked me as recent as 1 year ago, I'd against it. FWIW, I have changed my mind; I think Intel should spin off TMG. The question is, how to make it work? https://www.linkedin.com/... Matthew S. Wilson / @_msw_ : @jonmasters I personally I think that “control your own destiny” is a potential trap. You need to consider economies of scale and ecosystem effects. Those are not going to be 💯% in your control. And if you think you have control, you might miss important shifts. It's the ultimate “lock-in”. @jonmasters : The article misses a few obvious things a number of us predicted years ago: 1. Integrated manufacturing is done. It's now an albatross. The path is to standardize design rules and follow an outsource foundry model. 2. The industry is verticalizing. With commodity CPUs no less. @architect_pc : The real problem was “Murthy's law : “Every 18 Months, I delay a process technology” https://twitter.com/... https://twitter.com/... @fpiednoel : Sadly makes sense. Intel need sharks again, not managers and MBAs. https://twitter.com/...
Context & Ripple Effects
Michael Bruck, a former Intel executive, argues the company should go fab-less and spin off TMG, its manufacturing arm — a direct answer to Intel's own deliberations days earlier, when Bloomberg reported that considering outsourced manufacturing marked the end of an era of US-dominated chipmaking. The debate in his replies frames the stakes: Matthew S. Wilson warns that "control your own destiny" can be a trap once economies of scale and ecosystem effects favor the foundries.
Bruck's case leans on Intel's own history: the company's dominance came from pivoting out of memory into CPUs, and its decline began when it missed smartphones — the moment traced to around 2008, when Intel abandoned low-power phone chipsets to refocus on Atom and netbooks. The subsequent years tested both paths: Pat Gelsinger bet the company on becoming a foundry for chip developers, while by 2025 a former CEO argued against a breakup on the grounds that Intel's foundry could still compete with TSMC.
First-order effects
- Bruck puts TMG's spin-off on the table at the exact moment Intel leadership is weighing outsourcing, turning an internal manufacturing decision into a public strategy fight between integrated fabs and the fabless model.
- Jon Masters' counter-position in the thread — that integrated manufacturing is obsolete and design rules should be standardized for foundry outsourcing — gives the pro-fabless camp a concrete mechanism, not just a sentiment.
Second-order effects
- If Intel went fab-less, TSMC and other foundries would absorb its leading-edge volume, accelerating the scale advantage Wilson points to — and making any future return to integrated manufacturing harder to finance.
- A spun-off TMG would have to win external customers to justify itself, competing head-to-head with the very foundries Intel would rely on.
Third-order effects
- The thread anticipates the industry's central structural question of the decade: whether chip design and leading-edge manufacturing stay under one roof or split permanently — a split Gelsinger later tried to reverse from the opposite direction and struggled to execute.
- Whichever way Intel resolves it sets the template for every remaining integrated device manufacturer deciding between "control your own destiny" and foundry economics.
The trend: Leading-edge chipmaking is splitting from chip design, with Intel's fab-or-fabless choice as the decisive data point in whether integrated manufacturers survive at all.