Silicon & the Implausibility of Profit

by

Right. Intel. A perfectly respectable company, historically speaking. It’s been, shall we say, undergoing a period of…re-evaluation. The stock has doubled recently, which, in the grand scheme of things, is roughly the same as discovering a slightly larger pebble on a beach. A nice pebble, mind you, but still…a pebble. This, naturally, followed a rather pointed downward adjustment in expectations. (One begins to suspect that ‘expectations’ are merely elaborate constructs designed to facilitate disappointment.)

The U.S. government, in a move that suggests either remarkable foresight or a profound misunderstanding of economics, took a 9.9% stake. Followed by Nvidia, who, in a display of either generosity or strategic maneuvering, invested five billion dollars. (Five billion dollars. It’s a number, isn’t it? A large one. It could buy a lot of tea.) This infusion of capital is intended to propel Intel’s foundry business and its fledgling AI endeavors. The new CEO, Lip-Bu Tan, is attempting to streamline operations, reduce costs, and generally make the company less… bureaucratic. (Imagine a bureaucracy trying not to be bureaucratic. The inherent paradox is quite delightful.)

However, the recent pullback following the earnings report suggests a degree of… exuberance. The market, it seems, was anticipating miracles. (Miracles, naturally, being notoriously unreliable as a business strategy.) Revenue declined 4% to $13.7 billion, and the company reported a GAAP loss of $591 million. (Losses, one notes, are simply gains viewed from a different temporal perspective.) They were profitable on an adjusted basis, which is like being mostly alive.

First quarter expectations are… modest. Revenue between $11.7 and $12.7 billion. A “sharp sequential decline,” they call it. (One wonders if the sequence is actually sharp, or merely perceived as such.) Adjusted earnings per share are projected to be…break-even. (The universe, it seems, is determined to maintain a state of equilibrium.)

Now, for those of you hoping to profit from the AI infrastructure boom (and let’s be honest, who isn’t?), there’s a slightly more…reliable option. A competitor to Intel, if you will. Taiwan Semiconductor Manufacturing Corporation, or TSM. (A name that, admittedly, lacks a certain panache.)

A Track Record of… Existence

TSMC isn’t exactly a household name. They don’t produce shiny, branded products. They reside in Taiwan. (Taiwan, a remarkably resilient island nation with a knack for manufacturing things.) However, they are, by most metrics, one of the most valuable companies on Earth, boasting a market capitalization of $1.8 trillion. (A number so large it almost ceases to have meaning.) They are, essentially, the backbone of the global tech industry, manufacturing chips for Apple, Nvidia, AMD, Broadcom, and a host of others. They produce over half of all contract chips, and an estimated 90% of the advanced ones. (Advanced, in this context, meaning smaller and more complicated.)

This gives them a rather significant competitive advantage. And their results reflect that. Revenue rose 25.5% to $33.7 billion in the fourth quarter, with an operating margin of 54%. (54%! One can only dream of such margins. It’s almost…unfair.) They now derive 77% of their revenue from advanced chips (7nm or less). The stock trades at a P/E ratio of 32, which, compared to the S&P 500, is…reasonable. (Reasonable being a relative term, naturally.) Historically, the stock has traded at a discount due to its location and the ever-present geopolitical concerns. (Geopolitics, a fascinatingly complex game played with real lives and economies.) The stock has also risen over 1,000% in the last decade.

Loading widget...

Intel vs. TSMC: A Matter of…Fabrication

Intel operates in two segments: product design and manufacturing. Historically, they manufactured chips solely for themselves. However, they restructured in 2021 to open their foundry to outside customers, including Amazon. (A bold move, one might say. Like opening a very expensive, highly specialized bakery to the general public.) They’re betting heavily on advanced processes like 18A (1.8nm), recently launched. 18A and the upcoming 14A could potentially make Intel a genuine competitor to TSMC. (Potentially. The operative word being ‘potentially.’) However, they’re currently losing billions of dollars a year on the foundry business while building out these advanced processes. (Building, one notes, is an inherently expensive activity.)

Intel’s stock is currently more expensive than TSMC’s, despite flat revenue and GAAP losses. TSMC, on the other hand, expects revenue to grow at a compound annual rate of around 25% through 2029 and is currently keeping over half of its revenue as operating income. (The numbers speak for themselves, really. Though numbers, one must remember, are merely symbols representing quantities.)

Intel’s turnaround story is appealing, undoubtedly. But TSMC is the safer bet. It should continue to outperform the market as long as the AI boom persists. (Which, given the current trajectory, seems…likely. Though predicting the future is, of course, a fool’s errand.)

Read More

2026-01-29 07:33