Texas Instruments (TXN)

Texas Instruments is the global leader in analog + embedded semiconductors — an exceptional-quality franchise: ~80,000 long-lived products, a scale-and-low-cost moat from internal 300mm wafer fabs (a chip ~40% cheaper than on 200mm), sticky customer relationships, and 22 consecutive years of dividend increases. Chip content grows secularly across Industrial, Automotive, and Data centers (75% of the focus). It is cyclical (coming off a trough, recovering) and capital-intensive (manufactures internally), and today carries net debt of ~$9bn from the fab investment cycle.

A quality semiconductor is valued on EV/NOPAT (after-tax operating income, which expenses stock-based compensation and is not distorted by capex timing — unlike free cash flow, still mid-normalization today). It is EV-level: net debt of ~$9bn is subtracted from enterprise value. At ~$259, still-trough-depressed operating income implies ~39× EV/NOPAT (equivalent to ~34× the forward P/E on the recovery) — well above a disciplined exit.

The base scenario normalizes to mid-cycle — neither the trough nor the peak: revenue recovering to ~$25.5bn over five years (cyclical recovery + secular content), and operating margin toward ~43% (higher fab utilization + the end of the capex drag, though below the 50% peak given the higher depreciation of the internal fab model). That yields NOPAT of ~$9.5bn, to which a 22× EV/NOPAT exit multiple is applied — the top of the [18,24] band for a quality analog semiconductor, disciplined down from today's ~34×. The result is ~$239/share → a total return of +1%/year. Not even the favorable scenario (a full recovery + a sustained 25× multiple) clears today's price by a comfortable margin.

No margin of safety: at this price capital is preserved, but it is not bought below its value. At ~$259, after running ~35-64% to near highs, Texas Instruments is a superb business priced for perfection. The verdict is Preserves value: the franchise's quality is not in question — it is among the best in semiconductors — but the price already discounts the cyclical recovery + the free cash flow inflection + a sustained premium multiple. Normalizing to mid-cycle with a disciplined exit multiple, the base case ($239) falls short of the price, and even the favorable scenario barely reaches it. There is no discount: no absent buyers or motivated sellers are identifiable — on the contrary, it is one of the most beloved stocks in the sector. The adverse scenario (the recovery stalls or reverses, China's probe bites, the fabs sit underutilized, the multiple compresses to 17×) has material downside; the favorable one is already in the price. Value discipline says to watch it closely and wait for a cycle trough — not to buy it near highs with the inflection already paid for.

Three things. The free cash flow inflection — the entire thesis depends on capex falling to $2-3bn and free cash flow per share compounding; if capex proves stickier (or Silicon Labs, the ~$7.5bn acquisition closing in 2027, diverts capital), the inflection dilutes. Cyclicality + margin — revenue is coming off a trough; a relapse in the semiconductor cycle or a stall in the gross margin recovery (57% → toward the low 60s) breaks the base case; extrapolating today's recovery is the risk. And China — ~50% of revenue ships to China, with an anti-dumping probe (concluding ~Sep-2026) + a rule penalizing U.S. manufacturing; a direct overhang on the strategy. At this price, the best entry point likely comes with a cycle trough or a China scare — not near highs.