Software vs Chips: How I Am Positioning For The 2026 Rotation

One of the cleanest trends in markets this year has been money walking out of software and into chips. It is worth understanding the mechanics before deciding whether to chase it.

The rotation in plain terms

The thesis is simple. AI infrastructure spending is visible, enormous, and landing in the income statements of chipmakers right now. Software, by contrast, faces an open question about whether AI agents compress its pricing model. Faced with certainty on one side and a question mark on the other, large funds rotated toward the certainty. Microsoft sits well below its highs while the memory and accelerator names keep printing.

Where the rotation can overshoot

Rotations are momentum trades, and momentum overshoots. Chips are cyclical, and parts of the complex are priced for a supercycle that never ends. Software meanwhile is being treated as if AI only threatens it and never helps it, which ignores that the largest software firms are also the ones selling AI. When a trade gets this consensus, the easy money is usually already made.

My framework

I keep core exposure to the AI infrastructure winners because the spending is real, but I am not adding into vertical moves. On the software side I am building a watchlist of quality names sold off with the sector, waiting for signs the per seat fear is overdone for that specific business. The point is not to pick software or chips as a team. It is to own the durable winners on both sides and let the panic hand me better entry prices.

Bottom line

The rotation is real, but late stage consensus trades reward discipline over conviction. I would rather be early on quality software being oversold than late on chips being overbought.