Semiconductor Price Targets Keep Rising. Ratings Mostly Don't.

Reading analyst research well requires separating two different acts. Raising a price target is a mechanical consequence of pushing a valuation model forward a quarter or marking up an estimate. Changing a rating is a statement about expected relative return. August produced a great deal of the former and almost none of the latter.

What a target change is worth

Empirically, target revisions carry information mainly when they accompany estimate revisions. A target lifted because the analyst rolled a discounted cash flow forward twelve months tells you about the calendar, not the company. A target lifted alongside a raised revenue forecast tells you the analyst's channel work has changed.

The current chip complex sits with consensus targets roughly 10 to 18 per cent above spot prices — a spread that has been broadly stable all year even as prices rose, which is itself evidence that models are following prices rather than leading them.

Where the genuine disagreement is

Two names carry real dispersion. Intel's coverage is split between a foundry-turnaround case and a structural-decline case, with targets from $20 to $50. Tesla's spread runs from $150 to $500 depending almost entirely on what value the analyst assigns to autonomy. Wide dispersion is more useful than a high average: it identifies the stocks where the outcome, not the multiple, is the argument.