JPM AI Capex Deep Dive; Underappreciated AI Names; HDDs Rally; QCOM; MLCCs

研报英文原文证据摘录

JPM AI Capex Deep Dive; Underappreciated AI Names; HDDs Rally; QCOM; MLCCs

reflecting – in their view - persistent

skepticism about its AAPL dependence (losing – by their accounting - $5b/11% of revenue as AAPL shifts to in-house modems)

and its identity as primarily a smartphone chip company. The bull case, they argue, rests on diversification (automotive revenue

up from $240m to $1.3b in five years, ~45% growth forecast next quarter; IoT up 9% YoY) plus a credible AI data-center push—

existing inference chips, a custom AI chip deal with a major tech company via the Alphawave acquisition shipping later this year,

with more details expected at next week’s investor meeting.

Elsewhere the HDD names have been on a tear (WDC is up another 8% premarket, and STX +6.7%). The proximal cause is a

rash of sell-side upgrades (which included this one from Samik late last week), including one from a competitor yesterday

suggesting pricing is so good as to imply the stocks may only be trading on low dd% earnings; that competitor hosted WDC on an

NDR, and also apparently left saying management could see EB supply compound at 40% over the next 3~5 years. But more than

that yesterday I got questions all day long about the strong outperformance of WDC versus STX (which continues today). I don't

see any cut and dry reason, but here are a few things that stick out: 1) the above-referenced report highlights a WDC NDR, so that

commentary is most immediate there; 2) until yesterday, WDC had been a YTD underperformer (closed the gap yesterday,

though), 3) STX had been better bought (and a more crowded HF long), I think in part on an exaggerated view that HAMR 2.0

will drive strong margin outperformance, and 4) we see ongoing short covering (even though WDC is still a net long, shorts are