Bloomberg Intelligence: Broadcom Drops on Chip Outlook (Podcast)

Broadcom shares slid after its two-year AI chip sales forecast failed to impress investors, signaling the shift away from Nvidia remains early. Offsetting strength came from Snowflake, which jumped after raising its annual sales outlook and citing rapid adoption of its AI-assisted coding tool. Elsewhere, Nvidia agreed to acquire Hugging Face for $12.93B, while Hewlett Packard Enterprise sank on disappointing sales growth despite boosting revenue guidance for the year ending October (+34% to +37%). Campbell’s shares also fell after its fourth straight quarterly sales decline, dividend cut, and a $500M cost-savings plan to offset higher supply-chain and tariff costs.

Analysis

The cleanest read-through is not that AI demand is fading, but that the market is becoming more selective about where it pays for AI exposure. Broadcom is getting treated as a proxy for merchant-silicon share gains without yet proving the same scale economics investors assign to the category leader; that tends to compress multiples first, then force the company to either accelerate cadence or concede the narrative. The second-order effect is favorable for vertically integrated hyperscaler chips and for software names that can show direct monetization, while intermediate infrastructure suppliers face a tougher hurdle rate.

Snowflake’s reaction suggests capital is rotating toward software where AI can be tied to near-term usage, not just future TAM. That helps names with clear workflow attachment and hurts “AI by association” trades, especially if enterprise buyers are still rationing infrastructure spend. For hardware, HPE’s issue is actually more important than the stock move: supply constraints with visible demand usually mean the revenue deferral is real but temporary, which can set up a later catch-up quarter if supply normalizes; the stock will likely stay range-bound until investors see conversion, not just bookings.

Campbell’s is a reminder that defensive staples are not defensive when input inflation and tariffs squeeze pricing power. Dividend compression is usually the market telling you the earnings model has to reset, and that can spill over to other packaged-food names if the consumer is trading down faster than companies can re-price. Contrarian view: the broad AI pullback may be overdone in the short term because the selloff is punishing “good but not dominant” exposure more than actual demand deterioration; the real signal is a dispersion trade, not a sector de-rating.

Key Decisions for Investors

• Short AVGO on rebounds over the next 1-2 weeks; treat it as a valuation/expectations reset rather than a demand collapse trade. Falsifier: a follow-up guide that shows accelerating AI revenue conversion or renewed margin expansion.
• Buy SNOW call spreads or a small tactical long for the next 1-3 months; the stock can re-rate if AI-assisted product adoption translates into better consumption trends. Falsifier: any evidence that usage growth stays choppy into the next earnings print.
• Short CPB into strength for the next 1-2 quarters; the dividend action signals lower earnings quality and less pricing power than the market had assumed. Falsifier: stabilizing free cash flow and a credible pass-through of cost inflation.
• Watch HPE, do not chase the down move; if supply constraints ease before the next print, there is a potential catch-up trade from deferred revenue recognition. Falsifier: backlog fails to convert even after supply normalizes.