Broadcom Stock Is Down More Than 22% From Its Peak and Worth a Closer Look

Broadcom shares fell more than 22% from their early-June high after management did not raise its fiscal 2027 AI chip revenue forecast, even though the company still expects AI chip revenue to exceed $100 billion in fiscal 2027 versus under $64 billion of total revenue in fiscal 2025. The article argues the pullback is an overreaction and highlights strong demand from hyperscalers for custom AI chips, TPU orders worth $21 billion from Anthropic, and Broadcom's networking/CPO advantages. Valuation has compressed to about 19.5x fiscal 2027 earnings estimates, which the author views as attractive.

Analysis

The market is treating the guide non-update as an information event, but the deeper signal is simply that the AI supply chain is moving from narrative scarcity to execution scarcity. In that phase, the winners are the vendors who sit at the control point between hyperscaler capex and deployment bottlenecks: custom silicon plus networking plus optical interconnects. That makes AVGO less a “chip” trade and more a toll-road on the AI buildout, with pricing power reinforced by design-in stickiness and multi-year platform transitions.

The second-order implication is that any slowdown in Nvidia displacement is not necessarily negative for AVGO; it may just stretch the duration of the trade. Custom accelerators still need network monetization, so even if customer ramp schedules slip, the networking/CPO attach rate can cushion revenue timing. The real competitive pressure is likely to fall on generic AI inference hardware vendors and on smaller interconnect names that lack system-level breadth.

Consensus is probably underestimating how much of AVGO’s value is now tied to capex visibility rather than near-term guide hygiene. A forward multiple around the high-teens is not demanding if hyperscaler AI budgets remain in an expansion regime for the next 4-8 quarters. The risk is not valuation compression from “missed” guidance; it is a later-stage digestion period in which customers optimize existing clusters before the next TPU/custom ASIC wave, which could create 1-2 quarters of air pockets even while the secular thesis stays intact.

Key Decisions for Investors

• Add AVGO on further weakness over the next 1-3 weeks; use a staged entry rather than all at once, because the near-term setup is still sentiment-driven. Target 12-18 months, with upside if AI networking/CPO attach continues to re-rate the business as infrastructure rather than semis.
• Pair trade: long AVGO / short NVDA for 3-6 months as a relative-value expression of custom-ASIC share shift and networking monetization. Risk: if GPU demand reaccelerates or inference economics favor Nvidia again, the spread can reverse quickly.
• Buy AVGO call spreads 6-12 months out to capture a bounce in sentiment while capping theta risk. A 1:2 or 1:3 risk/reward structure fits the thesis that the selloff is more about disappointment optics than fundamental deterioration.
• Stay cautious on smaller AI hardware and interconnect names that rely on one product cycle or one hyperscaler. Broadcom’s multi-product control point makes it the safer way to own AI capex, while niche vendors are more exposed if hyperscaler procurement pauses.
• Use any further AVGO weakness as a hedge against an overextended NVDA-heavy book; the trade is less about outright bearishness on Nvidia and more about owning the part of AI infrastructure with the best revenue capture per deployed dollar.