This AI Chip Giant Quietly Became Worth More Than Tesla, and Many Investors Still Overlook It

Broadcom’s AI revenue more than doubled to $8.4B last quarter, with total revenue up 29% year over year to a record $19.3B and fiscal Q2 guidance calling for 47% revenue growth to $22B. The company also returned $10.9B to shareholders via buybacks and dividends and authorized another $10B repurchase, while management sees a path to more than $100B in AI chip revenue by 2027. The stock has surged about 85% over the past year and trades at roughly 87 times earnings, leaving limited room for disappointment despite strong operating momentum.

Analysis

The market is treating AVGO less like a cyclical semiconductor vendor and more like a toll collector on the AI buildout. That re-rating is justified as long as the company remains embedded in the architecture decisions of hyperscalers, because custom silicon and network interconnect are two of the hardest spend categories to rip out once clusters are designed around them. The second-order effect is that AVGO’s strength reinforces capex durability across its largest customers, which should keep demand shockwaves moving upstream into HBM, advanced packaging, and foundry capacity rather than into consumer-facing AI devices.

The bigger winner may be GOOGL and META, which can use bespoke chips to defend AI margins against NVDA’s premium pricing and to lower inference costs at scale. If AVGO’s roadmap truly extends through 2028, that implies the bargaining power has shifted toward the largest cloud platforms, not away from them: they get supply assurance and cost control, while smaller AI players face tighter access to leading-edge compute and networking. INTC is the structural loser here because every dollar of custom-accelerator spend that bypasses general-purpose x86 weakens its relevance in the most valuable end market.

The main risk is not near-term execution, but duration: the stock is implicitly discounting multi-year AI spend persistence, so any slowdown in hyperscaler capex growth could compress both the multiple and the growth rate at the same time. The first place to watch is forward order commentary over the next 1-2 quarters; if customer concentration starts to look like customer dependency, the valuation can de-rate quickly. TSLA is only a marginal comparison point, but its lower AI sensitivity makes it a useful relative short against the beneficiaries of infrastructure spending.

Consensus appears to be missing how much of AVGO’s upside is already tied to a small set of very sophisticated buyers. That improves visibility in the near term, but it also means the market is effectively underwriting continued strategic commitment from a handful of CFOs and CTOs who are under pressure to optimize ROIC. In that context, the risk/reward is less about product demand and more about whether the same customers continue to justify multi-year silicon roadmaps without pausing to digest prior deployments.

Key Decisions for Investors

• Stay long AVGO only on pullbacks into the next earnings window; use a 6-12 month horizon and treat a 15-20% drawdown as the entry zone, because the risk is multiple compression rather than business deterioration.
• Pair trade: long GOOGL / short INTC for 3-6 months. GOOGL benefits from lower AI cost structure and supply priority, while INTC remains structurally exposed to AI share loss; target relative outperformance of 10-15%.
• Consider a tactical long AVGO vs short NVDA pair for 1-2 quarters if investors rotate toward infrastructure names with clearer customer-specific roadmaps; AVGO has more visible contractual demand, while NVDA is more exposed to broader market expectations.
• Buy AVGO downside protection rather than outright shorting: 3-6 month puts or collars into earnings can express concern about valuation compression while limiting risk if AI capex stays hot.
• Overweight META and GOOGL over smaller AI exposure names for the next 12 months, as custom silicon lowers cost per inference and should translate into better AI margin capture than hardware-only beneficiaries.