Here's Why Nvidia and Broadcom Are Still Leading the Pack for AI Investing

Nvidia reported revenue growth of 73% to $68.1 billion in its latest quarter, while analysts expect 79% growth next quarter and 85% in the following quarter. Broadcom's fiscal Q1 AI semiconductor revenue jumped 106% year over year to $8.4 billion, and management sees its custom AI chip business reaching $100 billion or more by 2027. The article argues both stocks remain attractive long-term AI beneficiaries as global data center capex is expected to rise through 2030 to $3 trillion-$4 trillion.

Analysis

The market is still underestimating how much of the AI capex cycle is becoming a two-horse race between a general-purpose platform and a custom-silicon toll collector. That matters because the spending is no longer just about incremental model training; it is shifting into repeat procurement, networking, and replacement cycles, which creates a longer-duration revenue stream than the market typically assigns to “equipment” winners. The key second-order effect is that each wave of AI deployment increases software lock-in and switching costs for hyperscalers, making the incumbent chip ecosystem harder to displace even if performance parity emerges.

The bigger mispricing may be in the supply chain, not the obvious leaders. If custom AI accelerators keep taking share inside large customers, it raises demand for adjacent beneficiaries in advanced packaging, high-bandwidth memory, optics, and foundry capacity while pressuring legacy merchant silicon vendors that lack a platform position. At the same time, the growth rate implied here is so elevated that any moderation in order cadence could trigger sharp multiple compression, especially if investors start discounting 2027+ numbers too aggressively today.

The contrarian risk is that the consensus is extrapolating unit growth while ignoring cycle math: the first phase of AI build-out is being pulled forward, so headline revenue growth can decelerate even if end demand remains strong. The most important catalyst to watch over the next 1-2 quarters is whether hyperscaler capex guidance broadens beyond the top names or stays concentrated; concentration would increase near-term upside for NVDA/AVGO but also raise the probability of a digestion phase later. A secondary risk is export controls or foundry bottlenecks, which could convert demand strength into timing slippage rather than lost demand, creating violent but temporary drawdowns.

Key Decisions for Investors

• Stay long NVDA and AVGO on 3-6 month horizon, but size with discipline: use 5-10% pullbacks as entry points rather than chasing strength; upside remains tied to capex revisions, while downside is mostly multiple compression.
• Express the relative winner via long AVGO / short a basket of legacy semiconductor merchants over 6-12 months; custom silicon and platform integration should capture more of the AI wallet than undifferentiated chip names.
• Add a call spread on NVDA into any post-earnings consolidation if implied vol resets; the risk/reward is favorable because near-term order visibility remains tight, but premium should be financed by selling farther-dated upside.
• Monitor and potentially pair long NVDA/AVGO against short a high-multiple AI infrastructure basket if capex headlines begin to fade; the leaders should hold up better than second-tier beneficiaries once the market shifts from growth scarcity to growth durability.
• Set a catalyst watch on hyperscaler capex commentary for the next earnings season; if spending broadens rather than concentrates, increase exposure to the semiconductor supply chain broadly, especially packaging and memory, where second-order demand can lag but compound.