Investors Rotating Out of Tech Are Making a Mistake. This AI Growth Stock's Revenue Just Hit a Record.

Broadcom posted a record $63.9 billion in fiscal 2025 revenue and saw AI chip sales surge 65% to $20 billion, with AI revenue projected to rise to $60-$90 billion by fiscal 2027. Analysts expect revenue and adjusted EBITDA to grow at 47% and 46% CAGRs, respectively, from fiscal 2025 to fiscal 2028, reinforcing Broadcom’s role as a bellwether for AI demand. The article argues the AI market still has substantial room to grow, with industry CAGR estimates of 30.6% from 2026 to 2033.

Analysis

The market is still treating AI as a single trade, but this setup is really a bifurcation between training economics and inference monetization. Broadcom’s mix implies that the next leg of capex is less about frontier-model bragging rights and more about hyperscalers optimizing unit economics, which favors custom silicon, networking, and software bundles over pure-play GPU exposure. That means AVGO’s strength is not just a company-specific rerating; it is a signal that the buyers of AI infrastructure are moving from experimental spend to industrialized deployment.

The second-order winner is likely the supply chain around AI networking, interconnect, and data-center power efficiency, because inference at scale is constrained by throughput and cost per query rather than model quality alone. If hyperscalers keep reallocating capex toward ASICs, NVDA’s share of wallet can remain enormous, but its growth multiple may be more sensitive to every incremental sign of custom-chip substitution than the market is pricing. META and GOOGL benefit operationally from lower long-run inference cost, yet the equity read-through is trickier: lower AI unit costs can improve margins without necessarily expanding headline growth enough to justify further multiple expansion.

The main risk is that investors extrapolate the current AI capex cycle too far ahead of monetization. If enterprise adoption lags while hyperscalers tighten budgets, the market could punish anything with AI embedded in the narrative, and the first stocks to de-rate would be the ones with the most levered expectations rather than the best near-term fundamentals. Over a 3-6 month horizon, the catalyst to watch is whether hyperscaler commentary confirms sustained custom-silicon procurement into 2026; if not, this looks like a momentum trade vulnerable to abrupt factor rotation.

The contrarian read is that the move may be underpriced in duration but overowned in positioning. AVGO is increasingly a toll collector on the AI stack, so its downside is less about AI demand vanishing and more about digestion after a very strong run; that creates opportunity in relative value rather than outright chasing. The highest-conviction expression is to prefer monetization enablers over beta-heavy AI names until the market proves that inference spending is broadening beyond a handful of cloud giants.

Key Decisions for Investors

• Stay long AVGO on pullbacks over the next 1-3 months; use it as the cleaner AI monetization proxy, but size smaller than a fresh entry because positioning and valuation are already extended after the run.
• Pair trade: long AVGO / short a basket of higher-beta AI winners or semis with less direct inference exposure over 1-2 quarters; the relative thesis is that custom-silicon share gains and software bundling provide more durable earnings power than pure narrative-driven AI beta.
• Buy 3-6 month downside protection on NVDA into strength if implied vol remains cheap versus realized; the risk/reward improves if hyperscaler commentary turns more cautious on custom-chip substitution or capex pacing.
• Overweight networking and data-center infrastructure beneficiaries versus compute-only names over the next 6-12 months; if inference scales as expected, the bottleneck shifts toward bandwidth, power efficiency, and system integration.
• If AVGO rips another 10-15% without a corresponding upward revision cycle, trim into strength and wait for a post-earnings reset; the trade is attractive on fundamentals but vulnerable to multiple compression after a strong momentum move.