3 Beaten Down AI Infrastructure Stocks to Buy in June

NVIDIA, Broadcom and Microsoft are all trading below their 52-week highs even as AI revenue growth remains strong, with NVIDIA Q1 revenue up 85% YoY to $81.61B, Broadcom AI semiconductor revenue up 143% YoY to $10.80B, and Microsoft AI business run-rate above $37B. Guidance remains constructive: NVIDIA guided Q2 revenue to $91.0B, Broadcom sees Q3 AI semiconductor revenue at $16.0B, and Microsoft’s commercial RPO reached $627B. The setup is positive for AI infrastructure investors, though valuation, capex intensity and export restrictions remain key risks.

Analysis

The market is treating this as a synchronized de-rating, but the setup is actually a leadership rotation inside the AI capex stack. NVDA remains the purest beta to spend, while AVGO and MSFT are increasingly the “plumbing” and monetization legs; that matters because when the spend cycle matures, the winners tend to shift from compute scarcity to integration and network efficiency. Second-order, the pullback in AVGO is especially interesting because it may be signaling that hyperscalers are becoming more selective on incremental accelerator spend, which can pressure merchant GPU sentiment before it shows up in absolute demand.

The biggest overhang is not demand, but timing and customer mix. NVDA and AVGO both rely on a handful of hyperscalers for outsized incremental orders, so a single quarter of digestion can compress multiples faster than fundamentals deteriorate. MSFT is the cleaner duration asset here: its backlog and operating leverage can absorb a capex-heavy quarter, but if AI monetization lags infrastructure outlays for another 1-2 quarters, the stock can keep underperforming despite strong reported growth.

Contrarianly, the consensus may be too eager to buy the dip in the most crowded names while underestimating how quickly margins can normalize once supply bottlenecks ease. The better asymmetry is probably not “own all AI,” but own the name with the best balance between earnings visibility and valuation support, then use the higher-beta names tactically. The next catalyst window is the next earnings cycle and guidance update; until then, these are trading vehicles on capex confidence rather than clean long-duration compounders.

Key Decisions for Investors

• Go long MSFT on weakness over the next 2-6 weeks; best risk/reward in the group because the valuation reset already reflects capex skepticism, while backlog and cloud growth give the cleanest downside support.
• Buy NVDA only if it reclaims prior support after the next guide/earnings catalyst; otherwise keep position size smaller than AVGO/MSFT because its beta makes it the most vulnerable if AI capex pauses for even one quarter.
• Use AVGO as a tactical mean-reversion trade: initiate a partial long position with a 1-3 month horizon, but size below normal because premium valuation and customer concentration make it the easiest name to de-rate again on any hyperscaler slowdown.
• Pair trade: long MSFT / short NVDA or AVGO into the next quarter if you expect capex digestion; the trade benefits if the market rewards monetization and penalizes higher-beta infrastructure exposure.
• For options, consider buying 3-6 month MSFT calls financed by selling upside in NVDA or AVGO; this expresses a view that the market will pay up for durability over raw growth if AI returns remain uneven.