2 Top Artificial Intelligence Stocks to Buy Right Now

Arista Networks and Broadcom are benefiting directly from the AI-driven data center buildout: Arista reported Q1 2024 revenue growth of 16% year-over-year with net income up 46% and gross margin rising by over 4 percentage points, and roughly 39% of Arista revenues come from Meta and Microsoft. Broadcom posted fiscal Q2 2024 revenue of $12.5 billion, up 43% YoY, with AI-products revenue up 280% and comprising ~25% of sales (excluding the VMware acquisition, growth would have been ~12%); management expects about $11 billion in AI product revenue for the full year and called that a conservative estimate. Both names trade at elevated valuations despite strong fundamentals, so the piece views them as buys but recommends a cautious, dollar-cost-averaging approach given rich multiples.

Analysis

Market structure: AI-driven data‑center buildouts create clear upstream winners — high‑bandwidth switch vendors (ANET), ASIC/SoC suppliers (AVGO, NVDA), and hyperscaler software/hardware integrators (MSFT, META). ANET’s 39% revenue concentration with Meta+Microsoft and AVGO’s reported 25% AI revenue (280% growth) shift pricing power to suppliers able to deliver 400G/800G ports and AI silicon, supporting margin expansion of 200–400 bps in good quarters. Demand still outpaces supply for cutting‑edge GPUs and 400G/800G optics, implying tight ASPs and continued capex intensity for hyperscalers over the next 12–36 months.

Risk assessment: Key tail risks are (1) a hyperscaler capex pause that produces >20% YoY revenue misses for ANET/AVGO within 1–2 quarters, (2) regulatory/antitrust action or failed VMware integration that compresses AVGO multiples, and (3) hyperscalers internalizing networking silicon over 12–36 months, which would reduce supplier pricing power. Short‑term (next 30–90 days) volatility will hinge on quarterly guides; medium term (6–12 months) on supply chains and product ramps; long term (3–5 years) on architecture shifts to bespoke hyperscaler hardware.

Trade implications: Tactical allocations — overweight AI infrastructure (ANET, AVGO, NVDA, MSFT) and underweight legacy enterprise networking/software without AI share. Use DCA over 6–12 weeks into longs, hedge ANET customer concentration with puts sized ~30–50% of position for 3 months, and express leveraged upside in AVGO via 6–9 month call spreads (buy ATM, sell 20–30% OTM). Implement pair trade long AVGO / short CSCO equal notional to isolate AI infra exposure; trim if spread moves >15% or AVGO AI guidance misses by >5%.

Contrarian angles: Consensus underweights integration/quality noise — AVGO’s headline AI growth is partially M&A‑driven (VMware); upside surprises are possible if management’s $11bn AI guide proves conservative (>+10% upside). Conversely ANET may be over‑priced for its customer concentration: if ANET gross margin falls >200 bps sequentially or Meta/MSFT reduce orders by >15% it could re-rate down 25–40%, so size positions conservatively and set objective re‑eval triggers.

Key Decisions for Investors

• Establish a 3% portfolio long in AVGO via dollar‑cost averaging over 6–8 weeks; complement with a 6–9 month call spread (buy ATM, sell 20–30% OTM) sized at 50% notional to cap cost. If Broadcom’s next quarterly AI revenue guide > $12bn, increase position to 5%; if guide < $9bn, reduce to 0%.
• Accumulate ANET to a 2% portfolio weight over 3 months (DCA in 3 equal tranches); buy 3‑month puts at ~10% OTM sized 40% of the long position to hedge customer‑concentration risk. Liquidate or re‑assess if ANET reports sequential gross‑margin contraction >200 bps or if Meta/MSFT order commentary turns negative.
• Implement a pair trade: long AVGO / short CSCO equal notional sized 2% of portfolio to capture AI‑infra outperformance. Rebalance or close if the relative spread moves >15% or if AVGO AI revenue guidance misses consensus by >5% on the next earnings call (30–45 day catalyst).
• Reduce exposure by 1–3% to legacy enterprise networking/software names lacking AI infra exposure; redeploy proceeds into the above positions. Monitor three leading catalysts over next 30–90 days — Broadcom AI guidance, Arista customer commentary, and NVDA supply updates — and adjust weights if any catalyst changes expected revenue by >10%.