If I Had $10,000 to Put Into AI Stocks Right Now, Here's Exactly How I'd Invest It

The article argues for a three-stock AI portfolio, favoring Alphabet, AMD, and CoreWeave, with the largest weights in Alphabet and AMD at about $4,000 each and $1,900 in the speculative CoreWeave position. Alphabet is highlighted for $403 billion in 2025 revenue (+15%) and $132 billion in net income (+32%), while AMD posted $35 billion in revenue (+34%) and $4.3 billion in net income. CoreWeave remains highly risky, with $21.4 billion of debt and a $1.22 billion net loss, but its $66.8 billion backlog and forecast 142% revenue growth in 2026 support the bullish long-term case.

Analysis

The cleanest read-through is not “AI winners keep winning,” but that capital is starting to discriminate inside the AI stack. Alphabet and AMD sit in the part of the market where AI is monetizing through operating leverage, while CoreWeave is the opposite: growth is being bought with balance-sheet stress. That matters because the next leg of performance is likely to come less from model announcements and more from who can fund inference capacity, lower unit costs, and convert demand into durable cash flow.

Alphabet looks like the lower-beta compounding vehicle because AI is increasingly embedded in products with existing distribution, which reduces the need for external hype. The market’s underappreciated risk is that success there may suppress upside volatility: if AI merely preserves search/share and modestly expands cloud margin, the multiple may grind higher rather than re-rate sharply. The bigger second-order winner is likely the supplier ecosystem around AI inference and networking, where demand can persist even if end-user enthusiasm cools.

AMD’s setup is more interesting than the headline suggests. If it is gaining credible share in accelerators, the real implication is not just a stronger revenue line, but a more competitive pricing environment for Nvidia and more procurement optionality for hyperscalers. That can compress Nvidia’s pricing power over a 12-24 month horizon even if absolute demand keeps rising; the winner may be the buyer community, not the vendor that wins every benchmark.

CoreWeave is the speculative leg, but the market may be underpricing refinancing and dilution as the primary failure mode, not demand destruction. If backlog converts as expected, the equity can re-rate hard; if growth slows even modestly, debt service and capex can force a capital raise well before the business model is proven. This makes CRWV more of a macro-liquidity trade than a pure AI adoption trade, with asymmetric downside over the next 6-18 months.

Key Decisions for Investors

• Go long GOOG vs. long-only AI basket for 3-6 months: favor quality compounders where AI monetization is already embedded; target modest upside with lower drawdown than the more crowded AI trade.
• Initiate a pair trade long AMD / short NVDA over 6-12 months: if accelerator share continues to broaden, AMD offers higher delta to share gains while NVDA faces incremental pricing pressure; watch gross margin commentary as the key stop-loss signal.
• Avoid initiating fresh long CRWV until the market prices in financing risk more fully; if already involved, consider buying put spreads 3-6 months out to hedge dilution/refinancing risk while retaining upside.
• For aggressive accounts, sell cash-secured puts on AMD into weakness rather than chasing strength: the payoff is favorable if growth stays intact, and the implied volatility should remain supportive.
• Use any post-earnings rally in GOOG to add via call spreads rather than stock, targeting a 6-9 month horizon where AI-driven operating leverage can lift earnings without requiring a major multiple expansion.