3 Cheap Artificial Intelligence Stocks to Buy Right Now

Microsoft, Meta Platforms, and Nvidia are presented as undervalued AI leaders despite strong operating momentum. Microsoft’s AI business is running at a $37 billion annualized rate and grew 123% last quarter, while Azure revenue rose 40%; Meta’s most recent quarterly revenue grew 33%; and Nvidia is trading at 23x estimated earnings even as analysts forecast 41% growth next year. The article argues these stocks are priced at a discount relative to their growth prospects, making the overall message bullish but primarily opinion-driven.

Analysis

The market is treating AI capex as a zero-sum tax on returns, but the setup here is more nuanced: the leaders are turning infrastructure spend into measurable monetization faster than investors are willing to underwrite. That matters because once a hyperscaler shows sub-scale AI spend converting into visible operating leverage, the multiple compression usually stops before the earnings inflection fully shows up in consensus. The second-order beneficiary is the AI supply chain: semis, networking, and power infrastructure should continue to see demand durability even if headline sentiment rotates away from the mega-caps.

The clearest contrarian opportunity is that all three names are being forced into the same valuation box, despite very different cash flow sensitivity. Microsoft is the cleanest “discount to quality” expression because the core business can absorb AI investment without needing a near-term product surprise; the risk is mostly time horizon mismatch, not fundamental decay. Meta is the highest beta on execution: if AI monetization proves real, the rerating can be abrupt, but if spend rises faster than ad pricing, the stock can de-rate first and recover later.

Nvidia remains the most misread. The market still seems to be anchoring on current-cycle earnings rather than the next leg of compute buildout, which means the stock can re-rate without a single estimate raise if investors start pricing a longer duration of elevated demand. The main risk is not demand failure but pacing: any sign that cloud capex is front-loaded or delayed by customer digestion could create a 1-2 quarter air pocket, even if the multi-year thesis stays intact.

From a tape perspective, this is less a “buy AI” trade than a “buy earnings visibility before consensus catches up” trade. The better-risked entry is on pullbacks or post-print volatility rather than chasing strength, because the narrative already has enough sponsorship that incremental upside will come from estimate revisions, not multiple expansion alone. In contrast, the broader AI basket still offers relative value versus the mega-cap leaders if investors want exposure without paying for single-name execution risk.

Key Decisions for Investors

• Long MSFT on any 5-8% pullback over the next 2-6 weeks; target a re-rating toward market-plus quality premium as fiscal 2027 estimates become the focal point. Risk: if AI monetization growth decelerates sharply, the stock can stay rangebound despite strong fundamentals.
• Long META vs short SPY as a 3-6 month pair trade; this is a high-upside rerate if AI spend shows incremental ad monetization, but downside is capped by the core ads franchise. Cover if operating expense growth outpaces revenue growth for two consecutive quarters.
• Buy NVDA on weakness rather than momentum-chasing; horizon 6-12 months. The setup favors multiple expansion if 2027 growth starts getting priced in, but the trade should be sized for cyclical digestion risk over the next 1-2 quarters.
• Express a relative-value long basket: MSFT/META/NVDA vs short a broad tech ETF only if AI sentiment overheats; otherwise stay directional long the leaders. This is a hedge against a sector-wide de-rating while preserving upside to earnings revisions.
• For options-sensitive accounts, use NVDA call spreads 6-9 months out to capture the repricing of out-year growth while limiting blowup risk if capex pauses or guidance comes in conservatively.