Billionaire Stanley Druckenmiller Still Isn’t Buying Nvidia. But He Just Opened a Position in a Challenger That’s Soared More Than 100% This Year.

Stanley Druckenmiller exited Nvidia in 2024 on a “rich” valuation but in the latest quarter opened a new position in AMD, buying 72,900 shares and assigning it a 0.8% portfolio weight. The article highlights AMD’s AI momentum (doubling data center revenue) and cites Nvidia revenue of ~$96B vs AMD’s ~$11B, arguing AMD could offer more upside growth. Despite the bullish AI narrative, AMD trades at 60x+ forward earnings, which is framed as a valuation headwind for new investors.

Analysis

This is more a positioning signal than a fundamental inflection. Druckenmiller’s move tells you the market is still willing to pay up for AI second-derivative exposure, but it does not answer the harder question: whether AMD can convert design wins into durable earnings power without giving away margin to buy share. At 60x+ forward earnings, the stock is already pricing in a multi-quarter beat-and-raise cycle; that leaves the asymmetry skewed toward disappointment if data-center growth normalizes faster than bulls expect.

The real second-order beneficiary is hyperscale customers and enterprise buyers, not AMD alone. A credible alternative to NVDA can restrain GPU pricing and improve procurement leverage across the AI stack, while also lifting adjacent names like SMCI, DELL, and HPE if rack-scale deployments broaden. The loser is less “NVDA demand” and more NVDA’s pricing power; if AMD takes share, the market may start valuing the entire AI semiconductor basket on a lower terminal margin assumption.

Near term, the catalyst is the next one to two earnings prints: investors will care less about headline AI enthusiasm and more about whether data-center revenue acceleration continues and gross margin holds while product mix shifts. If AMD’s AI contribution doesn’t visibly outgrow the rest of the business over the next two quarters, the multiple should compress rather than expand. Contrarian view: the market may be underestimating the CPU angle for inference and agentic workloads, but it may also be overestimating how quickly that translates into profits versus revenue.

Key Decisions for Investors

• Hold off on initiating a core AMD long at current valuation; prefer a starter position only after a 10-15% pullback or after the next earnings report confirms continued data-center acceleration. Time horizon: 1-3 months. Falsifier: any guide that implies AI revenue growth is decelerating faster than the broader business.
• For tactical exposure, use a small long AMD / short NVDA pair for 4-8 weeks to express a second-source adoption thesis. Risk/reward is attractive only if the market rotates into relative valuation; stop out if NVDA reaccelerates order growth or AMD’s next guide disappoints on AI mix.
• Buy the basket instead of the single name: prefer SOXX or SMH over concentrated AMD risk if the objective is AI semiconductor beta over the next 6-12 months. This reduces the risk that AMD’s valuation multiple compresses even if the sector remains strong.
• Set an alert on AMD’s next quarter for data-center growth and gross margin trend. If data-center growth stays strong but margin deteriorates, fade the stock on the view that revenue share is being bought at the expense of earnings power.