Peter Thiel’s investment vehicle, Thiel Macro, has allocated 42% of its portfolio to two artificial intelligence-related stocks, Amazon and Vistra, both acquired during the second quarter. The fund outperformed the S&P 500 over the last six quarters, a track record that includes two quarters in which it held no stocks. This concentration in specific AI beneficiaries offers a distinct perspective on where capital is flowing within the sector.
Amazon serves as Thiel Macro’s largest position. Although classified in the consumer discretionary sector due to its retail operations, the company is frequently grouped with major technology firms because of its cloud computing arm. In retail, Amazon has deployed hundreds of generative AI applications to optimize inventory placement, demand forecasting, and last-mile delivery. The company also operates the largest fleet of industrial mobile robots, using AI to enhance efficiency. Its next-generation Proteus robots can interpret natural language, allowing human workers to issue instructions conversationally.
In the cloud sector, Amazon Web Services addresses AI demand across infrastructure, platform, and software layers. AWS has developed custom AI accelerator chips that serve as a lower-cost alternative to Nvidia GPUs. It also provides development tools such as Bedrock and ready-to-use software like Kiro, an AI agent designed to automate coding. During the second quarter, AWS revenue rose 37%, marking the fastest growth in more than four years. CEO Andy Jassy stated that the company is unusually well-positioned for the AI inflection due to its leading price-performance chips, specifically Trainium for AI and Graviton for CPU.
Amazon currently trades at 21 times earnings. Wall Street estimates suggest earnings will grow at 21% annually over the next three years, a pace that most analysts view as supporting an undervalued stock. The median target price stands at $330 per share, representing 27% upside from the current price of $259.
The second holding, Vistra, is positioned to benefit from surging power demand driven by AI infrastructure. Goldman Sachs estimates U.S. data center power demand will reach 66 GW next year, more than double the 31 GW recorded last year. Broader U.S. electricity demand is estimated to compound at 3.5% annually through 2030, up from roughly 0% annually in the previous decade.
As the largest competitive power producer in the U.S., Vistra generates electricity from nuclear, natural gas, coal, and renewable sources. It sells power into wholesale markets and directly to residential and commercial customers. Earlier this year, Vistra secured a deal with Meta Platforms to supply 2.6 GW of nuclear generation over a twenty-year term. The company also signed a similar agreement with Amazon. These contracts provide long-term financial clarity, and Vistra could pursue additional contracts with other hyperscalers as the AI sector expands.
For valuation purposes, investors should look at price-to-operating-cash-flow rather than price-to-earnings for Vistra. Price-to-earnings metrics can be distorted by non-cash adjustments related to derivatives used to hedge natural gas and electricity price volatility. Operating cash flow offers a clearer view of core business performance. Vistra currently trades at 10.8 times operating cash flow, a discount to its 10-year average of 12. The company’s operating cash flow is forecast to increase at 19% annually through 2030. The median analyst target price is $210 per share, implying 29% upside from the current share price of $162.