AI Stocks: One Buy, One to Avoid

History often provides valuable lessons for Wall Street observers. Although the past isn't a perfect predictor, market patterns tend to repeat. Crestmont Research indicates that the S&P 500 has consistently delivered positive total returns across all rolling 20-year periods since 1900, including dividends. This suggests that holding an S&P 500 index fund for two decades between 1900 and 2005 would have always yielded profits.

Trillion-Dollar Stocks and AI

Historical data can also guide us when looking at major companies. To date, only ten publicly traded U.S. firms have reached trillion-dollar valuations, including the "Magnificent Seven", Berkshire Hathaway, Broadcom, and Taiwan Semiconductor Manufacturing. Examining this group through a historical lens suggests that one trillion-dollar stock, significantly boosted by the artificial intelligence (AI) boom, faces potential risks. Conversely, another appears to be a strong buy.

Nvidia's Rise and Potential Bubble

Since the start of 2023, Nvidia has been among the market's top performers. The company's stock has surged nearly 1,200%, increasing its market capitalization by roughly $4.2 trillion. Nvidia's growth is primarily attributed to its graphics processing units (GPUs). Its Hopper (H100), Blackwell, and Blackwell Ultra AI-GPUs are crucial for AI-accelerated data centers and large language model (LLM) training. The demand for these AI-GPUs has allowed Nvidia to command prices of $40,000 or higher per unit.
Nvidia's CUDA software platform is key, enabling developers to fully utilize its GPUs and develop LLMs. Without CUDA, Nvidia might struggle to retain developers within its ecosystem. The company's dominance in AI infrastructure suggests a promising investment. However, history reveals potential pitfalls. Over the past three decades, nearly every major technological advancement has experienced a bubble.

Examples include the internet, genome decoding, nanotechnology, 3D printing, blockchain, and the metaverse. While AI could be an exception, it seems unlikely. Currently, most businesses are still optimizing their AI solutions and are not seeing returns on their investments, echoing the patterns seen before previous tech bubbles burst. If an AI bubble were to materialize, Nvidia could face significant challenges. Furthermore, competition in emerging technology markets is unavoidable. Despite the focus on external GPU developers, Nvidia's primary threat could come from its largest customers. Many members of the "Magnificent Seven" are developing their own AI chips for internal use. Although these chips are currently slower, they are more cost-effective and can reduce the scarcity of Nvidia's GPUs.

Adding to the concern, Nvidia's price-to-sales (P/S) ratio is unsustainably high. Historically, megacap companies at the forefront of technological advances have peaked at P/S ratios between 30 and 40. Nvidia's P/S ratio has fluctuated between 25 and 42 over the last two years.

Meta Platforms: A More Promising AI Play

In contrast to the risks associated with Nvidia, history suggests a positive outlook for Meta Platforms, despite the AI boom benefiting the "Magnificent Seven". Meta is investing heavily in AI data center infrastructure, including purchasing hardware from Nvidia. However, AI is a relatively small component of Meta's overall business. While Nvidia's data center segment accounts for nearly 90% of its sales, Meta derives roughly 98% of its revenue from advertising. Meta's suite of apps, including Facebook, Instagram, WhatsApp, Threads, and Facebook Messenger, attracts an average of 3.48 billion daily users, giving Meta Platforms substantial ad-pricing power.

Meta is primarily leveraging AI for application-based solutions, offering businesses generative AI tools to customize ads. These AI solutions enhance Meta's pricing power and improve click-through rates. A potential AI bubble burst would not significantly impact the ability of businesses to utilize these solutions. More importantly, advertising-driven businesses benefit from economic cycles. Over the past eight decades, U.S. recessions have averaged just 10 months, while economic expansions have lasted approximately five years. As advertising is cyclical, Meta Platforms is more likely to thrive during economic expansions.

Meta's strong balance sheet also supports its stock. As of June, the company held over $47 billion in cash and marketable securities and is on track to generate more than $99 billion in cash from operations this year. This financial strength allows Meta to invest in long-term projects like the metaverse. Additionally, Meta's forward price-to-earnings (P/E) ratio of 24 is only slightly above its five-year average. If historical patterns hold, Meta stock is well-positioned for future gains.