Over the past 30 years, Wall Street has always had a next-big-thing trend to captivate investors' attention. At the moment, no trend is garnering attention quite like the rise of artificial intelligence (AI).
AI involves the use of software and systems to handle tasks that would normally be overseen or undertaken by humans. What makes AI special is the incorporation of machine learning (ML), which is what gives AI-driven software and systems the ability to learn, evolve, and become more efficient at their tasks over time. AI and ML have broad-based application across virtually all sectors and industries, which is a big reason why professional services company PwC is forecasting a $15.7 trillion lift to global gross domestic product from AI by 2030.
Image source: Getty Images.
But as with all next-big-thing investments, the outlook for the companies riding their wave is mixed. Moving beyond the hype that often accompanies hot trends, one AI stock stands out as a clear-cut buy in October, while another high-flying AI stock could be headed for disappointment.
The artificial intelligence stock that's a screaming buy in October: Intel
Among the vast sea of companies that have AI ties, semiconductor giant Intel (INTC -9.20%) stands out as the top buy in October.
A quick look at Intel's multiple operating segments shows that it's not generating a substantive portion of its sales from AI chips and graphics processing units (GPUs) at the moment -- and this is precisely what makes the company such a no-brainer investment.
Every next-big-thing trend for the past 30 years has gone through an initial bubble phase where investor expectations outpace the actual uptake of a new product or service. Since Intel has numerous other revenue channels to rely on as it builds out its AI product line, the company can continue to benefit from its legacy operations while waiting for the AI use case to mature a bit.
To be certain, Intel wants to be a major infrastructure player behind the AI movement. The company's Xeon processors are being used in high-compute data center servers today, with the Falcon Shores GPU expected to make its debut in 2025.
Meanwhile, Intel's 14th generation central processing unit (CPU), Meteor Lake, is slated to hit the market in 2024. The new chiplet design for Meteor Lake, which includes a vision processing unit (VPU), meaningfully reduces the compute requirements for AI inferencing and allows the CPU, GPU, and VPU to individually leverage each of their strengths.
While Intel has dealt with weaker CPU demand in the quarters following the worst of the COVID-19 pandemic, it continues to account for the lion's share of CPU sales in personal computers, data centers, and mobile. Even with chief CPU rival Advanced Micro Devices (AMD 2.37%) chipping away at its market share lead, Intel remains well ahead of AMD in these categories. In other words, Intel's legacy segments are sustained cash cows, which are allowing it to invest aggressively in higher-growth initiatives.
In addition to its AI ambitions, Intel has a chance to become the world's No. 2 foundry by the end of the decade. The company is spending $20 billion on two chip-fab facilities in Ohio, which are set to open in 2024, and it intends to build a $33 billion chip plant in Germany. A third of the cost of this latter plant is being covered by the German government.
Lastly, Intel's valuation makes a lot of sense following a significant pullback over the past two-and-a-half years. With Wall Street forecasting well over $3 per share in earnings in 2026, Intel's needle looks to be pointing decisively higher.
Image source: Getty Images.
The artificial intelligence stock that appears poised to tumble: Nvidia
However, not all artificial intelligence stocks will necessarily reward their shareholders. After more than tripling in 2023 on the heels of AI hype, GPU kingpin Nvidia (NVDA 6.18%) is the AI stock that appears most likely to tumble in the not-too-distant future.
For the moment, Nvidia is on top of the AI world. The company's A100 and H100 GPUs are the infrastructure backbone of AI-driven enterprise data centers. According to estimates, Nvidia accounts for around 90% of high-compute GPUs currently in use in enterprise data centers.
Nvidia has also seen its revenue and profit forecasts soar as demand for its GPUs has picked up. Heading into fiscal 2024 (Nvidia's fiscal year begins in late January or early February), Wall Street had been looking for high-single-digit to low-double-digit sales growth from Nvidia. Given the strength of demand for its AI-inspired GPUs, the expectation is now for the company's net sales to more than double in fiscal 2024.
While there are certainly catalysts that explain why Nvidia has outperformed this year, a host of headwinds could completely derail its momentum in calendar year 2024.
To start with, Nvidia is set to face a widening field of competitors. AMD recently unveiled its MI300X AI-inspired GPU, which looks to be the first major competitor to Nvidia's A100 and H100 GPUs. Though AMD is delivering the MI300X to a handful of customers this year, the plan is for a widespread rollout in 2024. As noted, Intel's Falcon Shores GPU geared at AI-accelerated data centers should hit the market in 2025.
What could really make matters more challenging for Nvidia is that it's set to lose the bulk of its pricing power as it ramps up sales of its A100 and H100 GPUs. With world-leading chip-fab company Taiwan Semiconductor Manufacturing increasing its chip on wafer on substrate capacity, less scarcity of Nvidia's leading AI GPUs should cause Nvidia's pricing power to plunge. This may quickly reverse the gross margin gains Nvidia has enjoyed during the first-half of fiscal 2024.
As I pointed out earlier, every next-big-thing investment over the past three decades has gone through an initial bubble period, and it's highly unlikely that artificial intelligence will be the exception. Whereas well-diversified businesses like Intel would be minimally affected if the AI bubble bursts and the technology takes years to mature, Nvidia's stock could be crushed.
The final concern is Nvidia's valuation. Despite a reasonably low forward price-to-earnings ratio of 29, Nvidia is trading at a historically high multiple relative to its cash flow. History suggests that mammoth gains of the magnitude that Nvidia's shares have experienced this year are rarely, if ever, sustainable.
