Analysts at Goldman Sachs expect big tech giants to spend over half a trillion dollars on AI-related capital expenditure in 2026. And much of that money is going to data center hardware like AI accelerator chips, high-bandwidth memory devices, and various types of networking equipment.
The eye-popping level of hardware spending contrasts with low profits and often huge losses on the consumer-facing software side of the industry. This dynamic means investors can maximize returns by betting on pick-and-shovel providers like Micron Technology (MU +4.87%) and Broadcom (AVGO 0.74%).
Let's discuss why these potential millionaire makers could make great picks in February and beyond.
NASDAQ: MU
Key Data Points
Micron Technology
With shares up by over 300% in the last 12 months, Micron Technology's bull run is already in full swing. Investors are getting more excited about the memory hardware specialist as ravenous AI needs send demand for its hardware soaring. With a rock-bottom valuation and memory shortages projected to continue for the next few years, the company's long-term rally might be just getting started.
While Micron's stock is doing great right now, it wasn't always like this. In fact, the stock is known for generating practically no growth between the dot-com bubble in 2000 and the year 2020. Shares have historically underperformed because memory hardware is commoditized. Micron's products are not well differentiated from chips created by its rivals, which can lead to substantial price competition and boom-and-bust cycles as production capacity rises to meet demand.
Generative AI has shaken up this old paradigm by sending memory demand up much faster than supply can keep up. Micron is already benefiting from this trend, with fiscal first-quarter revenue jumping 57% year over year to $13.6 billion -- driven mainly by purchases from cloud data centers. With memory shortages expected to continue until 2027, Micron is poised to enjoy a near-term windfall, which it could use to reinvest in its business or return cash to shareholders through stock buybacks.
With a forward price-to-earnings (P/E) multiple of just 12, Micron stock trades at a dramatic discount to AI hardware alternatives like Nvidia, whose forward P/E is 22. The low price tag suggests Micron's stock still has room for continued growth.
Broadcom
As mentioned earlier, the consumer-facing software side of generative AI is generally not profitable, with industry leaders like ChatGPT and Anthropic expected to burn tens of billions in 2026 alone. This dynamic can be partially blamed on the high cost of Nvidia graphics processing units (GPUs) and other types of compute hardware. Broadcom's application-specific integrated circuits (custom chips) offer an alternative.
Custom chips allow Broadcom's customers to use hardware optimized for their specific workloads, without features they don't need. This can make them cheaper to buy and operate compared to general-purpose hardware. Recently, Broadcom has secured several high-profile contracts to work with AI leaders.
Image source: Getty Images.
In October, the company announced a strategic partnership with ChatGPT creator OpenAI to deploy 10 gigawatts of OpenAI-designed accelerators. The company is also a major partner of industry leader Alphabet's Google, which designs a proprietary AI chip called the Tensor Processing Unit (TPU) to compete with Nvidia's GPUs. The market for custom chips looks set to continue expanding as more clients aim to cut costs and make their AI efforts profitable.
Broadcom's fourth-quarter revenue jumped 28% year over year to $18 bilion, driven by strength in its AI semiconductor revenue, which soared 74% because of custom chip demand along with demand for Ethernet AI switches. With a forward P/E multiple of 31, Broadcom's shares trade at a premium over the market average. But this looks reasonable, considering the company's relatively safe pick-and-shovel business model and potential to capture market share from Nvidia as the AI industry matures.
