2 Artificial Intelligence (AI) Stocks to Buy for $1,000 and Hold for Decades

Last week, Oracle Chairman Larry Ellison gave investors some new insights into the current state of artificial intelligence (AI). He said he sees no slowdown in business spending on AI development. Indeed, he believes the industry will expand significantly over at least the next 10 years.

Oracle currently has 162 data centers online and under construction, but Ellison believes that number could reach 2,000 in the long term.

If he’s right, this could be a fantastic time to buy AI stock, especially after the recent selloff in the tech sector. Investors with $1,000 to put to work now might want to split it equally between shares of the chip giant Micron technology (MU 4.60%) and cloud computing company DigitalOcean (Doc 0.40%).

1. Micron Technology: A leader in memory chips for AI

The AI ​​is trained and powered by data center servers that use powerful graphics processing units (GPUs), most of which have been provided by Nvidia. Memory chips are complementary to GPUs because they store information in a ready state so it can be called up instantly, which is especially critical in data-intensive AI workloads.

Micron’s HBM3e has been selected by Nvidia to be included in its new H200 GPU. That high-bandwidth memory architecture offers several benefits over previous generations of chips, including higher capacity and a smaller physical footprint. In addition, Micron’s HBM3e is 30% more energy efficient than competing chips, and power consumption is a key consideration for data center operators when selecting hardware.

Micron recently told investors that it has already sold all the data center memory it could ship in 2024 and 2025, which isn’t surprising given that demand for Nvidia GPUs significantly outstrips supply. Some industry sources said that Micron is preparing to produce a new 12-layer HBM3e solution, which implies a 50% increase in capacity, which will be ideal for Nvidia’s upcoming Blackwell-based GPUs, such as B200. Therefore, investors should not expect demand for Micron’s high-bandwidth memory to slow down anytime soon.

However, Micron’s opportunities transcend the data center, as some of the hardware to power AI applications is also making its way to PCs and personal devices, thanks to increasingly powerful neural processing units (NPUs) from the likes of Apple and Advanced microdevices. Micron says AI smartphones, for example, require up to twice as much memory capacity as their predecessors. This will be a direct tailwind for the company’s earnings.

In the third fiscal quarter of 2024, which ended May 30, Micron saw an 85% increase in revenue from its data center and a staggering 94% increase in revenue from its mobile segment, driven by AI demand. The company will report its final results for fiscal 2024 at the end of September, and analysts expect total revenue of $25 billion, up 61% from fiscal 2023. Wall Street’s early predictions for fiscal 2025 points to $38.8 billion in revenue, representing further growth of 53%.

But it gets better. Analysts forecast that Micron could post earnings of $9.41 per share in fiscal 2025, placing its stock at a forward price-to-earnings (P/E) ratio of just 9.6. That means Micron stock would have to triple to trade in line with Nvidia stock, which currently has a forward P/E ratio of 29.1. Given how closely related the two companies are, Micron looks like an excellent value at the current price for investors looking to hold onto the stock for the long term.

2. DigitalOcean: An AI cloud provider for small and medium businesses

Returning to Ellison’s comments, the race to build data centers is fueled by growing demand for cloud AI services. Suppliers like Amazon Web Services and Microsoft Azure is the leader in this space, but primarily targets large organizations with deep pockets. DigitalOcean, on the other hand, has carved out a profitable niche serving small and medium-sized enterprises (SMEs).

DigitalOcean’s initial success came from a portfolio of cloud services such as data storage, web hosting, video streaming and software development tools. It offers cheap and transparent pricing, a simple dashboard with one-click deployment features, and personalized support, all of which are important for SMBs.

However, last year DigitalOcean acquired Paperspace, which operates multiple data centers with a range of GPU options designed for AI developers. DigitalOcean recently announced that it will allow customers to access fractional GPU capacity on demand, which has never been done before. In other words, SMBs can use between one and eight Nvidia H100 GPUs to power their AI workloads, a small scale that cloud leaders simply won’t offer because it’s not cost-effective for them .

We already know that Paperspace can be up to 70% cheaper than equivalent services from Microsoft Azure, for example, because it offers per-second billing without lock-in contracts. In addition, the company has a low cost structure due to its narrow portfolio of services, so customers are not paying for things they may never use. With these services combined with DigitalOcean’s new fractional GPU offering, even the smallest companies can start implementing AI.

DigitalOcean generated revenue of $192.5 million in the second quarter, a 13% increase over the prior-year period. However, its AI-attributable revenue grew by 200%, suggesting that SMBs are latching onto the company’s new services at a rapid pace.

The stock is now trading at a price-to-sales ratio of 5.1, which is 41% below its average of 8.7 since the company went public in 2021, so it looks like a good value right now. As one of the only cloud providers that truly focuses on the needs of SMBs, DigitalOcean could be a great addition to any portfolio as the AI ​​opportunity expands over the long term.