The heavy-duty computation required to train the most advanced artificial intelligence (AI) models and deliver those models as services is mostly being done in cloud data centers. Cloud giants like Amazon Web Services, Alphabet's Google Cloud, and Microsoft Azure are ramping up their AI offerings, and in some cases, designing their own AI chips.

While the cloud giants will win plenty of AI workloads, smaller players are having success, as well. Oracle (ORCL 2.25%) disappointed investors this week with its quarterly report but is bringing generative AI workloads onto its cloud platform at a blistering rate. And DigitalOcean (DOCN 2.43%), which is going through a management upheaval, should play well with developers looking for simplicity with its recent AI acquisition 

Oracle

While Oracle's cloud infrastructure business is still dwarfed by the biggest players, the company is clearly gaining traction. Infrastructure-as-a-service revenue rose 66% year over year to $1.5 billion in Oracle's first quarter, which ended Aug. 31. The growth rate has slowed a bit, but Oracle is still growing faster than its competitors.

One way that Oracle has attempted to differentiate itself has been a strong focus on making its cloud platform ideal for generative AI workloads. That strategy appears to be paying off. Oracle offers Nvidia GPUs through its cloud platform, and Chairman Larry Ellison claims that Oracle customers can train AI models twice as fast and at less than half the cost of other cloud platforms.

Ellison has a tendency to exaggerate, but the proof is in the pudding. Oracle had signed contracts for more than $4 billion of AI training capacity at the end of the first quarter. That number more than doubled since the end of the previous quarter.

Oracle's cloud buildout to support AI workloads is an expensive endeavor. Being primarily a software company, Oracle has historically not spent much on capital investments. Capital spending reached $8.7 billion in fiscal 2023, nearly double the previous year, and the company expects capital spending in the current fiscal year to be approximately the same.

Even with investments in cloud infrastructure and AI rising, Oracle continues to grow its bottom line. Adjusted earnings per share rose 16% year over year in the first quarter, while free cash flow shot up 21%. Cost-cutting has helped the cause. Oracle's total GAAP expenses were up less than 4% year over year in the first quarter, compared to 9% revenue growth.

Oracle is still a small player in the cloud infrastructure market, but the company's focus on AI could help it grow faster than its peers for years to come.

DigitalOcean

Up until recently, DigitalOcean's cloud platform offered nothing for customers looking to train or run AI models. That changed with the acquisition of Paperspace, an AI cloud infrastructure start-up that aims to keep things simple.

DigitalOcean paid $111 million for Paperspace, which has built an easy-to-use platform that enables developers to train AI models and deploy them to scalable API endpoints. DigitalOcean has made a series of acquisitions over the past few years, including managed hosting provider Cloudways and backup solutions provider SnapShooter, and a common theme has been a focus on simplicity. Paperspace certainly fits into that mold.

DigitalOcean could have built its own GPU-powered AI platform, but it would have taken too long in a market that's moving incredibly fast. Using Paperspace as a foundation, DigitalOcean can invest in expanding its AI capacity and win customers it would have otherwise lost to competitors.

DigitalOcean is funding these investments with its impressive free cash flow. The company expects to convert as much as 22% of revenue into free cash flow this year and is targeting a mid-to-high 20s percentage in the long run. Paperspace will reduce free-cash-flow margins by a few percentage points as DigitalOcean snaps up expensive GPUs, but the company can easily afford to take the hit.

One thing to note: DigitalOcean suddenly and unexpectedly announced last month that CEO Yancey Spruill would step down once a successor was found. This could be a sign that the core business is suffering a deeper downturn than expected or that some of the company's acquisitions aren't living up to their promise. Regardless, it's something to be aware of.

With the Paperspace acquisition, DigitalOcean is going after customers who want a simple platform for AI training and inference. While it won't win many enterprise customers, it fits in well with the company's overall strategy to target developers and small businesses.