The artificial intelligence (AI) market is full of hype and speculation, with early AI leaders like Nvidia or Microsoft trading at nosebleed valuations. Whether the AI market darlings deserve their rich valuations or not, three Fool.com contributors believe it's time to look beyond the usual suspects.
They put their heads together to uncover existing tech giants like International Business Machines (IBM -1.05%), Cloudflare (NET 1.44%), and Salesforce (CRM 0.42%) -- three great companies with deep AI expertise. This trio offers compelling value and potentially significant long-term returns from their AI initiatives.
Investing is a marathon. IBM, Salesforce, and Cloudflare have been running the AI race for years, and they won't stop anytime soon.
IBM is the big, blue AI workhorse
Anders Bylund (IBM): Generative AI tools and flashy consumer tech often dominate the headlines. Meanwhile, good old tech titan IBM is quietly positioning itself as the backbone of enterprise-grade AI solutions.
The company's hybrid cloud focus, emphasis on explainability and trackable data sources, and deep industry relationships might not seem as sexy as the latest ChatGPT upgrade, but these are the very elements that should make IBM an enduring force in the AI space.
IBM's focus on long-term partnerships and sustainable solutions aligns with the needs of businesses wanting to use AI in an impactful yet responsible way. As concern grows around AI systems' potential for harmful biases, misleading output, and black-box decision-making, IBM's approach inspires a deeper trust.
The long-term contracts Big Blue signs with enterprise-class customers will depend on that professional-grade trust. "Hallucinations" and data mishaps are common themes in today's large language models (LLMs), but IBM isn't interested in playing that game. The likes of OpenAI, Microsoft, and Alphabet don't seem ready to compete from the transparency and quality angle so far.
Critics might argue that IBM's "slow and steady" approach is a disadvantage. But in the realm of enterprise solutions, patience and thoroughness should be seen as strengths. The slow and heavy vetting IBM's prospective clients apply to the Watson AI systems today will result in lucrative, hard-to-replace, multiyear contracts.
"The early work for clients around data architecture, security, and governance is critical and hard, and we think consulting expertise is going to be crucial here," CEO Arvind Krishna said in the recent fourth-quarter earnings call.
Moreover, IBM's current valuation presents a potential value opportunity compared to some of the AI market's highfliers. The stock trades at a modest 2.7 times sales and 13 times free cash flow, making even Alphabet look expensive in comparison with the same ratios clocking in at 5.6 and 25, respectively.
If the company can continue to prove that its AI solutions add tangible, measurable value to businesses, investors who focus on long-term potential may find IBM to be the unsung workhorse of the AI revolution. If nothing else, you're looking at a generous dividend payer with torrential cash flows and a centennial business history.
IBM's big AI bet is starting to pay off right now, but even if the business-grade AI strategy fizzles, it's still a great income-generating stock to own in the long run. That's a big "if," though. AI and hybrid cloud computing have emerged as leading growth drivers for the ultra-experienced tech giant. I expect good things and big wins from Big Blue's AI business in the coming years.
Could this edge computing play be an under-the-radar AI star?
Billy Duberstein (Cloudflare): The past 18 months in artificial intelligence have been all about training, which is the act of building extremely large language models out of billions of data parameters. But now the AI industry is moving to an "if you build it, they will come" phase. In layman's terms, companies are now actually putting these models to work.
When a trained model is prompted and then goes to work on finding answers or completing a task, that's called inferencing. And inferencing applications are about to take off.
To see how that shift is unfolding, look no further than Nvidia's recent earnings report. CEO Jensen Huang said he estimates about 40% of Nvidia's data center revenue over the past year was actually for inference, not training. That's somewhat surprising, as Nvidia is thought to be the only game in town for training, whereas it's possible inferencing could be done with some lower-power processors from competitors.
One company that could be an under-the-radar inference play is Cloudflare, which began as a software company that helped speed up websites, software applications, and content delivery.
But Cloudflare's management was forward-thinking, taking the time and cost to build an integrated software stack from the beginning, whereby all of its new products and services can be run on the same type of commodity hardware. That has paid off in the form of an ever-increasing portfolio of products, including cybersecurity, network management, and developer platforms that run on its lightweight servers in internet service provider data centers close to customers at the edge.
Today, Cloudflare's servers are in 310 cities in over 120 countries connecting 13,000 networks, with its network now within 50 milliseconds of 95% of the world's population. And that makes Cloudflare a potentially ideal place for businesses to run their inference workloads. That's because inference workloads are likely too computing-intense to run on end devices such as PCs or phones, but are likely "lightweight" enough so that they don't have to be run all the way back in a large, cloud-based data center.
Management began to capitalize on this potential last September when Cloudflare released Workers AI, an inferencing-as-a-service platform for developers in Cloudflare's servers, which could potentially put Cloudflare up against the big cloud computing players for inference workloads.
And the new offering is showing early signs of success, with Workers AI requests growing 9x between September and December, according to Cloudflare's recent fourth-quarter conference call with analysts. And even though the new inference offering just came out and contributes little to revenues today, Cloudflare still saw a very healthy 32% revenue growth in the last quarter, driven by its core products. Of note, Cloudflare is really penetrating larger enterprise customers now, with the company recording its largest customer win, as well as its largest customer renewal ever last quarter.
With much-loved offerings across its core content delivery, network, and cybersecurity markets that are still growing strong, Cloudflare's new inference products could bolster its already-strong growth rate for years to come.
A transition from growth to value
Nicholas Rossolillo (Salesforce): After a couple of years of unwinding out-of-control spending that cropped up during the pandemic, customer data and relationship management platform Salesforce (CRM 0.42%) is totally back. The software giant reported revenue growth of 11% this past year, despite a big slowdown in corporate spending as companies looked to tighten up as the economy slowed. Salesforce is predicting some of this soft spending among its users will continue, but is still nevertheless predicting revenue growth of about 9% for 2024 (fiscal year 2025 for Salesforce).
As the business has matured, Salesforce made a quick pivot from all-out growth (it had been putting up 20%-plus sales growth for the duration of its existence as a publicly traded stock over the last two decades) to a more value-generating business. CEO Marc Benioff and the top team ratcheted up generally accepted accounting principles (GAAP) operating profit margins this last year to 14.4%, and expects that figure to jump to 20.4% for full-year fiscal 2025.
Data by YCharts.
Along the way, Salesforce started repurchasing its own stock as part of a new strategy to begin returning cash to shareholders. And starting this year, Salesforce will also begin paying a quarterly dividend -- which, as of this writing, is worth 0.5% in annualized yield. That's not too exciting, but it nevertheless signals this data management software company's successful evolution into a more mature business.
Of course, customer data is more important than ever in a new era of AI-fueled computing technology. Salesforce's own development of AI throughout its platform isn't translating into any pick-up in growth just yet. However, that tees up the possibility of management underpromising and overdelivering this year on financial guidance. Plus, Benioff and company said its own internal use of its AI tools should fuel the profit margin expansion story for years to come.
Salesforce stock now trades for about 45 times expected GAAP earnings per share (EPS), or about 30 times EPS on an adjusted basis (which corresponds a bit more closely to free cash flow). Shares have been on a hot run in the last year, but there are still years' worth of profitable returns left in the tank for this cloud software leader. I'm happy to keep holding.
