There's no denying that Nvidia (NVDA 3.46%) is on a roll. The chip designer has knocked every earnings report out of the park since OpenAI unveiled the ChatGPT artificial intelligence (AI) tool, powered by the company's AI accelerator chips.
Trailing sales are up by 126% in just five quarters. Free cash flow soared 610% higher over the same period, and Nvidia's stock price more than quadrupled.
Golf claps and finger snaps aplenty, please. Nvidia has made many investors a lot of money in this AI boom, including yours truly.
But this isn't the right time to buy more Nvidia shares. The stock is trading at incredibly lofty valuation ratios as the company's investors bet it will continue to dominate the AI hardware arena for years to come. Since I'm not so sure about that, I recently took some Nvidia profits off the table to reinvest in some lower-priced but equally exciting growth stocks.
You may not want to sell Nvidia stock today. In any case, I wouldn't recommend buying more of it right now.
Due to Nvidia's inflated share price, I suggest you steer clear of buying this stock until it cools down a bit. But if you insist on buying AI stocks right now, take a closer look at digital design software veteran Autodesk (ADSK 2.50%) and semiconductor pioneer Intel (INTC 1.28%).
Autodesk's AI-driven strategy
Autodesk is leaning into the AI opportunity. The company is already a leader in generative AI tools, particularly in the field of professional-quality 3D images.
"We can already generate 3D representations from images 10 times faster and with vastly higher quality than currently available 3D AI," CEO Andrew Anagnost said in February's fourth-quarter earnings report.
To build on this industry-leading expertise, the company is rebuilding its core products around a shared, AI-driven design engine. The master plan is all about product lifecycle management, where a computer-aided design (CAD) can move from one team to another within each client's operating structure without running into red tape and speed bumps. At the same time, software improvements and new features created to serve one target market can easily find use cases across other customers and project types.
Again, Autodesk relies on high-quality machine learning and generative AI tools to deliver these planned efficiencies.
These are early days in a "multiyear process," and the upgrade path hasn't been smooth so far. Sales are growing but not skyrocketing. Free cash flow is down in recent quarters -- and Autodesk has delayed the filing of year-end financial documents, as auditors found inconsistencies in the fourth-quarter's cash-flow calculations.
As a result, investors took a step back from the late paperwork issue, and Autodesk's stock is down 17% in the last month. It trades at 8.6x trailing sales and 36x free cash flow. That's a bargain next to Nvidia, with respective ratios of 33.8 and 76, respectively.
Yes, Autodesk is growing slower than Nvidia right now, and I don't love the potentially flawed financial reporting. Still, I'm downright excited about the company's long-term business prospects as the modular, AI-powered software platform evolves.
I bought my first Autodesk shares two years ago, as the industry veteran's stock looked undervalued from a long-term perspective. After a few wild swings, the stock price is back where it was in April 2022, but trailing sales are up by 20% and the AI catalyst should add more value over time.
Intel's strategic shift toward AI and chipmaking services
Intel's business model is also transforming. The company is still the leading name in PC and server processors, but Advanced Micro Devices is gaining ground in both markets. On the other hand, Intel taps into the AI boom from several angles, and the recently launched processor manufacturing business doubled its revenue in 2023.
This ain't your grandfather's Chipzilla anymore. Some investors aren't comfortable with Intel's shifting strategy, expecting the company to defend its traditional processor business at all costs. But I would much rather own stock in a company willing to change with the times than watch an outdated titan sticking to its old guns. Butch Cassidy and the Sundance Kid doesn't have a happy ending, y'know?
The Intel Gaudi 2 may not be the fastest AI accelerator in the world, but AI system builders often don't look for maximum performance per chip. Intel delivers where it counts.
"Our Gaudi 2 AI accelerators continue to demonstrate price performance leadership compared to the most popular GPUs," CEO Pat Gelsinger said in January's earnings call. "In a recent blog published by Databricks, Gaudi 2 was shown to clearly deliver the best training and inference performance per dollar based on public cloud pricing."
Moreover, the next-generation Gaudi 3 accelerator will launch later this year with quadruple the processing power and double the networking speed of the current chip.
Image source: Getty Images.
The Xeon line of server processors stands front and center in the AI systems market, too, since the supercomputers that train the beefiest large language models (LLMs) use these chips to control each accelerator module's data-crunching activities. Five of the 10 fastest supercomputers in the world today rely on thousands of Intel Xeon processors. The largest one, Argonne National Laboratory's Aurora system, even uses Intel GPUs instead of Nvidia's or AMD's. Those chip sales add up in a hurry when each processor costs at least $11,600.
The new and improved Intel is a serious AI competitor, with important side gigs in traditional processor sales and semiconductor manufacturing. Intel is currently consuming cash due to costly upgrades and new construction of chipmaking facilities. But the stock trades at just 2.7x sales, more than 30% below its 52-week highs.
Investors looking for a promising AI strategy and a flexible operating model might find Intel an intriguing choice, especially at its current valuation. And that's why I'd much rather buy Intel and Autodesk shares than double down on my Nvidia position right now.
