The surest sign there’s an generative artificial intelligence bubble is Super Micro Computer (NASDAQ:SMCI) stock, up 242% so far this year. It’s worth about $60 billion with 2023 sales of $9/25 billion. It trades at 76 times earnings.
SMCI doesn’t write AI software. It doesn’t make AI chips. Its business is assembling servers. It’s taking other companies’ gear and installing it in racks, with relevant software. This is a low margin business, and even in 2023 SMCI saw just 8% of its revenue, $732 million, become net income. Yet here we are.
Why Super Micro Stock?
Super Micro’s offices are just a few miles from those of Nvidia (NASDAQ:NVDA). Most of the servers SMCI sells carry Nvidia chips and software. CEO Charles Liang is very close to Nvidia CEO Jensen Huang, a fellow Taiwanese-American.
The product mix is why SMCI is worth so much more, based on sales and earnings, than rivals like Dell Technologies (NASDAQ:DELL) and HP Enterprise (NYSE:HPE). Everything they do is considered high margin. Business is exploding. (Super Micro is now worth more than HP Enterprise and HP (NYSE:HPQ) itself put together.)
Analysts expect 50% growth in SMCI sales this year, with earnings ballooning to $22 per share. But even if it hits that figure, you’re paying 44 times earnings to own it.
If SMCI had a huge moat, if no one else were doing what it does, that might be a fair price. But other companies can do what SMCI does. It’s mainly SMCI’s product mix, and Fear of Missing Out (FOMO) on the profits from that advantage, that hold up the stock price.
Yet plenty of analysts are still pounding the table for it. It’s called an AI Compounder. It’s what to buy if you missed out on Nvidia. Bank of America (NYSE:BAC) is now on the SMCI train.
Buy high, sell low?
The Limits on SMCI Stock
I don’t doubt analyst expectations on SMCI for this year, or even for next year.
I just know that there’s a limit.
So does management. Right after the stock fell 6% on joining the S&P 500 (buy the rumor, sell the news), the company filed to sell 2 million new shares. The stock lost 9% of its value after the announcement.
The problem is that moat. Nvidia has a wide moat because it controls the hardware and software considered standard for building AI applications. The Cloud Czars have fat moats because they’ll host the applications and have consumer apps of their own.
Super Micro’s moat, if one exists, is that it is quick to market. It’s basically an assembly business. That’s not much of a moat.
Besides, Super Micro has gotten valuable, fast. When ChatGPT was announced in November 2022, Super Micro was worth about $5 billion. Can superfast assembly capabilities and quick turnaround turn $60 billion into $600 billion?
I doubt it.
The Bottom Line
Super Micro is the “canary in the coal mine” of the GenAI boom.
Its advantages will disappear before those of other companies in the GenAI space, because its moat is thinner and the company itself is smaller. The company has just over 4,000 employees.
That doesn’t mean Super Micro is about to fall immediately. The AI Infrastructure boom has legs, and Super Micro is well positioned for it. But at some time in the next two years, we’re going to stop focusing on infrastructure and start focusing on software. We’re going to be looking for revenue and earnings from AI applications.
That’s when you’ll want to be out of SMCI stock. Software scales in ways that hardware can’t.
As of this writing, Dana Blankenhorn had a LONG position in NVDA. The opinions expressed in this article are those of the writer, subject to the InvestorPlace.com Publishing Guidelines.