Intel just posted its fastest revenue growth in 15 years and beat earnings by almost two times, with strong guidance. The stock still sold off, down about 3.3% on the day. Part of that is a weak day for tech overall.
The bear case: the story is played out
I would not start a trade in Intel now. The stock has had a great year, but over the last 3 years earnings have been flat to negative, basically no growth. When an old-line name like Intel, a Silicon Valley giant for decades, makes a move that big, most of the big money has already been made. At best it is a hold, and probably a sell. The one exception would be getting it at a stupidly cheap price if you truly believe in it for the long haul.
There is no dividend, so the real question is: why buy this stock now? The answer is you shouldn't yet. The market runs on "what have you done for me lately." No one would have predicted how the stock caught fire from around March 30th or April 1st, and that trade got missed. To turn more positive, I need to see consistent earnings, not one strong quarter after three flat years.
The bull case: extend your time horizon
In the short run, agreement holds. When a stock is up that much, trees don't grow to the sky. Over a longer view the picture changes. Intel's plant in the US is not online yet and will eventually come online. The AI spend is real.
The data center business grew 59% last quarter. One quarter is not enough to prove Intel is a meaningful AI infrastructure player, but for the long haul it looks good. If you already own Intel, keep holding it. I own it personally and at my firm. I might not add a big position, though on more weakness I probably would buy. Over the next 3 to 5 years Intel can beat market returns. You will not triple your money in 5 years, but there is still room to the upside.
Spenders versus receivers
The AI trade is far wider than chips. There is a way to think about it as big spenders versus receivers. The spenders are the hyperscalers, and they are not slowing down. The receivers provide everything inside the data center: chips, hardware, software, networking, cybersecurity, and cooling. Intel is part of that receiver group.
The hyperscalers will spend the money. What is unclear is whether they will earn the return people expect. So the smarter move in this environment is to shift money away from the hyperscalers and toward the companies that supply the gear that makes AI work, and let the spenders prove later whether they can make money on this build-out.
Several years ago an ETF was built for exactly this, ticker TRFK, holding everything inside the data center, hardware, software, networking, cybersecurity, and cooling. Another firm owns everything in a data center across several ETFs, from chips to hardware to software to networking to cooling.
Overlooked corners of the build-out
Cybersecurity gets missed. These sites need to be cyber fortresses.
Cooling is the bigger blind spot. Almost half the energy cost in a data center goes to cooling. Charge an iPhone and it feels warm afterward; now picture a building three football fields long packed with rack after rack of servers and the heat that produces. These machines must stay cool, and few people have caught on to cooling as the real bottleneck. Vertiv is one name here. Comfort Systems, an HVAC company, is another good play in the AI data center space. HVAC as an AI play would have drawn laughs 3 or 4 years ago.
Preferred chip name
In data center chips, AMD stands out. It keeps racking up one major win after another in AI, landing Oracle, Anthropic, OpenAI, and Microsoft. Basically everyone big in AI is beating a path to AMD's door.
The broader pattern
Intel fits a theme running through this earnings week and likely the whole season: spend now, prove it later. Alphabet and Tesla drew similar reactions. The check writers and the check cashers have not converged this quarter, and it doesn't look like they will.
