Just caught the aftermath of PSIX getting absolutely hammered yesterday - down 27.7% after earnings dropped. Wild move, but there's actually something interesting happening under the surface here if you dig into the numbers.

So Power Solutions is basically pivoting hard from legacy industrial and transportation engine stuff into the data center power systems space. That's a smart move given where all the AI capex is flowing right now. Revenue jumped 32.5% to $191.2M in Q4, which is solid. But here's where it gets messy - gross margins cratered from 29.9% year-over-year down to 21.9%. That's the kind of margin compression that makes investors nervous, even when top line growth looks good.

Management's explanation was pretty straightforward though. They're ramping up production way faster for data center products, and that's creating operational inefficiencies. Basically, they're scaling too quick and it's eating into profitability right now. Classic growing pains situation. On top of that, they just acquired MTL Manufacturing & Equipment to fill some tech gaps - switchgear subbases, electrical enclosures, fuel tanks, that kind of stuff. It makes sense strategically for rounding out their data center product suite, but investors apparently didn't love the timing given everything else going on.

The tax situation also played a role in the earnings miss. EPS actually dropped 31% to $0.71, but that was mostly because they're now a full taxpayer after using up prior-year loss carryforwards. So the real operational picture isn't quite as bad as the headline numbers suggest, though the margin squeeze is definitely real.

What caught my attention though is the valuation now. Stock's trading at like 13.9x forward earnings for 2026 based on analyst estimates of $4.46 EPS. That's pretty cheap compared to most other AI-adjacent industrial plays. And the thing is, the data center buildout isn't slowing down anytime soon. All the big tech companies are still throwing massive capex at AI infrastructure, so demand for power solutions should stay strong.

Management didn't want to give specific guidance, which obviously spooked people, but they did say they expect continued full year sales growth in 2026 with moderate margin improvement. They mentioned some headwinds from oil and gas, which makes sense as they're transitioning away from that segment. But honestly, the data center tailwind should more than offset that.

I get why the market sold off - margin compression is never fun to see, and the lack of specific guidance feels evasive. Plus you had the broader AI and data center stock selloff happening anyway with fears about energy costs and rates. But stepping back, this looks like a company successfully executing a transition to a higher-growth market, just with some near-term pain on profitability as they scale. The kind of thing that could look pretty smart in hindsight if they nail the execution.

For small-cap value investors looking to get exposure to the AI infrastructure build, this might actually be worth considering on the dip. The valuation's attractive, the market opportunity is real, and the business fundamentals are moving in the right direction even if the margins are taking a hit. Not saying it's a slam dunk - execution risk is real, especially with the margin pressure - but it's definitely more interesting at these levels than it was before the selloff. Definitely keeping an eye on how the data center ramp progresses over the next couple quarters.