Wall Street Is Missing the Forest for the Trees: Why I'm Still Buying Super Micro Computer
Super Micro Computer $SMCI has been a battleground stock in 2024 revered by AI bulls, doubted by valuation hawks, and whiplashed by shifting expectations. But here’s the thing: while Wall Street seems overly focused on short-term turbulence, I believe the bigger picture remains intact. In fact, the market may be mispricing what could be one of the most critical AI infrastructure companies of this decade.
Let me walk you through why I’m still bullish and why I’m holding on through the noise.
A Company Built for the AI Boom
$SMCI isn’t your average server company. It’s a hardware architect built for the AI era designing high-performance, energy-efficient systems optimized for large-scale inferencing and training workloads. Roughly 95% of its revenue comes from its Server and Storage Systems segment, which includes the AI-optimized racks that hyperscalers and enterprises need to run Nvidia’s $NVDA H100s, B100s, and beyond.
Yes, it’s capital-intensive. Yes, it’s reliant on chipmakers like Nvidia. But that’s the nature of selling picks and shovels during an AI gold rush. The question is: can Super Micro scale profitably, sustain momentum, and navigate the bumps ahead?
What the Market Is Overreacting To
The recent sell-off tells me the market is reacting more to headlines than to fundamentals. Let’s dissect what actually happened:
Guidance revisions: The revenue guide for FY2025 has come down three times from $30B to as low as $21.8B. On the surface, that looks alarming. But dig deeper: even at the low end, this implies 45%-plus revenue growth year-over-year. That’s not a collapse. That’s still hyperscale expansion.
Delayed shipments: Much of this guidance cut is due to customer pause ahead of Nvidia’s Blackwell launch. This is not a loss of demand it’s a deferral. Clients aren’t cancelling orders; they’re waiting for the next-gen GPU. When those chips start flowing, SMCI will be ready to rack them.
Nasdaq compliance drama: A late 10-Q filing led to governance fears, and shares dipped. But SMCI has since confirmed it’s back in compliance. I see this as a technical hiccup, not a red flag.
Tariffs and macro noise: A 10% tariff on Chinese imports could cost $SMCI ~$200–300M annually. That’s a hit, but not fatal. Every peer in the sector faces the same headwind.
In short: these are timing issues, not structural breakdowns.
Still a Growth Powerhouse (Even If the Street Doesn’t See It)
From FY2019 to FY2024, SMCI has grown revenue at a 27% CAGR, right in line with the AI infrastructure industry’s average. It’s not just riding Nvidia’s coattails it’s building customized systems faster, cheaper, and more flexibly than legacy players like HPE, Dell, or Lenovo.
While cash from operations (CFO) dipped in FY2024 due to inventory buildup and customer payment delays, the company bounced back sharply with $626.8M in operating cash flow in Q1 FY2025 alone. That shows me this isn’t a liquidity crunch. It’s working capital cycling through a high-growth business.
Yes, debt is up convertible notes surged ~661% last year but I’m not overly concerned. The interest coverage ratio sits around 28x, and the Debt/Equity ratio remains manageable at ~42%. They’re borrowing for expansion, not survival.
Valuation: Yes, It’s Not Cheap But It’s Not Crazy Either
On a trailing P/E basis, $SMCI might look expensive. But forward-looking metrics like the PEG ratio (price/earnings to growth) suggest the valuation is actually reasonable relative to future earnings power.
Here's where it gets more compelling:
Return on Invested Capital (ROIC) remains high showing me this company is putting capital to work efficiently.
Net income, EBITDA, and revenue growth continue to outpace many traditional OEM peers.
EV/EBITDA is higher than peers, but again, that’s the cost of being a first-mover in a hypergrowth market.
If you're purely a value investor looking for GAAP perfection, this might not be your stock. But if you’re looking for asymmetric upside tied to long-term AI infrastructure demand this is exactly where you want to be early.
The Risks Are Real, But Manageable
Let’s not sugarcoat it. SMCI isn’t without risk:
It’s a margin-thin business.
It doesn’t control the semiconductors it depends on (mainly from Nvidia).
Tariffs, delayed chip shipments, and demand lumpiness will continue to cause volatility.
Operating leverage cuts both ways just look at the decline in gross and net margins during FY2024.
But these are execution risks, not existential risks. The market is treating them as the latter.
My Take: The Price of Uncertainty Is Creating Opportunity
I think the pullback in SMCI is an overreaction. Investors are panicking about temporary delays and guidance cuts, while ignoring the company’s execution, positioning, and balance sheet strength.
In a world where AI demand is outstripping supply, I want to own the companies that build the physical foundation for this transformation. Super Micro may not be a household name like Nvidia, but it's arguably just as critical to the AI buildout story just behind the scenes.
Is the stock overvalued on traditional metrics? Maybe. But great companies rarely look cheap in the middle of their growth curves.
Final Thoughts
I’m not saying to go all-in. There’s volatility ahead. But I am saying that SMCI deserves a spot in any forward-looking portfolio. Personally, I’ve added modestly to my position on weakness, with the expectation that FY2026 and beyond will reward patient investors.
Wall Street can’t seem to decide whether Super Micro is a fad or a foundational AI player. I’ve made my bet and I’m betting this story is far from over.
@stockcharts808 Buying TSVT would be a smart move for $BMY. Unlock shareholder value by strengthening the CAR-T portfolio and take advantage of the manufacturing partnership with Cellares.
@SqueezeMetrics So if $TSLA declines after the $SPY add, as many predict, then what is preventing the entire S&P to get dragged down with it with such a large weighting? Please comment on if you think Tesla inclusion bursts the S&P bubble? That would be more problematic in my mind.