Seeing a stock drop from $118 to $22 over the course of about two years can cause a certain kind of vertigo. Not the typical vertigo caused by market corrections, where you find solace in charts and past averages. The other type, where the decline is related to something more difficult to measure, like trust, rather than just valuation. This is where Super Micro Computer finds itself in early April 2026, trading at $22.51 on the Nasdaq, down more than 65% from its near-term high of last July, and bearing the burden of an institutional exodus that doesn’t appear to be slowing down, a co-founder indictment, and a margin collapse.
At least from the outside, the San Jose headquarters, a low-slung campus that formerly hummed with the energy of a company riding the wave of AI infrastructure, now has a different atmosphere. One of the most obvious beneficiaries of the surge in AI spending not too long ago was Supermicro.
Super Micro Computer, Inc. (SMCI) — key information
| Full name | Super Micro Computer, Inc. (doing business as Supermicro) |
| Ticker symbol | NASDAQ: SMCI |
| Headquarters | San Jose, California, USA |
| Founded | September 1993 |
| CEO / Founder | Charles Liang (CEO since founding) |
| Current stock price | $22.51 USD (as of April 1, 2026) |
| Market capitalization | ~$13.5 billion |
| 52-week high / low | $62.36 / $19.48 |
| All-time high | $118.81 (March 2024) |
| P/E ratio (TTM) | ~17.16 |
| Q2 FY2026 revenue | $12.68B (+123.36% year-on-year) |
| Gross margin (latest) | 6.3% (down 550 basis points) |
| Key legal issue | Co-founder Wally Liaw indicted on export control violations (China) |
| Wall Street consensus | Hold; mean price target ~$35 |
| Official investor relations | ir.supermicro.com |
For a period in 2023 and early 2024, investors priced it appropriately because it manufactures the servers and rack systems that data centers require to run Nvidia’s chips. In March 2024, the stock reached a record high of $118.81. The narrative was tidy. AI required computation, which required servers, which Supermicro produced. Easy.
Subsequently, a series of complications began to emerge. The Nasdaq delisting threat and past-due financial filings were eventually resolved, but not before making a lasting impression. The most recent setback was the formal indictment of Yih-Shyan “Wally” Liaw, the company’s co-founder, on charges of circumventing US export regulations in transactions involving China. After the accusations, Liaw resigned. In a letter dated March 26, CEO Charles Liang swiftly announced new compliance procedures and a new acting chief compliance officer, portraying the move as proof of institutional seriousness. Observing all of this, the market was not overly impressed.
Tortoise Capital quickly clarified its stance by selling all of its SMCI holdings in the Tortoise AI Infrastructure ETF. The fund’s senior portfolio manager, Rob Thummel, was straightforward about the rationale: the indictment was the decisive element. Chief market strategist Brian Mulberry described the stock as “uninvestable” due to C-suite involvement in the legal issues. Zacks Investment Management had already left the company in 2025. Institutions appear to be deterred more by this uncertainty than by any particular piece of information. It is still unclear how far the regulatory scrutiny goes and what investigators might discover as they delve deeper.
On its own, the company is still doing something amazing: expanding. Revenue for the second quarter of fiscal 2026 was $12.68 billion, a 123% year-over-year increase that exceeded projections by over 22%. For the entire fiscal year, the company hopes to generate more than $40 billion in revenue.
Approximately 90% of Supermicro’s revenue now comes from AI GPU platforms; in a single quarter, the OEM and large data center segment alone brought in $10.7 billion. These figures don’t represent a business that has lost its commercial footing. In an odd way, they contribute to the complexity of the story. Supermicro’s revenue trajectory shows a company that is truly succeeding in AI infrastructure, but at the same time, the governance picture deteriorates in ways that make that revenue seem more precarious than it actually is.
The financial issue is most apparent in the gross margin. In an AI industry where leaders consistently maintain margins above 60%, the number is startling at 6.3%, down 550 basis points from previous levels. This is partly due to competitive pressure from Dell Technologies and Hewlett Packard Enterprise, both of which are aggressively entering the AI server market with their own Nvidia partnerships and, in HPE’s case, cloud-like consumption models that appeal to business clients.
Supermicro has responded to this by emphasizing speed and integration, positioning itself as a full-stack AI infrastructure provider rather than merely a hardware vendor. However, it is more difficult to maintain this differentiation when your biggest single client generates 63% of your quarterly revenue. Even without everything else, some institutional investors would be uneasy about that concentration alone.
Even though the price fell more than 60% during that time, it’s difficult to ignore the fact that insiders haven’t reported buying any stock in the last six months. Silence has a way of speaking. Meaningfully, but not definitively. A rebound toward $34 or $35 is conceivable if the AI demand cycle continues, according to some optimists who point to the forward earnings multiple hovering around 7x, far below the ten-year historical average of 12x. Longtime holder Louis Navellier has speculated that Liaw’s exit could ultimately be advantageous. He might be correct. Businesses have weathered more severe governance crises than this one.
However, it appears that Supermicro is juggling two distinct realities that haven’t yet been fully reconciled as it navigates 2026. One is a company that is building the infrastructure that the AI economy truly depends on, operating on almost every commercial cylinder. The other is a company that consistently gathers reasons for investors to be hesitant, and in markets such as this one, hesitancy builds up rapidly.

