Super Micro Computer stock remains one of Wall Street’s most heavily shorted large AI-infrastructure names, with roughly 91.2 million SMCI shares sold short as of August 14, equal to about 18.1% of the public float. Yet the bearish positioning is colliding with a powerful operating backdrop: Supermicro just reported $11.1 billion in quarterly revenue, more than $60 billion in new orders, record backlog and fiscal 2027 sales guidance as high as $72 billion, while Dell’s blockbuster AI-server results on September 2 provided another fresh signal that demand for Nvidia-powered data-center systems remains extremely strong.
That makes Super Micro Computer stock a very different short-squeeze candidate from the speculative small caps that typically dominate heavily shorted lists. SMCI is a roughly $25 billion company operating directly inside the AI data-center spending boom, and the company has real revenue, profits and a growing order book. The bearish case instead revolves around whether margins can remain healthy, whether enormous customer commitments convert into revenue on schedule, and whether governance and export-control concerns continue to hang over a business that is otherwise expanding at extraordinary speed.
Super Micro Computer Has More Than 91 Million Shares Sold Short
The size of the bearish position is significant. MarketBeat’s latest data show 91.22 million SMCI shares sold short as of August 14, representing approximately 18.08% of the public float and roughly $3.6 billion of bearish exposure at the reporting-date price. Short interest has declined from more than 103 million shares at the end of June, when about 20.5% of the float was short, but it remains unusually elevated for a company of Supermicro’s scale.
There is one important reason not to overstate the squeeze setup: SMCI trades very heavily. Its short-interest ratio is only about 1.4 days, meaning the current short position could theoretically be covered relatively quickly based on average trading volume. That makes SMCI less vulnerable to the kind of mechanical squeeze that can occur in thinly traded stocks where short sellers need ten or fifteen days of normal volume to exit.
Still, 18% of float being short is substantial. If Supermicro produces another major earnings beat, raises revenue guidance again or wins large next-generation AI deployments, short covering could amplify the upside even if it does not produce a classic meme-stock squeeze.
Related articles
The more interesting question is why so many investors remain willing to bet against the company despite extraordinary AI demand.
Supermicro’s Revenue Growth Is Exploding
Supermicro’s latest results make the bear case harder to frame purely around demand.
For the fiscal fourth quarter ended June 30, the company reported net sales of $11.1 billion, up from $5.8 billion in the year-earlier period. Net income surged to $1.18 billion from $195 million, while diluted earnings reached $1.62 per share compared with $0.31 a year earlier. Gross margin also improved sharply to 17.5% from 9.5%.
For the full fiscal year, revenue reached $39.1 billion, up from $22.0 billion in fiscal 2025, while net income more than doubled to $2.2 billion. Management said Supermicro generated more than $60 billion in new orders during the fourth quarter and entered fiscal 2027 with record backlog.
That order figure is critical because it suggests customers are committing capital well ahead of delivery. Supermicro has increasingly moved beyond selling individual servers toward complete rack-scale and data-center infrastructure systems, which can make each customer deployment much larger.
The company now expects fiscal 2027 revenue of $65 billion to $72 billion, far above the $39.1 billion generated in fiscal 2026. Reuters noted that the forecast also came in substantially ahead of Wall Street’s roughly $52.5 billion consensus estimate at the time.
If Supermicro delivers near the top of that range, annual revenue could almost double again.
Dell Just Strengthened the Bull Case for SMCI
One of the most important developments for Super Micro Computer stock did not come from Supermicro itself.
Dell reported exceptionally strong AI-server results on September 1 and raised its full-year outlook after demand continued to exceed expectations. Dell said AI-optimized server revenue reached $16.4 billion in its latest quarter, up 100% year over year, while its AI-server backlog climbed to approximately $95 billion. The company also increased its fiscal-year revenue forecast to $192 billion from $167 billion.
Dell shares surged following the report, and Super Micro was among the other AI-server names that benefited as investors interpreted the numbers as evidence that infrastructure spending remains exceptionally strong.
That matters because one of the biggest risks for SMCI is a sudden slowdown in hyperscaler and AI-cloud capital expenditures. Dell’s numbers point in the opposite direction.
The demand appears broad enough that competing server vendors are simultaneously building massive backlogs rather than simply stealing business from one another. Dell said it had received more than $130 billion in AI-server orders over the previous year and raised expected fiscal 2027 AI-server revenue to $74 billion.
For Supermicro, that industry read-through is clearly positive.
Nvidia’s Outlook Makes the AI Server Story Even Stronger
Nvidia has provided another major tailwind.
The chipmaker recently projected roughly 70% sales growth for its next fiscal year and forecast $108 billion in revenue for its current quarter, signaling that the AI-infrastructure boom could continue far longer than skeptics expected. Nvidia’s data-center revenue more than doubled to roughly $89 billion in its latest quarter, while management said demand for next-generation Vera Rubin systems remains extremely strong.
Supermicro is closely tied to that product cycle.
The company competes partly on its ability to bring servers built around the newest Nvidia accelerators to market quickly. Its liquid-cooled rack systems and Data Center Building Block Solutions are designed to support increasingly dense AI clusters where power consumption, cooling and deployment speed are becoming just as important as the GPUs themselves.
Supermicro has already outlined infrastructure designs for Nvidia’s Vera Rubin NVL4 platform capable of supporting more than 1,100 Rubin GPUs in a 3.2-megawatt scalable unit.
That gives SMCI direct exposure to Nvidia’s next product cycle rather than leaving it dependent on older AI hardware.
If Rubin demand develops anything like Blackwell demand did, Supermicro could have another enormous infrastructure refresh ahead.
Margins May Be the Most Important Number for SMCI Stock
Revenue growth alone is not enough.
Supermicro’s margins have historically been one of Wall Street’s biggest concerns because AI servers contain extremely expensive Nvidia GPUs and other components, meaning billions of dollars of revenue do not necessarily translate into equally impressive profits.
Fiscal Q3 highlighted that risk. Supermicro generated $10.2 billion of revenue but produced only a 9.9% gross margin. The company had reported an even weaker 6.3% margin in the previous quarter.
Then Q4 changed the picture dramatically.
Gross margin rebounded to 17.5%, substantially exceeding the company’s earlier expectation of 15% to 17%. Management attributed the improvement primarily to a more favorable mix of customers and products.
That recovery may be more important to SMCI investors than the headline revenue figure.
If Supermicro can maintain margins in the mid-teens while revenue moves toward $70 billion, earnings power could rise dramatically. If margins return toward single digits because competition becomes more aggressive or giant customers demand better pricing, revenue growth may look much less attractive.
The short sellers are effectively betting that margin normalization will eventually expose the limits of Supermicro’s business model.
Customer Concentration Is Another Risk
The company’s growing dependence on extremely large orders creates another complication.
Supermicro said nine customers generated more than $1 billion each during fiscal 2026, compared with only four in the previous year. That illustrates the scale of demand, but it also means revenue can become more dependent on a relatively small number of enormous customers.
Large AI-cloud providers and hyperscalers possess significant negotiating power. They can delay installations, shift orders between suppliers or demand lower pricing in exchange for enormous commitments.
Supermicro itself warns that increasingly targeting larger customers could raise customer concentration, increase costs and make quarterly sales less predictable.
This partly explains the stock’s extreme volatility.
A deployment slipping from one quarter into the next can move billions of dollars of revenue without necessarily changing the underlying customer relationship.
Investors therefore need to look beyond individual quarterly misses and determine whether backlog is disappearing or merely moving between reporting periods.
The Export-Control Investigation Still Matters
Supermicro also carries a governance and regulatory overhang that many competitors do not.
In March 2026, two employees and one contractor associated with the company were indicted over an alleged conspiracy involving export-control violations. Supermicro itself was not named as a defendant and has said it was not accused of wrongdoing. The three individuals are no longer associated with the company.
On August 20, Supermicro announced that an investigation led by independent members of its board had been completed. The review used outside legal counsel and an independent forensic accounting consultant, while the company said it was continuing to enhance its export-compliance program.
That development removes some uncertainty, but the issue remains relevant because Supermicro’s stock has previously been highly sensitive to governance concerns.
For investors who lived through earlier accounting and filing controversies around the company, even a strong AI growth story does not entirely eliminate trust risk.
That history helps explain why SMCI can trade at a much lower earnings multiple than some AI peers despite extraordinary revenue growth.
SMCI Valuation Looks Surprisingly Modest
At roughly $36 to $37 per share on September 2, Super Micro carries a market capitalization near $25 billion and trades at roughly 17 to 18 times trailing earnings.
That valuation is remarkably modest compared with many other AI infrastructure companies.
The reason is that investors are applying a substantial discount for execution risk, thin historical margins, customer concentration, governance concerns and the cyclical nature of server hardware.
There is also intense competition.
Dell is scaling AI infrastructure rapidly and now has a $95 billion backlog. Hewlett Packard Enterprise is pursuing the same market, while cloud customers increasingly design portions of their infrastructure internally. Server assembly is not protected by the same semiconductor-level barriers that benefit companies such as Nvidia.
Supermicro must therefore maintain an advantage through speed, liquid cooling, customization and complete data-center solutions rather than relying on proprietary AI chips.
At current valuation levels, the market appears to be saying that the growth is real but the durability of those economics remains uncertain.
Is Super Micro Computer Stock a Short-Squeeze Candidate?
SMCI clearly has enough bearish positioning to benefit from short covering, but it is not an ideal mechanical squeeze.
About 18% of float is short, yet only 1.4 days of average volume would be required to cover those positions. The stock is extremely liquid, which gives bearish investors far more ability to exit than they would have in a smaller company.
The better bullish argument is fundamental.
Supermicro is guiding toward as much as $72 billion in annual revenue, has more than $60 billion of recent orders, recovered gross margins sharply in Q4 and is directly exposed to Nvidia’s next-generation AI systems. Dell’s latest results provide independent confirmation that AI-server demand remains exceptionally strong.
If those trends persist, short sellers could still add fuel to rallies.
But the stock does not need a squeeze to work. It needs Supermicro to prove that its enormous backlog can convert into sustainable earnings.
Outlook: The Next Battle Is Revenue Versus Margins
The bull-versus-bear debate around Super Micro Computer stock is unusually straightforward.
Bulls see a company sitting directly inside one of the largest infrastructure spending cycles in history. Nvidia says AI demand remains powerful, Dell is posting record server orders, Supermicro itself has more than $60 billion in recent orders, and management sees fiscal 2027 revenue potentially reaching $72 billion.
Bears argue that server hardware remains fiercely competitive, margins can collapse quickly, customer concentration is rising and Supermicro still carries governance and execution baggage.
The next several quarters should reveal which side is closer to reality.
Investors should focus especially on gross margin, cash generation, backlog conversion, customer concentration and the ramp of Nvidia’s Rubin systems.
More than 91 million SMCI shares are still sold short, but the real threat to the bears is not a meme-style squeeze. It is the possibility that Supermicro keeps turning the AI spending boom into tens of billions of dollars of real revenue—and proves that its recovered margins are here to stay.
