DeepValue | The New Way Investors Work

At first glance, Super Micro Computer (SMCI) looks like one of the purest public-market plays on the AI infrastructure build‑out. Revenue has exploded, AI servers and GPU-optimized racks are flying out the door, and the company keeps talking about “plateau-high” demand into fiscal 2026 and beyond.

But when we dig into the filings, a very different picture emerges underneath the headline growth story: thin 6.3% gross margins, aggressive inventory build, negative operating cash flow, and a breathtaking level of customer concentration. According to the 10-Q (2026), p.38, SMCI’s latest quarter delivered $12.682B of net sales, mostly from data center customers and “a large design win from one customer,” but also came with $132.9M of inventory write-downs and sharply higher production and expedite costs.

Our read: at roughly $30.54 per share in this research framework, the market is already paying for sustained, record-level AI volumes and a clean margin recovery that has not yet shown up in the numbers. We rate SMCI a WAIT with a 4.0 conviction score, a base‑case value around $32, an attractive entry closer to $26, and a trim zone above $38, based on our scenario work.

If you want to analyze complex, filing-heavy stories like SMCI without spending days in 10‑Ks and 10‑Qs, you can use DeepValue to automate the deep research work.

Build a fully cited SMCI thesis in minutes instead of weekends by letting DeepValue parse the 10‑K, 10‑Q, and industry sources for you

Run Deep Research on SMCI →

In this article, we’ll walk through how we get to WAIT, what needs to improve for us to upgrade our stance, and how long-term investors can frame SMCI’s risk/reward inside the AI hardware cycle.

What Does SMCI Actually Do in the AI Value Chain?

SMCI is not a chip designer; it sits in the middle of the AI infrastructure stack as a fast-moving systems integrator.

According to the 10-K (2025), pp.3–4, SMCI:

Designs and sells GPU‑optimized servers, storage, and rack‑scale data center platforms
Offers both liquid- and air-cooled AI systems
Provides modular subsystems and accessories
Serves large enterprises and data centers across the U.S., Asia, and Europe

SMCI’s role is to take NVIDIA and AMD accelerators, CPUs, memory, networking, and its own chassis and cooling technologies, and turn those into fully configured racks and clusters that hyperscalers and large enterprises can deploy quickly. As the 10-K (2025), p.3 explains, this OEM/ODM-style model embeds services and support revenue inside system sales, but economically, the business still looks very much like hardware plus integration.

Recently, management has layered on two big strategic angles:

Data Center Building Block Solutions (DCBBS): Bundling power, cooling, facility equipment, and management services around servers to capture more margin and reduce pure-hardware commoditization. The new DCBBS line was announced in Supermicro’s Oct 14 2025 press release.
Liquid cooling (DLC‑2): Positioning liquid cooling as a total-cost-of-ownership lever for dense AI clusters. SMCI claims up to 40% power savings and up to 20% TCO reduction vs air-cooled setups, per the May 14 2025 DLC‑2 press release.

These are the right strategic directions if SMCI wants to escape low‑margin “metal bending.” The open question is whether those initiatives are already strong enough to offset intense price competition and ramp‑related costs we see in the recent financials.

The Numbers Behind the AI Hype: Revenue Surging, Margins Squeezed

SMCI’s top line has been extraordinary. According to the 10-K (2025), p.41, net sales jumped from $14,989,251 in fiscal 2024 to $21,972,042 in fiscal 2025, driven by GPU servers, HPC, and rack‑scale solutions. That momentum accelerated into FY26:

Q2 FY26 net sales: $12.682B for the quarter ended December 31, 2025
Driven by data center shipments and “a large design win from one customer,” per the 10-Q (2026), p.38

The problem is what the income statement and cash flow statement look like beneath that revenue.

From the 10-Q (2026), pp.38, 43:

Gross margin collapsed to 6.3%, down from 11.8% a year earlier
Drivers included:
Competitive pricing and unfavorable mix
Higher production and expedite costs
Tariff expenses from “new trade policies”
Inventory write-down adjustments of $132.9M in the quarter and $169.1M year‑to‑date

At the same time, the balance sheet ballooned:

Accounts receivable reached $11.0B
Inventory reached $10.6B
Net cash used in operating activities was $(941.4)M for the six months ended December 31, 2025

Our takeaway: SMCI is successfully winning AI deals, but it is paying for that share via:

Discounting and aggressive pricing
Rush production, freight, and expedite costs
Platform-transition friction that is causing large inventory write-downs
Heavy working‑capital investment that is currently consuming cash

From a value-investor perspective, that combination is the opposite of a “margin of safety.” We’re not just looking for growth; we’re looking for growth that comes with improving economics or at least stable, cash‑generating economics. Right now, SMCI doesn’t clear that bar.

Customer Concentration: One Hyperscaler Drives the Story

If you remember only one risk about SMCI, remember this: one customer effectively is the story today.

Per the 10-Q (2026), p.11:

Customer A represented 62.6% of net sales for the three months ended December 31, 2025
For the six‑month period, that customer was 45.1% of net sales

The Q2 FY26 call transcript summarized at Fintool, Feb 3 2026 corroborates that roughly two‑thirds of quarterly revenue is tied to a single hyperscale data center customer. That level of concentration turns everything into a binary bet:

Revenue is highly sensitive to one deployment schedule
Working capital is hostage to that customer’s payment terms and acceptance timing
Pricing power is structurally weak, because that customer knows how critical it is to SMCI

Management itself flags, in the 8-K (2026), that pursuing larger customers can further concentrate the base and pressure margins.

For our portfolio construction, that means we cap position size and require a stronger margin of safety than usual before we’d lean in. At the current setup, we don’t see that cushion.

Is SMCI Stock a Buy in 2026, Or Is It Time to Wait?

We frame SMCI’s valuation through three explicit scenarios from our research:

Base case (50% probability):

Implied value: $32
Revenue holds in a $12.0B–$13.0B quarterly band
Gross margin improves to 7%–8%
Free cash flow roughly flat
Key driver: expedite and new-platform ramp costs fade as Blackwell‑class builds standardize

Bear case (30% probability):

Implied value: $22
Revenue drops below $12.0B
Gross margin sticks around 6%
Cash flow stays negative
Key driver: deployment timing slips for the big customer, stretching receivables and inventory

Bull case (20% probability):

Implied value: $40
Revenue remains ≥$13.0B
Gross margin climbs to 9%–10%
Operating cash flow turns positive
Key driver: higher-margin facility and services attach through DCBBS

At around $30.54 in this framework, the stock is effectively asking investors to assume something between the base and bull case: plateau‑high revenue, fading ramp costs, and a clean margin recovery. But as the 10-Q (2026), pp.38, 43 makes clear, the most recent audited reality is:

6.3% gross margin
$132.9M quarterly inventory write-downs
Meaningful tariffs and expedite costs still very much in the picture

That’s why we land on WAIT rather than BUY. We’re not fighting the AI demand thesis; we’re pushing back on paying up before we see any evidence that this business can turn revenue into durable profits and cash.

If you’re trying to hold a watchlist of names like SMCI, Dell, and other AI-infrastructure plays and compare their unit economics in parallel, Read our AI-powered value investing guide to see how we use automation to scale that kind of cross‑company deep work.

What Needs to Happen Before SMCI Becomes More Interesting?

We think about SMCI through a simple 3–6 month falsification window. There are clear, testable conditions that would upgrade the setup from “wait” to “potentially attractive.”

1. Q3 FY26 Revenue ≥ $12.3B

Management guided to at least $12.3B in net sales for Q3 FY26 and ≥$40.0B for FY26 net sales, per the Supermicro IR release, Feb 3 2026. We treat that $12.3B as a line in the sand.

If Q3 FY26:

Prints below $12.3B
While customer concentration remains >50%
And guidance is cut meaningfully below $40B for FY26

Then our core underwriting — “AI volume winner” with plateau‑high demand — breaks. In that scenario, revenue, margins, and working capital all become more fragile at once. Our playbook is to cut or avoid exposure rather than “buy the dip.”

2. Gross Margin Starts to Recover and Write-Downs Fade

A single quarter at 6.3% gross margin with big write-downs could be written off as a transition blip. Two or three similar quarters in a row start to look structural.

We’ll be watching:

Sequential changes in:
Inventory write-downs vs the Q2 FY26 benchmark of $132.9M, per the 10-Q (2026), p.16
Management commentary on expedite, freight, and tariff expenses from 10-Q (2026), p.43
Evidence that gross margin is sustainably trending back toward 8%–10%

If we see another quarter or two with write-downs at or above Q2 levels, and no sign of easing ramp costs, we treat that as a thesis breaker. It would imply that SMCI’s platform transition process is destroying gross profit faster than volume gains can offset.

3. Working Capital Normalizes and Cash Flow Turns Positive

For a hardware-heavy integrator, it’s normal to see some working‑capital expansion during a spike in demand. But the magnitude here is striking:

AR of $11.0B and inventory of $10.6B
Operating cash flow of $(941.4)M over six months

We’re looking for:

AR and inventory to flatten or grow more slowly than revenue
Operating cash flow to flip positive, demonstrating that SMCI can collect efficiently and avoid over‑building stock

If those working-capital balances remain near current scale without a cash flow turn, we start to model in financing stress, tougher vendor terms, or forced discounting to clear aging inventory — all of which hurt equity value.

This is exactly the kind of multivariate pattern (revenue, GM, AR, inventory, cash flow) that DeepValue tracks in parallel across filings.

Let DeepValue surface red flags across SMCI’s cash flow and working-capital trends while you focus on portfolio decisions

See the Full Analysis →

Will SMCI Deliver Long-Term Growth and Profits?

The fundamental long-term question isn’t whether AI infrastructure grows — we think it does — but whether SMCI can translate that secular tailwind into a sustainable, high‑return business.

The Long-Term Bull Case

We see three key pillars to a constructive long-term view:

1. Sustained AI demand and platform cadence

SMCI has demonstrated it can ride GPU transitions quickly. From FY24 to FY25, net sales jumped ~47%, per the 10-K (2025), p.41, and Q2 FY26 revenue of $12.7B underscores its ability to capture new waves like NVIDIA’s Blackwell and AMD’s MI series. Tailwinds include:

Company guidance for FY26 net sales ≥$40B
Broader channel expansion such as Eviden’s partnership to distribute Supermicro AI SuperCluster racks, described in GlobeNewswire, Mar 11 2025

2. Higher-margin DCBBS and services attach

If DCBBS can meaningfully lift the mix of power, cooling, and management services, SMCI can shift some of its economics away from bare-metal hardware into more defensible, service‑rich solutions. Management’s push here, noted in Supermicro’s Oct 14 2025 release, is strategically sound.

3. Liquid cooling as a TCO differentiator

In dense AI racks, power and cooling become first‑order concerns. If SMCI’s DLC‑2 can truly deliver up to 40% data-center power savings and up to 20% TCO reductions vs air-cooled designs, as claimed in the May 14 2025 press release, SMCI gains a strong talking point to defend pricing and win complex deployments.

Put differently, there is a plausible path where SMCI grows into a structurally better business than a traditional server OEM.

The Bear Case: Structural Margin Squeeze and Governance Overhang

The counterpoint is equally clear, and it’s already showing up in the data:

AI servers may inherently carry lower gross margins due to intense competition and higher component content. Morgan Stanley’s note on Dell, summarized by Yahoo Finance, Nov 2025, highlights AI mix pressure across the OEM landscape.
SMCI is feeling that pressure more acutely: 6.3% gross margin with heavy ramp costs and write-downs, as detailed in the 10-Q (2026), pp.38, 43.
Platform transitions carry reliability and qualification risk. Industry coverage like Tom’s Hardware, Jul 2025 points to coolant leak issues with large‑scale liquid-cooled deployments. Those issues can delay shipments and increase rework and expedite costs — pressures we already see flowing through SMCI’s P&L.

On top of that, governance and controls remain unresolved:

As of December 31, 2025, SMCI’s CEO and CFO concluded that disclosure controls were not effective due to unremediated material weaknesses in internal control over financial reporting (IT general controls, segregation of duties, related-party disclosure controls, etc.), as stated in the 10-Q (2026), p.50.
The 10-K (2025), p.123 reiterates the requirement for sustained control effectiveness before these weaknesses can be considered fully remediated.

We view this as a meaningful valuation cap. The market can only put so high a multiple on a company that:

Has rapidly growing, low‑margin hardware revenue
Consumes cash on growth
And still hasn’t cleaned up material weaknesses in its control environment

We would need to see explicit, sustained remediation of ICFR issues — and management concluding that disclosure controls are effective — before we’d consider SMCI a fully investable “core” holding rather than a tactical, cycle‑sensitive name.

How We’d Tactically Trade or Accumulate SMCI

Given this setup, our stance is not “never touch SMCI.” It’s more nuanced:

Below ~$26 (our attractive-entry zone):

If Q3 FY26 revenue holds ≥$12.3B and we see improving margins and cash metrics, we’d start to accumulate with a 6–12+ month horizon.

Around $30–$32 (our base value):

We’re content to watch and wait. The price is roughly fair if everything goes right, but we don’t have enough evidence yet that things are going right structurally.

Above $38 (our trim zone):

We’d be inclined to take profits or reduce exposure unless we see clear, multi‑quarter confirmation of higher margins, healthier working capital, and resolved governance.

Time horizon matters here. Short-term traders can try to ride the AI momentum and quarter‑to‑quarter beats, but they are taking real event risk around guidance and customer behavior. Long‑term investors should let the next 1–2 quarters do the heavy lifting in proving (or disproving) the unit economics.

For that kind of event‑driven setup, having a fast way to digest every new 10‑Q and 8‑K is critical.

Use DeepValue to automatically parse each new SMCI filing, track margin and cash trends, and update your thesis before the crowd

Research SMCI in Minutes →

Key Risk Checklist: What Could Break the Thesis?

We maintain a concrete list of “thesis breakers” and early warning signs for SMCI. Here’s how we’d structure risk monitoring as investors:

Thesis Breakers

Any of the following would push us toward reducing or exiting exposure:

Q3 FY26 net sales < $12.3B and FY26 guidance cut below $40B, while Customer A still accounts for >50% of revenue

Inventory write-downs staying at or above $132.9M per quarter for another two quarters

Early Warning Indicators

Signals that would make us more cautious or cap position size:

Customer A remains ~63% of revenue with no meaningful diversification

AR and inventory remain near current levels without cash flow improvement

Management continues to cite higher freight, production, and expedite costs as key gross-margin drivers on subsequent calls

How Should Fundamental Investors Approach SMCI Today?

Our overall framework as the DeepValue team:

Acknowledge the real secular tailwind. AI infrastructure demand is strong, and SMCI has clearly proven it can win business and ramp fast.
Respect the financial red flags. Low single‑digit gross margins, heavy write-downs, negative operating cash flow, concentration risk, and unresolved ICFR weaknesses are not minor details; they are central to whether this is a compounding business or a cyclical trading vehicle.
Let the next 3–6 months do the work for you. We have very clear, testable conditions around Q3 FY26 revenue, margin behavior, working capital, and governance progress. Until those resolve, we’d rather protect capital than chase upside.

For investors who insist on exposure today, we’d suggest:

Keeping position sizes modest
Using explicit stop‑loss or thesis‑break rules tied to upcoming filings
Focusing on the quality of revenue and cash generation, not just the headline growth rate

SMCI may yet evolve into a high‑quality AI infrastructure platform with stronger economics. We just don’t think the current evidence base justifies paying up as if that transformation were already complete.

Super Micro Computer (SMCI) Deep Research Report: Waiting For Proof of Profits in the AI Server Gold Rush