SMCI · Forward model · Bull case
The Bull case, 20 quarters out
Each segment is projected from its own operating driver, rolled up into consolidated cash flow, and discounted back to a fair value per share. The assumptions below are editable — change them and every number on this page moves with them.
Super Micro reports as ONE operating segment. The two verticals here are the customer-type split the company itself discloses — OEM Appliance & Large Data Center, and Enterprise & Channel — and nothing has been invented beneath them. Only the basis quarter is a reported split: management put fiscal Q4 2026 at roughly 50/50 ($5.5B large-data-centre, $5.6B enterprise/channel), and the FY2026 full-year split at 69%/31% of $39.06B. The three earlier quarters are therefore APPORTIONED: the full-year totals less the reported Q4, spread across fiscal Q1-Q3 at the resulting 76.8%/23.2% ratio, and every one of them is flagged estimated. History starts at fiscal Q1 2026 (calendar 2025 Q3) rather than running back further because the FY2025 customer-type split is not in hand — a shorter honest history beats four more quarters of guesswork. Quarter labels are CALENDAR, not fiscal: Super Micro's FY2026 Q4 ended 30 June 2026 and appears here as 2026 Q2. Segment EBITDA margins and capex intensity are assumptions, not disclosure — Super Micro publishes no segment profitability at all. The model's largest known omission is working capital: free cash flow here is EBITDA less capex less tax, and Super Micro's cash is consumed by inventory and receivable builds ahead of large deployments, not by capex (FY2026 free cash flow was deeply negative in the quarter revenue nearly doubled). Read the fair value as the value of the earnings stream, then discount it yourself for the fact that growth at this company is cash-hungry in a way this engine does not charge for. netCash is set to zero deliberately: the balance sheet carries convertible notes against a cash position that swings by billions with the inventory cycle, and a single point estimate would be more misleading than none.
The DCBBS thesis works. Selling cooling, power, networking, management software and deployment with the compute lifts the take on every rack, and the enterprise share keeps rising alongside it, so the blended margin goes up while volume still fills the line. Management has pointed at DCBBS as the highest-margin layer and at >20% as what a full solution can carry. This case pays for that and re-rates the exit multiple toward where the market values a solutions business rather than an assembler.
Latest: $40.89B (2031Q2E)
| Period | Value |
|---|---|
| 2025Q3 | $5.02B |
| 2025Q4 | $12.68B |
| 2026Q1 | $10.24B |
| 2026Q2 | $11.12B |
| 2026Q3E | $13.36B |
| 2026Q4E | $15.83B |
| 2027Q1E | $18.59B |
| 2027Q2E | $21.73B |
| 2027Q3E | $25.35B |
| 2027Q4E | $28.73B |
| 2028Q1E | $29.59B |
| 2028Q2E | $30.43B |
| 2028Q3E | $31.25B |
| 2028Q4E | $32.06B |
| 2029Q1E | $32.87B |
| 2029Q2E | $33.68B |
| 2029Q3E | $34.50B |
| 2029Q4E | $35.34B |
| 2030Q1E | $36.20B |
| 2030Q2E | $37.08B |
| 2030Q3E | $37.99B |
| 2030Q4E | $38.92B |
| 2031Q1E | $39.89B |
| 2031Q2E | $40.89B |
What drives each segment
OEM Appliance & Large Data Center
Units × priceThe AI factory business: hyperscalers, neoclouds, sovereign programmes and the very large private AI builds, buying liquid-cooled GPU racks by the row. It is where the growth is and where the concentration risk is — a single customer was 28% of FY2026 revenue, and nine customers each spent over $1 billion. What limits it is not demand but how many racks can be allocated, built and energised in a quarter.
Latest: $20.67B (2031Q2E)
| Period | Value |
|---|---|
| 2025Q3 | $3.85B |
| 2025Q4 | $9.74B |
| 2026Q1 | $7.86B |
| 2026Q2 | $5.50B |
| 2026Q3E | $6.82B |
| 2026Q4E | $8.44B |
| 2027Q1E | $10.41B |
| 2027Q2E | $12.81B |
| 2027Q3E | $15.74B |
| 2027Q4E | $18.46B |
| 2028Q1E | $18.69B |
| 2028Q2E | $18.90B |
| 2028Q3E | $19.10B |
| 2028Q4E | $19.28B |
| 2029Q1E | $19.45B |
| 2029Q2E | $19.61B |
| 2029Q3E | $19.77B |
| 2029Q4E | $19.91B |
| 2030Q1E | $20.05B |
| 2030Q2E | $20.18B |
| 2030Q3E | $20.31B |
| 2030Q4E | $20.43B |
| 2031Q1E | $20.55B |
| 2031Q2E | $20.67B |
Assumptions & reasoning
- Modelled on units rather than growth because the binding constraint is physical and publicly stated: Super Micro puts global manufacturing capacity at 5,000 racks a month, of which 2,000 are direct-liquid-cooled. The ceiling here is the DLC figure — 6,000 a quarter — not the headline one, because liquid cooling is what an AI factory buys and it is the narrower constraint. Even that is generous: what actually rations output is GPU allocation and the customer's power and networking readiness, which is the reason management gave for the fiscal Q3 2026 revenue miss — customers not ready, not factories full.
- The rack count is NOT disclosed. It is backed out of the reported $5.5B at an assumed $2.75M per rack, so treat the level as a calibration and argue with the price instead. If a rack is really $2.0M the volume is 2,750 and the ceiling binds sooner; if it is $3.5M the volume is 1,570 and it binds later. The revenue path barely moves either way, which is the point of splitting price from volume rather than hiding both in a growth rate.
- Margin falls rather than rises in the base case, and that is a deliberate disagreement with management. The basis quarter carried a 17.5% gross margin against 10.8% for FY2026 as a whole, on what the company itself described as a favourable customer and product mix. Taking the peak quarter as the new normal would flatter every projected year. The DCBBS thesis — that selling cooling, power, networking and services with the compute lifts margin structurally — is real and is what the Bull case pays for; it is not what the base case assumes.
- This vertical carries the concentration risk for the whole model. One customer was 28% of FY2026 revenue and is widely linked to the xAI/SpaceX Colossus build. Nine customers were each above $1 billion. There is no diversification inside this line to fall back on: if the largest programme pauses, roughly a fifth of company revenue pauses with it, which is what the Concentration case tests.
Enterprise & Channel
Growth pathThe older, broader book: traditional enterprises, distributors and resellers buying CPU servers, storage, blades and smaller AI systems. Lower growth and far less headline risk, but structurally better margin and no single customer that matters. Management's stated path to a durable margin runs through growing this share, not through the AI factories.
Latest: $20.22B (2031Q2E)
| Period | Value |
|---|---|
| 2025Q3 | $1.17B |
| 2025Q4 | $2.95B |
| 2026Q1 | $2.38B |
| 2026Q2 | $5.62B |
| 2026Q3E | $6.54B |
| 2026Q4E | $7.40B |
| 2027Q1E | $8.19B |
| 2027Q2E | $8.92B |
| 2027Q3E | $9.61B |
| 2027Q4E | $10.27B |
| 2028Q1E | $10.90B |
| 2028Q2E | $11.53B |
| 2028Q3E | $12.15B |
| 2028Q4E | $12.78B |
| 2029Q1E | $13.41B |
| 2029Q2E | $14.07B |
| 2029Q3E | $14.74B |
| 2029Q4E | $15.43B |
| 2030Q1E | $16.15B |
| 2030Q2E | $16.90B |
| 2030Q3E | $17.68B |
| 2030Q4E | $18.49B |
| 2031Q1E | $19.34B |
| 2031Q2E | $20.22B |
Assumptions & reasoning
- A growth driver rather than units or capacity, and that is the honest answer here rather than a lazy one. This line is hundreds of enterprise customers and a distribution channel buying a wide mix of servers, storage and standalone subsystems; there is no single physical constraint and no published volume or price to build from. Inventing a unit and an ASP for it would look more rigorous and be less true.
- The basis quarter is unusually strong and the starting growth rate deliberately does not extrapolate it. Enterprise & Channel was half of fiscal Q4 2026 revenue against 31% of the full year, which management attributed to the timing of large data-centre projects slipping rather than to a step change in enterprise demand. The 14% opening rate assumes some of that mix holds; a reader who thinks it was pure timing should pull it toward the terminal rate immediately.
- Margin is set well above the large-data-centre line because the product mix is genuinely different — CPU servers, storage, networking and standalone subsystems, sold in smaller quantities to buyers without hyperscaler purchasing power. Super Micro discloses no segment profitability whatsoever, so the 17% is an assumption about mix, not a reported figure, and the gap between the two verticals is what the whole margin thesis rests on.
- This is the line the bull case actually needs. Revenue growth is not scarce at Super Micro; margin is. Every extra point of Enterprise & Channel share lifts blended margin without needing another GPU allocation, which is why management talks about DCBBS and enterprise mix in the same breath even though DCBBS also sells into the large accounts.
From cash flow to fair value
| Present value of free cash flow, 20 quarters | $36.30B |
| Terminal-year revenue | $157.70B |
| Terminal-year EBITDA | $18.54B |
| Exit multiple, on revenue | 0.6x |
| Terminal value | $94.62B |
| Discounted at 11.5% a year, terminal value becomes | $54.90B |
| Enterprise value | $91.21B |
| Net cash | $0 |
| Equity value | $91.21B |
| Diluted shares | 0.65B |
| Fair value per share | $141.00 |
| Against the current price of $36.50 | +286% |
0.4x terminal revenue on a roughly 8% blended EBITDA margin — about 5x EBITDA, which is where a thin-margin hardware integrator with real customer concentration belongs, not where a platform belongs. Super Micro trades near 0.65x trailing sales today, so the exit multiple is a de-rating, and deliberately: the terminal year in this model is a much larger, more concentrated, lower-margin business than the one being valued at 0.65x. The discount rate is 13% rather than the 10-11% used for the platform names, for three reasons that are all specific rather than atmospheric — one customer at 28% of revenue, a gross margin that has ranged from 6.3% to 17.5% across the last four quarters, and a working-capital pattern that consumes cash exactly when the business is winning. Move the exit multiple before anything else: at 0.25x and at 0.7x the answer moves by more than any operating assumption in the spec, which is the honest warning about valuing a business whose margin is this thin. Nothing here prices the >$60B order book as backlog; it shows up only as the growth rate it can sustain.
Read the other way round: at $36.50 the market is paying -0.1x terminal-year revenue, holding every other assumption on this page fixed. That is the number to argue about.
The projected path
| Quarter | OEM Appliance & Large Data Center | Enterprise & Channel | Revenue | YoY | EBITDA | Capex | FCF | R40 | PV of FCF |
|---|---|---|---|---|---|---|---|---|---|
| 2026 Q3E | $6.82B | $6.54B | $13.36B | +166% | $1.73B | $267M | $1.25B | +176 | $1.21B |
| 2026 Q4E | $8.44B | $7.40B | $15.83B | +25% | $1.99B | $317M | $1.43B | +34 | $1.35B |
| 2027 Q1E | $10.41B | $8.19B | $18.59B | +82% | $2.26B | $372M | $1.61B | +90 | $1.48B |
| 2027 Q2E | $12.81B | $8.92B | $21.73B | +95% | $2.55B | $435M | $1.80B | +104 | $1.61B |
| 2027 Q3E | $15.74B | $9.61B | $25.35B | +90% | $2.86B | $507M | $2.00B | +98 | $1.75B |
| 2027 Q4E | $18.46B | $10.27B | $28.73B | +81% | $3.15B | $575M | $2.19B | +89 | $1.86B |
| 2028 Q1E | $18.69B | $10.90B | $29.59B | +59% | $3.26B | $592M | $2.27B | +67 | $1.88B |
| 2028 Q2E | $18.90B | $11.53B | $30.43B | +40% | $3.37B | $609M | $2.35B | +48 | $1.89B |
| 2028 Q3E | $19.10B | $12.15B | $31.25B | +23% | $3.48B | $625M | $2.42B | +31 | $1.90B |
| 2028 Q4E | $19.28B | $12.78B | $32.06B | +12% | $3.58B | $641M | $2.50B | +19 | $1.90B |
| 2029 Q1E | $19.45B | $13.41B | $32.87B | +11% | $3.69B | $657M | $2.58B | +19 | $1.91B |
| 2029 Q2E | $19.61B | $14.07B | $33.68B | +11% | $3.80B | $674M | $2.66B | +19 | $1.92B |
| 2029 Q3E | $19.77B | $14.74B | $34.50B | +10% | $3.92B | $690M | $2.74B | +18 | $1.92B |
| 2029 Q4E | $19.91B | $15.43B | $35.34B | +10% | $4.03B | $707M | $2.83B | +18 | $1.93B |
| 2030 Q1E | $20.05B | $16.15B | $36.20B | +10% | $4.16B | $724M | $2.92B | +18 | $1.94B |
| 2030 Q2E | $20.18B | $16.90B | $37.08B | +10% | $4.28B | $742M | $3.01B | +18 | $1.95B |
| 2030 Q3E | $20.31B | $17.68B | $37.99B | +10% | $4.42B | $760M | $3.11B | +18 | $1.96B |
| 2030 Q4E | $20.43B | $18.49B | $38.92B | +10% | $4.56B | $778M | $3.21B | +18 | $1.97B |
| 2031 Q1E | $20.55B | $19.34B | $39.89B | +10% | $4.70B | $798M | $3.32B | +19 | $1.98B |
| 2031 Q2E | $20.67B | $20.22B | $40.89B | +10% | $4.86B | $818M | $3.43B | +19 | $1.99B |
Every row is projected. A year-over-year change is shown only where the quarter it compares against exists — an em dash means there is no comparable quarter, not a flat year.
Model revisions
Assumptions are marked to reality as each quarter prints. Every change is appended here, with the fair value the model produced at the time, so the model's own history stays visible.
| Date | Changed | Fair value then | Note |
|---|---|---|---|
| 2026-08-20 | — | Initial model. Two verticals on the disclosed customer-type split, basis fiscal Q4 2026 (calendar 2026 Q2) at $11.12B revenue, FY2027 base case calibrated to the $65-72B guidance range. |