SMCI — Super Micro Computer, Inc.

The tape is unambiguously bullish as of 2026-09-13: SMCI was up 31.65% over three months and 30.41% over six months, closed at $40.10 on a 7.28% single-day gain, and sat above all three key moving averages. The trigger was the 2026-08-11 Q4 FY2026 print - $11.1B quarterly revenue versus $5.76B a year earlier, gross margin snapping back to 17.5% from 9.9% in Q3, $1.68 diluted EPS, and disclosure of over $60B in orders booked in the quarter with FY2027 revenue guided to $65-72B. Post-print PT hikes (Wedbush to $40, Citigroup to $39, Rosenblatt to $51) confirm the re-rating. What is NOT bullish is the analyst rating distribution itself - 0 strong buy, 9 buy, 14 hold, 2 sell - and a $41.44 consensus target that was essentially the market price on 2026-09-13, so the market has warmed faster than the sell-side has.

The three-year quarterly series shows a long plateau followed by a violent inflection. Revenue sat at $2.12B in the Sep-2023 quarter, settled around $4.6-5.9B per quarter through FY2025 (Jun-2025: $5.76B), then went vertical in FY2026: $5.02B (Sep-25), $12.68B (Dec-25), $10.24B (Mar-26), $11.12B (Jun-26), summing to $39.1B for FY2026 versus about $22.0B for FY2025 - roughly +78% year over year. Margins tell the harder story: gross margin collapsed to 6.3% in the Dec-2025 quarter and 9.3% in Sep-2025 as low-margin hyperscale volume ramped, then recovered to 9.9% (Mar-26) and sharply to 17.5% (Jun-26); operating margin followed the same path at 3.6%, 3.7%, 6.1%, and 13.4%, and diluted EPS went $0.26, $0.60, $0.72, $1.68. The Jun-2026 quarter is a genuine inflection, not noise - but it is one quarter, and the trailing-twelve-month gross margin is still only 10.8%. Cash flow is the dissenting voice: operating cash flow was -$918M (Sep-25), -$24M (Dec-25), -$6.70B (Mar-26), then +$747M (Jun-26), a TTM operating cash flow of roughly -$6.8B against TTM net income of $2.23B. Two data caveats worth flagging: the reported 2023-12-31 and 2024-06-30 rows are numerically identical, as are the 2024-03-31 and 2024-09-30 rows, so 2023-2024 sequential comparisons in this dataset are unreliable; and stockholders' equity jumped from $7.58B (Mar-26) to $14.48B (Jun-26) - about $5.7B of non-earnings equity inflow - while long-term debt rose from $2.43B (Mar-25) to $6.68B (Jun-26). Growth is real and accelerating; it is being financed externally, not organically.

As of 2026-09-13, SMCI was cheap on every backward- and forward-looking earnings metric versus its own history. At $40.10, the TTM P/E was 10.99x against a five-year average of 19.29x, the TTM P/S was 0.664 against a five-year average of 1.47, and the forward P/E on consensus FY2027 EPS of $4.338 was just 9.24x. Price-to-book was 1.67x (1.93x on the screener's snapshot) with debt-to-equity of 0.461 and a current ratio of 3.872, backed by $7.52B of cash against $6.68B of long-term debt - a genuinely solid balance sheet after the recent financing. The PEG of 0.056 should be ignored entirely; it is a statistical artifact of a trailing growth denominator, not a signal. The honest framing: this is not a rich multiple that needs justifying, it is a discount that needs explaining. The discount is explained by a 10.8% TTM gross margin, an 8.9% ROIC, and roughly -$7.0B of TTM free cash flow. Cheap versus its own history, but the history was earned at higher margins and better cash conversion. The 'Cheap vs. its own history' label is accurate; the question is whether it is a mispricing or a warning.

As of 2026-09-13, at $40.10 the market was paying 9.24x consensus FY2027 EPS of $4.338 on $67.1B of consensus FY2027 revenue. Read that combination carefully: analysts are modeling about +72% revenue growth but only about +17% EPS growth (from $3.706 TTM), which implies a net margin of roughly 4.5% on FY2027 revenue - far below the 10.6% net margin the Jun-2026 quarter just printed. In other words, the market is giving SMCI full credit for the $60B backlog converting to revenue and almost zero credit for the 17.5% gross margin proving durable. That is the entire trade. If the June margin holds even halfway (say 11-12% gross margin at $67B revenue), consensus EPS is materially too low and the stock is priced for a disappointment that will not arrive; if the June quarter was a one-off mix or cost catch-up, consensus is roughly right and $40 is roughly fair. This is a de-risked revenue story with a completely undecided margin story - which means the stock trades on the gross margin line and the operating cash flow line, not on the top line, for the next several quarters. Do not anchor on the headline revenue beats.

As of 2026-09-13, SMCI was in a clear uptrend and somewhat extended. RSI(14) was 60.5 - elevated but not overbought - with price 5.87% above the 20-day SMA, 23.29% above the 50-day SMA, and 27.34% above the 200-day SMA. Returns were +1.29% over one week, +6.62% over one month, +31.65% over three months, and +30.41% over six months, and the $40.10 close sat in the middle of a $19.48-$58.78 52-week range. The structure is a healthy, confirmed uptrend rather than a blow-off, but the gap above the 50- and 200-day averages means a new buyer at $40.10 is paying up for a Hold-rated stock whose average analyst target was $41.44. Entry timing implication: there is no oversold or mean-reversion signal here. A pullback into the mid-to-high $30s (which is where the 20-day average and the prior $37.38 close sit) would offer meaningfully better risk/reward than chasing a 7.3% single-day pop, and the volatility (beta 1.996) means those pullbacks have historically arrived quickly.

• AI infrastructure capex cycle: SMCI is a derivative of NVIDIA's GB200/GB300 roadmap and hyperscaler, neocloud, and sovereign AI data-center spending. The $60B of orders booked in Q4 FY2026 and FY2027 guidance of $65-72B are only as good as the underlying capex budget cycle, which is itself the biggest single macro variable for the stock.
• Component supply and memory pricing: HBM, DRAM, NAND, and GPU allocation directly set SMCI's cost of goods and therefore its already-thin gross margin. A $12.9B inventory position means a sharp component price decline hits the balance sheet, while component inflation compresses margin on fixed-price racks.
• Cost of working-capital financing / rates: SMCI funds a large inventory build with borrowed money - long-term debt went from $2.43B (Mar-2025) to $6.68B (Jun-2026). Higher-for-longer rates raise the carry cost of that inventory and make the negative-free-cash-flow growth model more expensive; a genuinely accommodative rate environment would be a meaningful tailwind.
• Export controls and AI hardware regulation: US restrictions on AI server shipments and ongoing scrutiny of transshipment routes create revenue uncertainty and are a persistent discount factor on the multiple. Governance and compliance reviews left over from the prior filing-delay episode compound this.
• Sector rotation within AI hardware: capital has been rotating toward server and storage peers with cleaner cash conversion (notably Dell). SMCI has lagged those peers in 2026 despite faster revenue growth, and that relative-performance pattern tends to persist until cash flow inflects.

• Q1 FY2027 earnings, expected early-to-mid November 2026: the single most important datapoint. Watch three things - whether revenue lands near the ~$15B midpoint, whether gross margin holds at or above 17%, and whether operating cash flow is positive for a second consecutive quarter.
• Conversion of the $60B Q4 order backlog and improved customer prepayments: evidence that the growth can be partially self-funded would directly attack the bear case and could trigger multiple expansion.
• Liquid-cooling and Data Center Building Block Solutions (DCBBS) mix shift: if higher-margin integrated liquid-cooled racks become a larger share of revenue, gross margin structurally re-rates above the 10.8% TTM level and the whole earnings model changes.
• Broadening beyond concentrated hyperscalers into neocloud, sovereign, and traditional enterprise deployments via the VMware and Cisco partnerships, which would reduce customer-concentration risk.
• Sell-side rating migration: with 14 holds and 2 sells against 9 buys as of 2026-09-13, a cash-flow or margin inflection could drive a wave of upgrades and target revisions above the current $41.44 average.
• Downside catalyst of equal weight: another quarter of negative operating cash flow requiring new debt or equity, which would confirm dilution risk and likely compress the multiple further.

• Cash conversion is the central risk. TTM operating cash flow was roughly -$6.8B and TTM free cash flow roughly -$7.0B, driven by an inventory build to $12.9B; the March-2026 quarter alone burned $6.70B. The June-2026 quarter's +$747M was a relief, but one quarter does not resolve a structural pattern in which paper earnings routinely outrun cash generation.
• Financing and dilution risk is now visible in the balance sheet, not theoretical. Long-term debt rose from $2.43B (Mar-2025) to $6.68B (Jun-2026), and stockholders' equity jumped from $7.58B to $14.48B in the June-2026 quarter alone, implying roughly $5.7B of non-earnings equity inflow. Continued growth funded this way dilutes existing holders.
• Gross margin instability: 6.3% (Dec-2025), 9.3% (Sep-2025), 9.9% (Mar-2026), 17.5% (Jun-2026). If the June print does not repeat, FY2027 consensus EPS of $4.338 is too high and the 9.24x forward multiple is not cheap at all.
• Customer concentration and hyperscaler pricing power: the largest AI rack deals are competitively bid, and a single large customer repricing or slowing its deployment schedule can compress both revenue and margin simultaneously.
• Earnings-quality disconnect the deterministic score understates: the score's accruals ratio of 1.44% is computed on the June-2026 quarter only (net income $1.178B vs operating cash flow $747M against $29.9B of assets). On a trailing-twelve-month basis, net income of $2.23B against operating cash flow of about -$6.8B is a far more alarming gap than the score's single-quarter figure conveys.
• Cyclicality and high beta (1.996): AI server demand is tied to a capex cycle with no recurring revenue base. The stock fell as low as $19.48 within the past year; a capex digestion phase would hit both estimates and the multiple at the same time.

SMCI as of 2026-09-13 is a genuine 'cheap for a specific reason' situation, and the reason is cash, not demand. Revenue is close to pre-sold - over $60B of orders booked in a single quarter and FY2027 guided to $65-72B - yet the stock trades at 9.24x forward earnings because the company has repeatedly converted paper profits into negative cash flow, funds a $12.9B inventory position with rising debt and fresh equity, and just printed one quarter (17.5% gross margin, +$747M operating cash flow) that has to be repeated before anyone should believe the low multiple is a mispricing rather than a correct discount. The setup is asymmetric in the bulls' favor only if the margin holds; on the trailing twelve months it has not, and that is the honest reason the consensus rating is Hold. I would not chase this at $40.10, 23% above its 50-day average, right after a 7.3% pop - but this is exactly the kind of story worth owning into a second consecutive quarter of double-digit gross margin and positive operating cash flow, ideally bought on a pullback into the mid-$30s.

Forward P/E applied to a blended FY2027 (June-2027) EPS, not a restatement of the analyst target. I weighted consensus FY2027 EPS of $4.338 at 50% and an independently built mid-case at 50%: $67B revenue, an 11.0% gross margin (below the 17.5% just printed, above the 10.8% trailing-twelve-month figure), $2.1B of operating expense, and roughly 78% net-income-to-operating-income conversion consistent with FY2026, giving about $4.1B of net income on ~710M diluted shares, or $5.79. That blends to $5.06, and I applied a 10.0x multiple - roughly half the five-year average P/E of 19.29x, justified by a 10.8% trailing gross margin, an 8.9% ROIC, negative trailing free cash flow of about -$7.0B, and demonstrated dilution risk, but above the 9.24x the market was already paying on the lower consensus number. $5.06 x 10.0 = $50.60, which sits above the $41.44 analyst consensus and the $48 street high because I believe consensus EPS embeds a margin assumption the June-2026 quarter at least partially invalidated.