Lenovo Q1 FY26/27 — The Best Quarter It Has Ever Had, Bought in Advance

Start by grading the last one

The June dive on Lenovo's FY25/26 close ended on a single question: the demand was never in doubt, so the print was "a 'does the margin turn stick' question." Infrastructure Solutions Group had just turned profitable for a full year for the first time and printed a record $202M of quarterly operating profit, and the argument was about whether that was a re-rate or one good cycle.

Three months later the answer arrived at a volume nobody modelled. ISG earned $776.9M of operating profit in a single quarter — against an operating loss of $85.5M in the same quarter a year earlier, and against $73M for the whole of FY25/26. One quarter produced 10.6x the entire prior fiscal year's infrastructure profit on 44% of its revenue.

So the margin turn stuck, and then some. That closes the June question and opens a harder one, because a number that good arrives by one of two routes and the two have opposite futures. Either Lenovo crossed the scale threshold where a server business finally absorbs its own fixed cost — in which case 9.1% is a floor. Or Lenovo bought its component position before the market repriced it and sold through at the new price — in which case 9.1% is a timing win, and timing wins are collected once.

The balance sheet answers it, and the release does not mention the balance sheet.

The quarter, stated plainly

Revenue was $26,943M, up 43% year-on-year and a record. Gross profit was $4,452M, up 60%, and gross margin rose 1.8 percentage points to 16.5%. Adjusted profit attributable to equity holders was $1,075M, up 176%, past a billion dollars for the first time in the company's history.

Reported profit attributable to equity holders was a loss of $609M. Reported operating profit was $19M, down 98% from $785M. Basic earnings per share were US(5.04) cents against US4.12 cents a year ago.

Both of those are true at once, and the flow shows how:

The gap is one line. Operating expenses were $4,433M against $1,990M, up 123%, and the release is direct about the cause: the increase "was mainly attributable to a fair value loss of US$1,690 million on derivative financial liabilities relating to warrants." Strip that and the other non-cash items and adjusted operating profit was $1,523M against $631M, up 141%, at a 5.7% adjusted operating margin against 3.4%.

The non-HKFRS bridge deserves credit before it gets scrutiny. It runs both ways: it adds back the warrant loss, and it also strips out a $250.1M fair-value gain on strategic investments that flattered the reported line. A bridge that removes its own good news is not a bridge built to flatter. Take the $1,075M as the operating truth of the quarter.

Where the entire margin came from

Lenovo says gross margin "increased by 1.8 percentage points to 16.5 percent, primarily driven by improved profitability in the ISG business." That sentence is doing more work than it looks.

Read the segment note and there is no second engine. IDG turned $17,105.5M of revenue into $1,210.1M of operating profit — a margin of 7.07% against 7.06% a year earlier. Not "stable at 7.1 percent" as a matter of rounding; stable to two decimal places, on 27% more revenue, through a quarter Lenovo itself describes as carrying "rising commodity costs and heightened market competition." SSG improved from 22.18% to 24.16%, which is real, on a segment that is 10.7% of group revenue. Neither moves the group by 1.8 points.

ISG does all of it. Revenue went from $4,290.1M to $8,510.1M, and operating profit from $(85.5)M to $776.9M. That is an $862.5M swing on $4,219.9M of incremental revenue — an incremental operating margin of 20.4% in a business that assembles and ships servers. For scale: the segment's steady-state margin over the four quarters to March 2026 was 0.4%.

Dan Gorakhpurwalla, who runs ISG, was asked directly whether one-time factors were involved and said the "record operating income and margin are indicative of business performance, not transitory factors" transcript. Set against the series, that claim needs to survive three things the quarter also contains.

The first is the base. A year-ago comparator that is a loss makes any recovery look like a discontinuity, and the prior-year quarter is not clean: ISG's own commentary carves out "the impact of international GPU sales in China in the prior-year quarter" when it describes AI-server growth. A comparison that needs a carve-out to be stated is a comparison worth stating twice.

The second is Infinidat. Lenovo completed the acquisition of the enterprise-storage vendor in early April 2026, inside this quarter, and said at the time that it strengthened "ISG's high-end enterprise storage capabilities and unlocking long-term potential for margin expansion." The cash flow statement carries $427.3M for "acquisition of businesses and subsidiaries, net of cash acquired." This is the first quarter ISG consolidates a high-margin storage business, and the Q1 commentary on ISG's record margin does not mention it once. Find the segment whose margin actually moved before crediting the acquisition — here the instruction inverts. The acquisition is the one contribution management had every reason to name and didn't.

The third is the warehouse.

The warehouse is the position

Inventories closed the quarter at $15,741.9M against $11,720.9M three months earlier. That is $4,021.1M of inventory added in ninety-one days, on a balance sheet whose total assets are $70,112.1M. Year-on-year, inventory is up 80.1% against revenue up 43%.

An inventory build that outruns revenue is usually the worst signal a hardware company can send — it normally means product was made and not sold. This one is the opposite, and the category split is what separates the two readings:

Finished goods rose 45.2% year-on-year, which is revenue growth plus a rounding error. Unsold product is not accumulating. Raw materials and work-in-progress rose 118.3% — from $4,562.6M to $9,961.6M — nearly three times the rate of the revenue it supports.

The seasonal control makes it sharper still. Between March and June of 2025, raw materials and WIP grew 14.2%; between March and June of 2026 they grew 41.5%. Same three months of the fiscal calendar, same seasonal shape, roughly three times the build. A one-period change is a phase, not a decision, until you have the series — and the series here does not soften the claim, it hardens it. This is not the working capital a growing business drags along behind it. It is a position, taken deliberately, in the one input whose supply-demand balance Lenovo names four separate times in its own release.

Yuanqing Yang described the mechanism on the call without quite naming it: "Our scale, resilient global supply chain built on our global local model, and efficiency of our end-to-end self-controlled operating model turned those constraints into competitive advantages" transcript. Decoded: Lenovo secured components while its competitors could not, at costs struck before the current lane, and sold finished systems into a market that had already repriced. The 20.4% incremental margin in ISG is what that looks like when it lands in a segment note.

This is the position the warehouse is the position was coined for, and Lenovo is the largest name yet to demonstrate it. It is worth being precise about what kind of finding this is. It is not an accounting problem, and it is not a criticism. Buying the right input ahead of the curve at this scale is a genuine operating achievement and the single most valuable thing a hardware assembler can do in a tight component lane. It is also, mechanically, a trade — and a trade pays on the turn you took, not on the ones after it. The second turn of that inventory is bought at the market.

The record that did not produce cash

A position has to be funded, and the cash flow statement is where the funding shows up.

Net cash generated from operations was $718.3M, against $1,566.0M in the same quarter last year — down 54%. After interest and tax, net cash generated from operating activities was $382.3M against $1,219.1M, down 69%. Adjusted profit attributable rose 176% over the same span.

Three first fiscal quarters, and the strongest one on the income statement is the weakest one on the cash statement — weaker even than Q1 FY24/25's $1,091.6M, when group revenue was a fraction of today's. Receivables tell the same story from the other side: trade receivables went from $10,532.1M a year ago to $18,844.3M, up 78.9%, again against 43% revenue growth.

Lenovo funded the gap rather than absorbing it. Financing activities brought in $1,504.7M against an outflow of $1,087.2M a year earlier, and the largest single item is $2,000.0M of new convertible bonds issued on June 25, 2026 — six days before the quarter closed. Trade payables rose to $18,698.7M and notes payable to $5,698.9M, up 88.2% year-on-year, so the supply chain is carrying part of it too.

None of this is distress. Cash and equivalents rose to $5,907.5M. The point is narrower and it is about attribution: the quarter's profit and the quarter's cash disagree by a wide margin, and the reconciling item is the same inventory that produced the profit. Price-long, cash-late was coined for the memory module layer; at $26.9B a quarter, Lenovo is the largest instance of it on the tape.

The loss management calls an accounting item

On the reported loss, the release says the warrant revaluation "largely reflects changes in the Group's share price and is expected to remain a non-cash accounting item through the end of fiscal year 2027/28."

The first half of that is true and the second half is a choice, not a fact. The warrants were issued on January 8, 2025 — 1,150,000,000 of them, for gross proceeds of HK$1,645M (approximately US$212M). Note 13 states that Lenovo "has the option to satisfy such exercise rights by allotment and issue of the Company's shares, or through cash payments." That option is precisely why the instrument sits as a derivative liability rather than in equity. It is non-cash only for as long as Lenovo elects shares.

And the liability is not small any more. It closed the quarter at $1,787.1M against $96.8M at March 31 — a warrant book that Lenovo sold for about $212M now carries at roughly 8.4x the proceeds. FY25/26 recorded a fair-value gain of $230.4M on the same instrument; the direction reversed because the shares ran.

The larger cost is not the mark at all. It is the claim on the equity, and the release discloses every piece of it in notes 12 to 14 without ever adding them up:

Against 12,404,659,302 shares outstanding, the disclosed full-conversion counts total 3,677,071,572 potential shares, or 29.6%. Excluding the 2033 convertibles, which are struck above the current quote, the in-the-money stack alone is 3,238,727,980 shares — 26.1%.

Then there is the ratchet. Effective August 8, 2026, every conversion price moved down: warrants from HK$12.31 to HK$11.51 (~US$1.48), the 2028 bonds from HK$10.02 to HK$9.37 (~US$1.20), the 2029 bonds from HK$9.94 to HK$8.37 (~US$1.07), the 2033 bonds from HK$36.70 to HK$35.74 (~US$4.58). A lower conversion price on a fixed principal means more shares for the same money. The adjustment is contractual and ordinary — it is what an anti-dilution clause does when a dividend goes out — but its effect runs one way, and it ran that way this quarter.

Lenovo's own EPS disclosure prices this honestly, which is more than the headline does. Adjusted basic earnings per share were US8.90 cents; adjusted diluted were US7.39 cents. The record quarter is 17.0% smaller per share once the instruments are counted, and a year ago that wedge was 15.5%. It is widening.

There is a second-order note worth making. The 2033 convertibles were issued in part, per note 12, "to repay the existing debts, share buy-back and for general corporate purposes," and weighted average basic shares did fall, from 12,265,154,456 to 12,077,851,410. Lenovo is retiring shares in the denominator while issuing instruments that will re-create them at HK$35.74. That can be good capital allocation — the repurchase is at today's price and the issuance is struck above it — but it is not share-count reduction, and it should not be read as such.

What the services number actually recognised

SSG is the segment carrying the recurring-revenue story, and the release describes "continued progress in shifting toward higher-value, recurring revenue streams," with the mix from Managed Services and Projects & Solutions "reaching a record high" of 62.4% of SSG revenue.

The revenue-recognition note is the check on that sentence. Revenue recognised over time — the accounting definition of the recurring kind — was $914.8M against $828.9M, up 10.4% in a quarter when the group grew 43%. As a share of revenue it fell, from 4.4% to 3.4%. Revenue recognised at a point in time rose 44.5% to $25,897.2M.

SSG is growing and its margin expansion is real. But the segment's growth is overwhelmingly landing in point-in-time recognition — project and solution delivery attached to hardware — rather than in the ratable line. The mix statistic and the recognition note describe the same business and point in different directions; only one of them is in the highlights.

The same discipline applies to the AI framing. AI-related revenue was $9.3B, 35% of the group, growing 60%. Against the company's own prior disclosures, that share is down from 38% in Q4 FY25/26, and the growth rate has stepped down from 105% across FY25/26 and 84% in Q4 to 60% here. AI revenue is compounding more slowly than the group it is credited with driving. The quarter's growth came from PCs, servers and geography — the Americas up 58.4% to $9,908.9M and EMEA up 52.8% to $6,397.6M — more than from the AI mix shifting.

One more on that theme. The AI server pipeline was given as $54.0B, up 157% quarter-over-quarter from the $21B disclosed at Q4. Asked what it contains, management said it includes "customer engagements, deals, future orders that are uncommitted or unqualified" with normalised conversion rates transcript, and declined to guide the conversion rate or the timeline. That is a candid answer and it should be read as given: a pipeline defined that way is a measure of engagement, not of backlog. The increment is not the base — a 157% move in a quarter is a statement about how the funnel is being counted at least as much as about what will be shipped.

The strongest case against this read

The operating-leverage case is genuinely strong, and if it is right the whole timing argument collapses.

ISG spent years structurally sub-scale. A server business carries fixed cost (manufacturing capacity, engineering, a global service footprint) that does not absorb at $4.3B a quarter, which is why the segment lost money there. At $8,510.1M it is a different business, and high-single-digit operating margins are not exotic for infrastructure at scale; they are roughly what the segment ought to earn. On that reading 9.1% is not a spike at all, it is the first quarter Lenovo was ever big enough to print a normal number. The mix supports it: Enterprise & SMB, structurally the richer half, nearly doubled year-on-year alongside CSP, and Lenovo says its China Enterprise & SMB business delivered record first-quarter revenue with "accelerated profit growth and margin improvement." Infinidat adds enterprise-storage margin permanently rather than once. And on this reading the inventory build is not a trade at all — it is the ordinary working capital of a business whose infrastructure revenue doubled, secured early by a supply chain Gartner just ranked in the global top five.

It is a coherent case and part of it is certainly true; ISG has clearly crossed a scale threshold. Where I still land differently is on the shape of the move rather than its direction. Operating leverage arrives as a climb — a segment grinds from 0.4% to 2% to 4% as absorption improves, which is roughly the path from the FY25/26 full year through the 3.6% Q4 print. What printed instead was 3.6% to 9.1% in one step, a 5.5 percentage point move in ninety-one days, in the same quarter that raw materials and WIP grew three times their normal seasonal rate and operating cash flow fell by more than half. Leverage does not travel that fast. A favourable input cost against a repriced output does.

The honest position is that both are present and the split is not disclosed. What I am claiming is the weighting, and the weighting is testable.

How this reads across the box-makers

Every server OEM is buying from the same constrained lane, so Lenovo's quarter is a read on the cohort rather than a Lenovo-specific event.

• Dell — the closest comparison and the one this newsletter has already graded. Dell's Q2 FY27 showed the same structural pressure from the other end: margin flattered by an AI mix that had not yet converted, cash flow going the wrong way, and the composite hiding the component. Lenovo's print says a well-timed component book can offset that pressure for at least one quarter. It does not say for how many.
• HPE — carries the same input exposure with less of Lenovo's PC volume to spread procurement across. Whatever advantage sits in buying power and timing, Lenovo has more of it.
• SMCI — mix is the softest margin to capitalize was coined here, and Lenovo's quarter is the same caution at ten times the revenue. A margin that arrives with an inventory build attached is a margin to confirm, not to extrapolate.
• The component suppliers — the read-through is directional and aggregate. A buyer of Lenovo's scale pre-positioning this heavily is evidence that the supply lane was tight enough to be worth pre-positioning against, and that buyers with balance sheets have been willing to fund it. Lenovo itself expects the imbalance to persist, with Yang telling the call it lasts "at least by the end of next year" transcript.

For the PC side there is a second-order point in Lenovo's own guidance. Management expects the PC market to contract roughly 15% in units and smartphone shipments to fall about 20% in the back half of the fiscal year transcript, and expects to offset that with price and premium mix — IDG president Luca Rossi said Lenovo will "continue to offset the units decline with higher ASP, improved premium mix, improved AI PC mix" transcript. Read the ingredients, not the level: a revenue line held flat by price while units fall 15% is a different business from one growing 27%, and IDG's 7.1% margin has never been tested against falling volume and rising input cost at the same time.

Valuation, and what settles it

0992.HK traded at HK$32.38 (~US$4.15) on September 4, 2026, for a market capitalisation of about HK$399.7B (~US$51.2B), inside a 52-week range of HK$8.52 to HK$36.16 unverified. Rate used throughout: ~7.80 HKD/USD, approximate.

The multiple is not the interesting part; the denominator is. On 16,081,730,874 fully diluted shares rather than 12.4 billion, the same market capitalisation implies a company about 29.6% larger than the headline share count suggests — and the instruments creating that gap are struck between HK$8.37 and HK$11.51 for the in-the-money majority. Annualising a single quarter is not analysis, but it frames the question: the market is capitalising a run-rate that depends on whether $776.9M of quarterly infrastructure profit is a level or a peak, on a share count roughly a third larger than the one most screens display.

Equity attributable to owners actually fell over the quarter, from $7,635.0M to $7,213.4M, because the reported loss ran through reserves. A company can print its best operating quarter ever and end it with less book equity than it started with; this one did.

The dated test. The quarter that settles this is Q3 FY26/27 — the December quarter, reported around February 2027. Two conditions, and they resolve together:

• If ISG's operating margin holds above 6% while raw materials and WIP stop outgrowing revenue, the margin was scale and mix, the leverage case was right, and this read was wrong. Say so plainly at that point.
• If ISG's operating margin falls toward the low single digits while ISG revenue still grows, the margin was the warehouse, and the second turn of the inventory was bought at the market.

The intermediate print, Q2 FY26/27 in November, is the early tell rather than the verdict: watch whether the raw-materials-and-WIP line keeps compounding faster than revenue, and whether cash from operations recovers toward the $1.5B this business used to generate in a first quarter.

The trade

Positioning follows from the share count more than from the margin. Anyone sizing this name off a screen is sizing it on 12.4 billion shares; the disclosed instruments say the correct denominator is closer to 16.1 billion, and the difference is not a tail risk — it is contractual, disclosed, and mostly struck at a third of the current quote. Size on the diluted number.

For holders, the asymmetry is genuinely two-sided and sits entirely in the Q3 test above, which is why a structure that gets paid for time rather than for direction fits the setup better than an outright view does — an overwrite against the elevated post-run volatility, with the expiry chosen to sit past the February print rather than before it, converts the wait into carry. Nothing here is a recommendation, and the compliance line at the foot of this page is not decoration: this is analysis of a filing, not advice.

What would change the read fastest is not a price move but a disclosure — any quarterly split of ISG profitability between CSP and Enterprise & SMB, or any inventory-cost commentary that lets an outsider separate procurement timing from operating leverage. Lenovo does not currently provide either, and the absence is why this comes down to a dated test rather than a calculation.

Bottom line

This was the best quarter in Lenovo's history and the arithmetic behind it is not in dispute: $26,943M of revenue, $1,075M of adjusted profit, and an infrastructure business that earned more in ninety-one days than it had in the previous four quarters combined, several times over. The execution that produced it — securing components ahead of a tightening lane at the scale of a $100B-run-rate company — is a real competitive advantage and the hardest thing in this industry to do.

It is also, on the evidence of the balance sheet, substantially a timing win. Raw materials and work-in-progress grew 118.3% against revenue at 43% and finished goods at 45.2%; the March-to-June build ran at roughly three times its normal seasonal rate; operating cash flow fell 54% in the same quarter adjusted profit rose 176%. Those four facts describe a position, and a position pays once.

Management says the margin reflects "business performance, not transitory factors," and the release attributes the entire group gross-margin gain to ISG while never mentioning the acquisition that closed inside the quarter or the inventory that funded it. The claim may still prove right — the operating-leverage case is real and ISG has genuinely crossed a scale threshold. But it is a claim about the next four quarters, not a fact established by this one, and the company has given the market no disclosure that would let anyone separate the two before February.

The June dive asked whether the margin turn would stick. It stuck. The question this print leaves is narrower and sharper: whether Lenovo earned that margin, or bought it early.