AutoStore Holdings

Phase A — Understand the business

Lens 1 · Company Overview

What it is. AutoStore invented and commercialized cube storage — the densest of the four "light" AS/RS (automated storage & retrieval) architectures. Instead of racks and aisles, inventory lives in HDPE bins stacked directly on top of each other inside a self-supporting aluminium grid; battery-powered robots drive across the top of the grid, dig the required bin out of the stack, and deliver it to a worker (or a robotic picker) at a port. The pitch is ~4x the storage density of conventional shelving in the same footprint, with 99.8% uptime. It is the clear global category leader: ~1,950 installations across 65+ countries, 1,300+ unique customers, having crossed 300 installs in North America in 2025.

How it makes money — the load-bearing fact. AutoStore is a 100% indirect, partner-led OEM. It does not sell, install, or service systems itself. It manufactures the hardware (grid, bins, robots, ports, controller) plus the software, and sells through a network of certified distribution partners / system integrators — Element Logic (the original and largest), Swisslog (KION's integrator arm, 400+ AutoStore projects), Syncore (170+), Hörmann, Kardex, and others — who do the customer-facing solution design, integration, and lifecycle support. This is the single most important structural fact about the business: it is capital-light and high-margin precisely because the integrators absorb the labor, the project risk, and the customer relationship. The flip side (Lens 3, 13): AutoStore does not own the end customer.

Contract structure. Historically a capex sale — lumpy, project-based, recognized on delivery. The strategic shift since ~2024 is toward AutoStore-as-a-Service (ASaaS): 7–10-year recurring contracts that convert the box sale into a subscription, plus software subscriptions (the "Essentials" package, and the 2026 CubeVerse cloud platform) that layer annual recurring revenue on top of the installed base. 60% of 2025 revenue came from existing customers — a meaningful land-and-expand signal for a hardware OEM.

Customers. End users span e-commerce, grocery/retail, 3PL/logistics, healthcare/medical (Medline was the 300th NA install), industrial parts, and now cold-chain (2025 Multi-Temperature launch). A notable 2025/26 deal: a $200M framework agreement with a U.S. 3PL for 50 cube systems across 25 sites by 2026 under a RaaS model — exactly the recurring-revenue flagship the new strategy needs.

Lens 2 · Supply Chain

Map: upstream inputs → AutoStore → integrator → end customer.

• Upstream inputs (named where sourceable): aluminium extrusions (the grid), HDPE (the bins), electric motors / drives, batteries (the robots are battery-electric and famously low-power — "10 robots ≈ one toaster" ), bearings/wheels, controller electronics and chips. AutoStore's manufacturing is anchored in Norway (Nedre Vats), Poland, and the US (it expanded US assembly to cut North-American lead times). Specific tier-1 component suppliers are not publicly disclosed — n/a — not disclosed.
• The company: designs, assembles, and ships the standardized hardware "kit" plus the software. This is its chokepoint of value — the IP and the manufacturing scale.
• Integrator layer (the real distribution chokepoint): Element Logic (largest; acquired ABCO Systems US in Apr-2025 and BS Handling UK in Sep-2025 to expand reach), Swisslog (KION), Kardex, Syncore, Hörmann. These partners are simultaneously AutoStore's distribution moat and its single biggest dependency — Element Logic and Swisslog concentrate a large share of throughput, and Swisslog's parent KION (via Dematic) is a direct competitor in the broader AS/RS market. That is a structural conflict baked into the model.
• End customer: the warehouse operator.

Chokepoints / single-source risks: (1) integrator concentration — a falling-out with, or capacity constraint at, Element Logic or Swisslog hits revenue directly; (2) aluminium/HDPE commodity input cost (a margin, not supply, risk); (3) no second source on the core IP — AutoStore is the single source of the AutoStore system, which is the moat, but also means the whole P&L rides one product architecture. Names or it didn't happen — the named chain above is the differentiator vs. a generic "robots and bins" description.

Lens 3 · Competitive Advantages (moats)

The moat is real but narrower than the share price implies. Four pillars:

1. Density + a 25-year reference base. Cube storage is the highest-density light-AS/RS architecture, and AutoStore has the deepest installed base (1,950+) and operating-hours dataset in the category. For a risk-averse warehouse buyer spending millions on a 10-year asset, "the system that 1,300 other companies run at 99.8% uptime" is a powerful default.
2. Integrator ecosystem (the strongest moat). A certified network of integrators who have trained staff, spare-parts depots, and reference projects is extremely hard to replicate — it took AutoStore 20+ years. A new entrant must build that channel from zero. This is the switching-cost / distribution moat.
3. Switching costs at the customer. Once a cube grid is installed and a warehouse's processes are built around it, ripping it out for a competitor is a multi-million-dollar, multi-month disruption. Expansions go to the incumbent — hence 60% of revenue from existing customers.
4. IP — now de-risked. The Ocado patent war (Lens 10) ended in a 2023 cross-license of all pre-2020 patents, which paradoxically strengthened AutoStore: the existential "an injunction kills the product" risk is gone, and both incumbents are now insulated against each other while new entrants face both patent estates.

Bargaining power. Over suppliers: moderate-to-high (commodity inputs, multiple sources). Over customers: high at the expansion stage (lock-in), but weak at the initial-win stage because the integrator — not AutoStore — controls the customer relationship and could, in principle, steer a deal to Exotec or a KION/Dematic solution. AutoStore's pricing power is real on the installed base, thinner on net-new competitive bids.

Why the moat is narrower than bulls think (preview of Lens 13): cube storage is one architecture among several, the buyer ultimately compares total-cost-of-ownership across Exotec shuttles, Symbotic, Ocado, and increasingly cheap Chinese ASRS — and "highest density" is not always the deciding spec.

Lens 4 · Segments

AutoStore does not report multi-segment P&L by product line; it is effectively a single-product company. The meaningful cut is by geography and by revenue type.

By geography (FY25 / recent quarters):

• EMEA — the dominant region, >70% of revenue in Q3'25 (Q2'25: ~$98M of ~$134M).
• North America — 24% of FY25 revenue, but >30% of Q4'25 order intake — i.e. NA is the growth vector, mix-shifting up (Q2'25: ~$29M).
• APAC — small (~$7M in Q2'25), the underpenetrated long tail.

The trend: Europe is the mature base; North America is the accelerant (order-intake mix running ahead of revenue mix), and APAC is optionality. This geographic skew matters for the bull case — the US warehouse-automation market is larger and less penetrated than Europe.

By revenue type: transitioning from ~100% capex/box to a growing recurring/ASaaS + software layer (Essentials package, CubeVerse, RaaS frameworks). Exact recurring-revenue $ is not cleanly disclosed — n/a — not separately broken out; management frames it qualitatively ("building a base of predictable recurring revenue," 7–10yr contracts).

---

Phase B — Measure performance

Lens 5 · Earnings Result (latest print — Q1 2026, reported 2026-04-23)

The latest print is the recovery quarter the bulls were waiting for, but it must be read against a savagely depressed comparison base.

• Revenue: USD 165.8M, +92.9% YoY. The +93% is almost entirely a base effect — Q1'25 was the trough at $85.9M (−37.8% YoY). So the right read is "revenue recovered to a normalized run-rate," not "revenue nearly doubled."
• Order intake: USD 179.4M, +27.0% YoY, and book-to-bill > 1 (intake exceeded revenue by $13.6M). This is the genuinely bullish line: bookings are outpacing recognition and now include projects scheduled for 2027 shipment — i.e. customers are committing further out, a confidence signal.
• Backlog: USD 570.6M, a record, up from the FY25 close of ~$557M.
• Gross margin: 72.7%; adjusted EBITDA margin: 44.0% — best-in-class hardware economics, holding near peak even at a lower revenue base.
• Cash conversion: 81.9% — the "highly cash-generative" claim is genuine.
• Market reaction / what was priced in: coverage framed Q1'26 as an "expectation-arbitrage" print — strong absolutes but the question is whether the recovery was already in the ~+140% 12-month share move (52-wk range 5.17→13.90 NOK). The stock did not melt up on it, suggesting much of the recovery is priced.

Flag vs. own history: the standout anomaly is the violence of the 2024→2025 downturn then snap-back — Q1'25 −38%, then sequential rebounds (+56% QoQ in Q2'25), then Q1'26 +93% YoY. This is a cyclical, project-timing-driven revenue profile, not a smooth compounder. Margins, by contrast, are remarkably sticky (43–48% adj EBITDA across the cycle) — the operating-leverage story is real, but so is the cyclicality.

Lens 6 · Earnings Calls (sentiment trend)

No transcripts in the research layer; this is web-derived.

Tone arc across the last ~5 quarters:

• Q1–Q2 2025 (trough): defensive — "challenging market conditions," "cautious customer investment," project slippage; management leaning on backlog for visibility and emphasizing margin discipline.
• Q3–Q4 2025: stabilizing — "strong sequential growth amid uncertain market conditions"; recurring-revenue and product-launch narrative ramps (five new products in spring '25, ASaaS, cold-chain).
• Q1 2026: the most confident in two years — "securing longer-term commitments," intake into 2027, software/CubeVerse as the new growth axis.

Recurring phrases: "cash-generative," "market-leading margins," "partner ecosystem," "recurring revenue," "too early to predict full-year." What they stopped saying: the acute 2023–24 destocking/slippage language. What they will not yet say: a hard 2026 revenue guide — management has explicitly declined to guide the year ("too early to predict 2026 performance"), which is itself a tell that visibility, while improving, is not yet conviction-grade.

Lens 7 · Comps

Pure-play cube-storage peers are mostly private (Exotec) or embedded in conglomerates (Dematic in KION, Swisslog in KUKA, Intelligrated in Honeywell). The cleanest listed comp is Symbotic. Multiples are `` with source/date or n/a. No multiple below is fabricated.

Read: on an absolute basis AutoStore screens expensive — Simply Wall St flags ~32x P/E vs a ~21x peer average and EV/EBITDA in the ~91st sector percentile. The bull rebuttal is that AutoStore's margin structure (73% GM, 43% EBITDA) and asset-light model deserve a software-like premium the metals-bending peers don't. Both are true; the tension is the whole debate. The honest peer-multiple table is thin because the best comps are private/embedded — a real limitation of this lens for this name.

Lens 8 · Stock-Price Catalysts (>5% moves, last ~5 years)

Pattern from:

• Oct 2021 — IPO at a ~$12.4B valuation, Norway's biggest in two decades. The high-water mark.
• 2022 — down ~50% YTD by May despite earnings beats; the rate-shock/growth-derating + SoftBank Vision Fund writedowns dragged it. Lesson: macro/rates and the SoftBank overhang moved it more than fundamentals.
• 2023 — order-intake scares ("new order intake remains concerning," Sep-2023) repeatedly hit the stock even when revenue/margins were fine. Lesson: the market trades AutoStore on order intake, not the trailing P&L.
• 2023 — Ocado settlement (£200M payout) removed a tail risk.
• 2024–25 — the downturn dragged the stock to a 5.17 NOK 52-wk low.
• 2025–26 — the recovery doubled it back to ~12–14 NOK as order intake inflected.

What the market actually reacts to: (1) order intake / book-to-bill above all (the leading indicator of the cyclical box business), (2) macro/rates as a duration asset, and (3) the SoftBank overhang (any hint of a block sale). Earnings beats alone do not move it sustainably — this is an order-intake stock.

---

Phase C — Judge people & books

Lens 9 · Management

• CEO — Mats Hovland Vikse (since Jan-2023). A Norwegian operator who took the helm into the downturn. Track record: he has steered the company through the worst order-intake collapse in its history while defending 43–48% adj EBITDA margins and pivoting strategy toward software/ASaaS — a competent defensive-into-offensive transition, though he has not yet had to prove he can deliver durable top-line growth (the cycle did the 2026 rebound as much as he did). Pre-AutoStore he ran businesses in industrial/tech; he is a professional manager, not the founder.
• Founder note: the cube-storage concept dates to Jakob Hatteland (founder, Hatteland Group) in the 1990s; the business has long since passed through PE (THL/EQT) and SoftBank ownership into a professionalized public company. No founder remains operationally central — relevant because the "founder-zealot" energy is gone; this is a managed asset.
• Bench (2025–26 build-out): Parth Joshi (CPO, ex-Cisco/Eaton) hired to own the product roadmap — a clear signal the software pivot is being staffed seriously; Carlos Fernandez (Chief Solutions Officer), Mike Dickson (SVP Global BD), Anette Matre (CHRO). The CPO hire from networking/industrial software is the most strategically telling appointment.
• Skin in the game / insider ownership: management holds modest equity; the cap table is dominated by financial owners (SoftBank ~40%, ex-PE residuals), not insiders — insider ownership n/a — not separately sourced this run. This is a financially-owned company, not a founder-aligned one.
• Capital allocation: disciplined and conservative — leverage policy ≤2x Net Debt/Adj EBITDA, net debt ~$180M (well inside ~0.8x), no dividend yet (0.00% yield) and no buyback decided as of Q1'26; cash is going to deleveraging + reinvestment (software, new TAMs). The one large historical "allocation" was involuntary — the £200M ($256M) Ocado settlement (Lens 10), a self-inflicted cost from an over-aggressive patent strategy that lost in the UK High Court.
• Red flags: (i) the board is partly SoftBank-appointed (recent swaps: Andreas Hansson → Angela Du / Kevin Mok, Aug-2025) — governance reflects the financial owner, not public minority holders; (ii) the failed Ocado litigation is a genuine capital-allocation/judgment black mark; (iii) declining to guide FY26 — defensible given cyclicality, but it limits accountability.

Net: a competent, margin-disciplined professional-management team, correctly pivoting to software — but owner-aligned to SoftBank, not founders, and carrying one real strategic misjudgment (the patent war) on the record.

Lens 10 · Forensic Red Flags

Accounting risk surface (web-only — no filings to forensically tie out, which is itself a limitation; figures /):

• Revenue recognition / cyclicality dressed as quality. The biggest "forensic" issue is not fraud risk but interpretation: management and bulls frame AutoStore as a steady compounder, but the income statement is project-lumpy and cyclical (Q1'25 −38%, Q1'26 +93%). Order intake and backlog are the honest leading indicators; trailing revenue can flatter or mislead depending on where you are in the cycle. Watch for backlog quality — "intake into 2027" is good, but framework agreements (the $200M 3PL RaaS deal) can convert to revenue more slowly/uncertainly than a firm box order.
• ASaaS transition optics. The shift to 7–10yr recurring contracts defers revenue vs. an upfront box sale. In transition years this can understate reported revenue while building deferred/contract balances — the opposite of an aggressive-recognition risk, but it makes YoY comparisons noisier. Watch contract-liability / deferred-revenue growth as the real ASaaS scoreboard (not separately sourced this run).
• Margin sustainability. 73% gross / 43% adj EBITDA is exceptional for hardware. The "adjusted" EBITDA bridge (what's excluded — SBC, restructuring, the Ocado payment in prior years) deserves scrutiny; non-GAAP adjustments are where OEMs flatter. Adjustment detail n/a — not parsed this run.
• Cash flow vs. earnings. Here AutoStore is clean and reassuring: 76–82% cash conversion, FCF supports the deleveraging story — cash flow corroborates earnings rather than diverging from it. Positive.
• Goodwill/intangibles, FX, Bermuda domicile. Bermuda incorporation + Norway HQ + USD reporting introduces FX translation noise and a tax/governance structure worth understanding but not inherently a red flag. Goodwill from the Berkshire-Grey-linked / partner integrations is small relative to the balance sheet (n/a — not sized this run).

Regulatory findings (required sub-section).

• SEC (EDGAR LR + AAER): None possible — AutoStore has no CIK and does not file with the SEC. Per regulatory/regulatory-findings.md (fetched 2026-06-30), total_sec_findings: 0.
• Litigation of record: the AutoStore v. Ocado global patent war (filed Oct-2020) — AutoStore asserted six patents; the UK High Court ruled AutoStore's patents invalid (prior public disclosure to the Central Bank of Russia before filing), AutoStore lost, and the dispute settled July 2023 with AutoStore paying Ocado £200M (~$256M) over 24 monthly installments and a mutual cross-license of all pre-2020 patents; Ocado retained exclusive rights to the Single Space Robot. Material, resolved, fully (or nearly) paid by 2025. No new AutoStore–Ocado litigation surfaced in 2025–26 search.
• Non-SEC enforcement (FTC/DOJ/FDA/EU/etc.): web search surfaced no material enforcement actions, consent decrees, or fines against AutoStore Holdings.
• Conclusion: No material regulatory or accounting-fraud findings — verified via SEC EDGAR EFTS (LR, AAER, returned 0 / no-CIK), web search, and public litigation record as of 2026-06-30. The one historical material legal event (Ocado) is resolved. Treat the cyclical-revenue interpretation, the "adjusted" EBITDA bridge, and the ASaaS deferred-revenue mechanics as the items to watch — none rise to a forensic red flag on current public information.

---

Phase D — Project & stress-test

Lens 11 · Forward Projection

No forecast.ts create is logged — this is an unattended --watchlist run and the SKILL bars it (and bars a committed base case here). Figures are `` with arithmetic shown; not a tracked forecast.

Anchor (actuals, ):** FY25 revenue **$538.6M** (−10.4% vs FY24 $601.4M); FY23 **$645.7M** (cycle peak). FY25 adj EBITDA margin **43.3%** (~$233M ); FY25 net income $81.8M. Shares ~3.36B**; so FY25 EPS ≈ $0.024 (≈ NOK 0.25). Q1'26 annualized run-rate ≈ $663M revenue, though seasonality makes a straight ×4 a ceiling.

Three-year EPS path (FY26 / FY27 / FY28), every line labeled:

• Base — recovery continues but normalizes off the +93% Q1 base effect. FY26 revenue ~$620–650M (+15–20% YoY); adj EBITDA margin held ~43%; net margin ~15% → FY26 net income ~**$95–100M** → **EPS ≈ $0.028–0.030**. FY27 +12–15% rev on software/ASaaS uplift + NA share gains → EPS ≈ $0.033–0.036; FY28 +10–12% → EPS ≈ $0.037–0.041 ``.
• Bull — the cycle re-accelerates, ASaaS/CubeVerse adds a high-margin recurring layer, NA and cold-chain TAM expansion lift growth back toward the high-teens/low-20s%; margins tick toward 45%+ on software mix. FY28 revenue ~$800M+, EPS ≈ $0.050–0.055 ``.
• Bear — warehouse-automation capex stalls again (rates, weak consumer-discretionary/e-commerce capex), integrators steer share to Exotec/Chinese ASRS, software adoption is slower than hoped. FY26 flat-to-down, margins compress toward high-30s on under-absorption; FY28 EPS ≈ $0.018–0.022 `` — i.e. a return to the trough.

The number that matters: at 12.4 NOK ($1.23) and base-case FY28 EPS ~$0.038, the stock is on ~32x trailing / ~20x forward earnings. The market is paying a software multiple for a business that has just demonstrated capex-OEM cyclicality. The forward case only works if (a) the recovery sustains and (b) the software/ASaaS mix genuinely re-rates the durability of earnings. Both are plausible, neither is yet proven.

Lens 12 · Bull vs Bear

Bull case. AutoStore is a category-defining monopoly in the densest AS/RS architecture, with a moat (1,950 installs, the integrator ecosystem, switching costs) that compounds. It earns 73% gross / 43% EBITDA margins with 80% cash conversion — software-like economics from a hardware base — and is structurally under-penetrated: ~80% of the world's warehouses are still non-automated, the labor-cost case for automation only strengthens, and AS/RS is >30% of a warehouse-automation market growing ~18% CAGR toward ~$60B by 2030. The 2026 software pivot (CubeVerse, ASaaS, Essentials) is the re-rating catalyst — it converts a lumpy box into recurring revenue and a higher-quality multiple, with the CPO hire signaling intent. Order intake inflected (+27% YoY, book-to-bill >1, into 2027), North America (the bigger market) is mix-shifting up, and the existential patent risk is gone. Capital allocation is conservative (sub-1x levered, deleveraging, optionality for buybacks/dividends).

Bear case (2–3 permanent-impairment risks). (1) It is a cyclical capex-OEM, not a compounder — the −38%/+93% swing proves revenue rides customer capex cycles AutoStore doesn't control; a multi-year automation-capex winter (rates, weak e-commerce capex) caps the stock regardless of moat. (2) Architecture risk — cube storage is one approach; Exotec's shuttle system, Symbotic, Ocado, KION/Dematic, and rapidly-cheapening Chinese ASRS (Hai Robotics, Geek+) all compete for the same buyer on TCO, and AutoStore doesn't own the customer (the integrator does), so it can be designed out of bids. (3) The SoftBank overhang — a ~40% holder that has written this asset down and needs liquidity is a permanent supply-of-stock cap on any rally. Expectations baked in: ~20x forward / ~91st-percentile EV/EBITDA prices a clean software-grade recovery; a single soft-intake quarter de-rates it hard (the market trades it on intake — Lens 8).

Pre-mortem (it's late 2027, the thesis broke — what happened?). The 2026 recovery turned out to be a backlog-burn / base-effect blip; FY27 order intake rolled over as e-commerce/3PL capex paused; ASaaS adoption was slower and dilutive to near-term reported revenue; an integrator routed two flagship deals to Exotec's Next-Gen Skypod; and SoftBank placed a block at a discount, capping the stock. The "software re-rating" never arrived because CubeVerse stayed a feature, not a P&L line. Net: a great business, but bought at a cyclical-peak multiple.

Are multiples too high? Yes, on current evidence. ~32x trailing P/E and top-decile EV/EBITDA demand that the software/durability story is already true. It isn't yet. The business quality justifies a premium; it does not yet justify this premium at this point in the cycle.

Contrarian view (what the market refuses to see). The market is debating "great moat vs. expensive multiple." The thing it's underpricing is that AutoStore's distribution moat is also its ceiling: the integrator network that makes it capital-light also means AutoStore is a component supplier that can be substituted inside a partner's solution — so the software pivot (CubeVerse) isn't just upside, it's existential: owning the orchestration/data layer is how AutoStore stops being a swappable box and becomes the indispensable platform. If CubeVerse works, the bull multiple is justified. If it doesn't, this is a cyclical Norwegian metals-and-motors OEM trading at a tech multiple.

Lens 13 · Devil's Advocate (short-seller)

Dismantling the bull case.

• What structurally breaks the money machine: AutoStore sells through integrators it does not control, several of which (Swisslog/KION, Kardex) are tied to or compete with rival AS/RS platforms. The day an Element Logic or Swisslog decides Exotec or an in-house Dematic solution wins a deal, AutoStore loses it with zero visibility. The "moat" of the ecosystem is rented, not owned.
• Revenue concentration: concentrated in EMEA (>70%) and in a handful of large integrators. A European warehouse-capex slowdown or an integrator falling out hits revenue directly. Geographic and channel concentration are both real.
• Why the moat is weaker than bulls think: "highest density" is a spec, not a verdict — buyers optimize total cost of ownership, throughput, flexibility, and implementation speed, and Exotec explicitly markets faster, more modular deployment vs. traditional AS/RS. Meanwhile Chinese ASRS (Hai Robotics, Geek+, HIKRobot) are collapsing the price floor — the same dynamic that gutted other Western hardware categories. AutoStore's premium is exposed.
• Most dangerous underestimated competitor: Exotec (closest cube/shuttle analog, unicorn, Next-Gen Skypod) on the high end, and Chinese ASRS on price. Symbotic (Walmart-scale) and KION/Dematic (the largest player, $2.5B+) bracket it.
• Worst capital-allocation move: the Ocado patent war — AutoStore sued, lost in the UK High Court on its own prior disclosure, and paid £200M. That is a self-inflicted quarter-billion-dollar judgment error that says something about prior management aggression.
• Assumptions that must hold for today's price: (i) the 2026 recovery is a sustainable up-cycle, not a backlog-burn blip; (ii) ASaaS/software re-rates earnings durability and contributes margin, not just optics; (iii) integrators stay loyal; (iv) SoftBank doesn't dump. Break any one and ~32x P/E is indefensible.
• If growth disappoints 20–30%: revenue back toward the ~$430–480M trough zone, adj EBITDA margin compresses toward high-30s on under-absorption, EPS halves toward ~$0.018–0.020, and a ~32x multiple compresses toward a cyclical-OEM 12–15x — a 40–60% drawdown is entirely plausible, as the 52-wk low of 5.17 NOK (vs ~12.4 now) demonstrates the range.
• Single permanent-impairment scenario: a step-change in a rival architecture's TCO (Exotec or a Chinese entrant reaching parity-density at materially lower cost) that turns AutoStore from "category default" into "one option among equals," eroding pricing power on the installed base. Plausibility: low-to-moderate over 3 years, rising thereafter — the installed base and switching costs buy time, but not permanence.

Lens 14 · Management Questions (ordered by information value)

1. ASaaS now: what % of FY25/Q1'26 order intake and revenue was recurring/ASaaS vs. upfront capex, and what is the contract-liability/deferred-revenue balance — i.e. show us the recurring scoreboard, not the narrative.
2. What share of your deals are competitive bids (vs. installed-base expansions), and what is your win rate vs. Exotec and vs. Chinese ASRS in those competitive bids over the last 8 quarters?
3. How much of the $570M backlog is firm box orders vs. framework agreements (like the $200M 3PL RaaS deal) that can slip or not convert — and what's your historical framework-to-revenue conversion rate?
4. CubeVerse: is it a feature that helps sell hardware, or a standalone P&L line? What is its attach rate, ARPU, and contribution-margin target — and at what year does software become >10% of revenue?
5. Channel risk: what is your revenue concentration in your top 3 integrators, and what contractual exclusivity (if any) protects you when an integrator like Swisslog is owned by a competitor (KION/Dematic)?
6. SoftBank holds ~40%: what is your understanding of their intended holding period, and is there any orderly-distribution / lock-up framework that protects minority holders from a block overhang?
7. Walk us through the adjusted-EBITDA bridge: what exactly is excluded, how much is SBC, and what is the FY25 unadjusted operating margin?
8. The 2024–25 downturn: what specifically drove the −38% Q1'25 — destocking, project slippage, lost share, or end-demand? How much has structurally recovered vs. cyclically?
9. Geographic mix: EMEA is >70% of revenue but NA is >30% of intake — what investment is required to flip NA to your largest region, and what's the margin profile of NA vs. EMEA deals?
10. Pricing power: as Chinese ASRS prices fall, what is your premium vs. the cheapest credible alternative on a TCO basis, and is that premium widening or narrowing?
11. Capital allocation: with sub-1x leverage and 80% cash conversion, why no buyback or dividend yet — and what return hurdle are reinvestment opportunities (software, cold-chain, case-handling) clearing that returning capital wouldn't?
12. Multi-Temperature / cold-chain and case-handling: what is the realistic incremental TAM and the timeline to a material (>5% of revenue) contribution?
13. Post-Ocado: how have you changed your patent strategy and freedom-to-operate process so the £200M judgment-error can't recur, and how defensible is the current IP estate against new entrants?
14. What is the single biggest threat to AutoStore's business model over the next five years that the market is not pricing — and what are you doing about it?
15. Founder energy is gone and the company is financially owned: how do you sustain the innovation pace of a category creator as a professionally-managed, PE/SoftBank-owned public company?

---