Oxford Nanopore Technologies

Phase A — Understand the business

Lens 1 · Company Overview

Oxford Nanopore makes DNA/RNA sequencing technology built on a fundamentally different physical principle from the incumbent. Where Illumina reads short fragments by detecting fluorescent flashes as bases are synthesized ("sequencing-by-synthesis"), Oxford Nanopore threads a native single molecule of DNA or RNA through a protein nanopore and reads the electrical-current disruptions each base causes. That gives it three structural product attributes the chemistry of the incumbent can't easily replicate: arbitrarily long reads (up to a megabase vs. ~150–300bp short reads), real-time streaming of data as the molecule passes, and direct detection of base modifications (methylation/epigenetics) without a separate prep — because it reads the native strand, not a synthesized copy.

The business model is razor-and-blade. Customers buy a sequencing device once — from the $90 Flongle and the pocket-sized MinION, up the ladder to the desktop GridION and the high-throughput PromethION (which runs up to 24 flow cells at once) — and then repeatedly buy the consumable flow cells and reagent kits that the runs consume. Critically, the company deploys nearly all its improvements through the consumable, software, or firmware — "so that devices do not need to be changed". This means (a) the installed base is sticky and self-upgrading, and (b) consumables are the recurring revenue engine, not instruments.

Customers / end-markets. Management segments demand by funding source, not product: in H1 2025, ~68% Research (academic/government/public-health/grant-funded), ~13% Clinical (diagnostic/prognostic value), remainder Applied/industrial. The fastest-growing slices in FY2025 were Clinical +59%, BioPharma +30%, Applied/Industrial +27%, with core Research +15% — i.e. the mix is deliberately shifting toward higher-value, stickier applied and clinical uses and away from pure grant-funded research.

Anchor relationships are large population-genomics and public-health programs: UK Biobank (transitioned from pilot to production to generate the first large-scale methylome dataset), Genomics England (Cancer 2.0, PRECISE), and a landmark UK Government partnership to stand up a real-time, pathogen-agnostic biosurveillance network across up to 30 NHS England hospitals. The historical cautionary tale here is the Emirati Genome Program — a single large contract whose renegotiation triggered a 2023 margin warning (see Lens 8) — illustrating that flagship-program concentration cuts both ways.

Lens 2 · Supply Chain

n/a — research-layer supply-chain.md missing (kb/genomics/wiki/ not yet compiled); mapped from web disclosure.

Upstream inputs → ONT → end customer:

• The sensor. The heart of every flow cell is a bespoke Application-Specific Integrated Circuit (ASIC) paired with a printed circuit board (PCB), a moulded-plastic fluidic chamber, a fluidic gasket, and the arrayed nanopore sensor itself. The ASIC controls and measures thousands of simultaneous channels.

Chokepoints / single-source dependencies: (1) the un-named ASIC foundry — geopolitically exposed given the US/China semiconductor backdrop already cited as a China growth drag; (2) protein/chemistry production is single-source by design (in-house) — a quality excursion there (note the H1-2025 one-off £3.3m inventory charge ) flows straight to gross margin; (3) flagship-program concentration on the demand side (Emirati Genome, UK Biobank, Genomics England) — losing or pausing one materially moves a half-year.

Lens 3 · Competitive Advantages (moats)

The moat is real but narrow and technology-specific. Oxford Nanopore is effectively the only at-scale commercial nanopore-sequencing platform in the Western market, and it owns a category — long-read + real-time + native-modification — that the dominant incumbent (Illumina) structurally cannot serve with its existing short-read chemistry.

Durable moat sources, ranked:

1. IP + trade-secret estate around nanopore chemistry and the ASIC. ONT describes a "substantial portfolio of proprietary rights" and is actively litigating to defend it against BGI/MGI (Lens 10). Decades of protein-engineering know-how is hard to clone.
2. Installed-base switching costs + self-upgrading razor-and-blade. Once a lab's workflows, bioinformatics pipelines, and trained staff are built around nanopore data, and the device keeps improving via consumables, switching to short-read means re-tooling. PromethION grew ~59% / >40% YoY — the high-throughput tier is where lock-in deepens.
3. A category short-read can't enter. For structural variants (LRS found 2.86× more SVs than short-read, excelling >6kb), full-length isoforms, repetitive/complex regions, real-time field/clinical use, and direct epigenetics, nanopore is differentiated rather than substitutable.

Where the moat is weaker than bulls think: on the core accuracy axis that most routine clinical/SNV work cares about, short-read is still the gold standard (Illumina SNV F-measure 0.967 vs. ONT 0.954 in high-complexity regions), and ONT's direct rival in long-read, PacBio (HiFi), competes on accuracy. So ONT's moat is strongest in the "long/real-time/native-mod" niche and contested everywhere it tries to push into mainstream high-accuracy applications.

Bargaining power: modest. Against suppliers — strong on chemistry (in-house) but weak/unknown on the ASIC foundry. Against customers — strong with small labs (no alternative nanopore vendor), weaker with the giant national programs that can dictate price (the Emirati renegotiation proved buyers have leverage on the biggest contracts).

Lens 4 · Segments

Hard requirement noted: segments.csv is an empty stub — all figures, not.

By funding-source segment (FY2025, growth at constant currency):

By product: the PromethION range grew ~59% (>40% YoY) — the high-throughput tier is the growth driver; the "Life Science" product line reached ~£105m, +28% cc in H1.

By geography (FY2025, cc): EMEAI +26.3% (largest region) and APAC strong but with a China drag — APAC ex-India revenue ~£40.4m, +19%, explicitly held back by China. H1 2025 had EMEAI and APAC both >30%, so the H2/2026 message is a deliberate EMEAI deceleration as big strategic research projects conclude before new ones ramp (the core of the Jan-2026 guidance trim — Lens 8).

Read-through: the segment story is a company actively rotating its revenue base from lumpy grant-funded research toward recurring clinical/applied demand — strategically correct, but it means near-term growth is hostage to the timing of a handful of large programs (EMEAI project "air-pocket" into 2026).

---

Phase B — Measure performance

Lens 5 · Earnings Result (FY2025 — reported 2 March 2026)

The latest full-year print (FY ended 31 Dec 2025, prelims 2 Mar 2026):

• Revenue £223.9m, +24.2% cc (+22% reported) — ahead of the top end of guidance.
• Gross margin 58.6% reported; ~60.9% "see-through" (ex one-off inventory charge), an underlying improvement of ~+460bps vs FY24. H1 had been 58.2% (61% ex-charge) with a £3.3m non-cash inventory charge (−315bps) plus mix (−195bps) and FX (−80bps) headwinds, partly offset by +525bps of margin-initiative and new-pricing-model gains.
• Adjusted EBITDA loss improved by £31.2m (26%) YoY. → Implied FY25 adj. EBITDA loss ≈ £(89)m and FY24 ≈ £(120)m. H2 adj. EBITDA loss £(38.4)m, £9.9m better than H1 — i.e. the loss is narrowing H/H.
• Statutory loss after tax £(145.2)m (FY24: £(146.2)m) — broadly flat, including £22.6m of restructuring to realign strategic focus. The gap between the improving adj-EBITDA loss and the flat statutory loss is heavy D&A, SBC, and the restructuring charge.
• Cash £302.8m, zero debt; net cash outflow ~£101m, ~£50m lower YoY. (FY24 exit cash was ~£338m with ~£159m TTM burn — so burn is decisively falling.)
• Guidance: FY2026 revenue +21–25% cc, targeted 62% gross margin; reiterated adj-EBITDA breakeven in FY2027 and cash breakeven by 2028.
• Market reaction: negative. Despite the revenue beat, shares fell ~12–17% on the print because the medium-term growth bar was cut (FY2027 trimmed from 25–30% to 21–25%) and a new CEO was announced simultaneously (Lens 8).

Unusual vs. its own history: the combination of a beat on the headline yet a sell-off tells you the stock now trades on the glide path to profitability and forward growth, not the current-year number. A revenue beat no longer rescues the multiple.

Lens 6 · Earnings Calls (sentiment trend)

transcripts/ empty — sentiment read from web call summaries.

Trajectory across the last few reporting cycles:

• JPM Jan 2025: upbeat, expansive — "a proteomics-genomics thing of beauty"; the narrative was platform breadth beyond DNA/RNA.
• H1 2025 (Sep 2025): confident — 25–28% cc growth, margin initiatives landing, guidance beaten.
• JPM / FY2025 results (Jan–Mar 2026): reflective and transitional. Co-founder Gordon Sanghera handed the baton to incoming CEO Francis Van Parys; tone was legacy-framing ("deep tech firm that took 10 years to deliver its first products," "this company will be around in 100 years' time") combined with a sober medium-term reset.

Recurring phrases: "path to profitability," "adjusted EBITDA breakeven in 2027," "applied markets," "see-through gross margin," platform extension into proteomics and eventually small-molecule/chemical sensing. What they stopped saying: the aggressive 25–30% medium-term growth framing — quietly retired and replaced with 21–25%. That single deletion is the most important sentiment signal: management itself recalibrated the growth ceiling.

Net: tone has cooled from "visionary deep-tech compounder" to "disciplined execution toward breakeven." Credibility-positive (under-promising on growth, over-delivering on margin/cash), but it removes the blue-sky multiple.

Lens 7 · Comps (long-read / sequencing-tools peers)

Peers pulled from _index.json (genomics topic) — the relevant comps are the sequencing-instrument names (Illumina, PacBio), not the gene-editing/mRNA tickers in the same topic bucket (CRSP/NTLA/BEAM/MRNA/BNTX — different business model, excluded).

Reads:

• ONT trades at ~4.2x EV/Sales — a growth multiple (justified by +24% cc growth and 60% gross margin), well above its loss-making long-read rival PacBio (~2–2.5x) but below profitable, scaled Illumina (~5.6x EV/Sales on low-single-digit growth).
• No P/E for ONT or PacBio — both pre-profit; the only honest valuation lens here is EV/Sales and the path to breakeven. The Illumina P/E above is an unverified implied figure — n/a — consensus P/E not sourced; do not rely on it.
• The cleanest framing: ONT is priced as the growth long-read pure-play (4.2x, +24%) vs. PacBio as the distressed long-read pure-play (~2x, shrinking revenue $0.20bn→$0.15bn 2023→24 ) — a genuine long-read duopoly where ONT is decisively the stronger franchise.

Lens 8 · Stock-Price Catalysts (>5% moves, ~5y since IPO)

ONT IPO'd Sept/Oct 2021 at 425p, spiked ~+45% to as high as 619p intraday on debut ("one of London's best-ever debuts"). It now trades ~115p (£1.12bn mkt cap, 17 Jun 2026) — down ~73% from IPO, in a 52-week range of 104.0p–224.8p, i.e. near the lows.

Pattern of what actually moves this stock:

• 2022–2023: COVID-revenue cliff + Emirati Genome renegotiation. Shares fell sharply as pandemic-surge sequencing rolled off and the Emirati Genome Program purchase-agreement change triggered a margin warning; the stock was down ~26% in Q1 2023 alone and ~69% from IPO by then.
• Jan 2024: weak preliminary 2023 revenue (LSRT ~£169m vs ~£177.4m consensus) → −15% in a day. Breakeven target pushed from FY2026 → FY2027.
• Recurring China / export-control drag — every cautious guide cites China and US semiconductor trade rules.
• Jan–Mar 2026: the medium-term guidance trim + CEO change → −12% to −17% on the day despite a revenue beat (FY2027 growth cut 25–30% → 21–25%; Van Parys in, Sanghera out).

What the tape reveals: this market reacts to (1) the breakeven/medium-term growth trajectory far more than to in-period beats, (2) flagship-program/contract news (Emirati, big national programs), and (3) the China/funding macro overlay (NIH funding uncertainty + China export controls). It is a credibility stock — every guidance reset has been punished hard, which is why management's new under-promise posture matters.

---

Phase C — Judge people & books

Lens 9 · Management

The defining event is a founder-to-professional-operator CEO transition (effective 2 March 2026).

• Outgoing: Dr Gordon Sanghera — co-founder (with chemist Hagan Bayley and IP Group), CEO since the company's 2005 inception. ~20 years tenure; the archetype scientist-founder who took a deep-tech idea to a commercial, ~£224m-revenue, 60%-gross-margin platform — but also presided over a −73% post-IPO de-rating and two breakeven pushes. Remains as advisor through early 2027; directly owns 1.6% (£21.8m) — meaningful skin in the game but not founder-controlling.
• Incoming: Francis Van Parys — professional operator pedigree, deliberately chosen to commercialize/scale: currently President & CEO of Radiometer (acute-care diagnostics, a Danaher company), prior senior roles at Cytiva and GE Healthcare; 20+ years running multi-billion-dollar life-science businesses. The Danaher Business System background is the tell: the board is buying operational discipline and a path to profit, not more vision. (Note: a search snippet conflated bioMérieux — no direct bioMérieux role is evidenced; his line is GE/Cytiva/Danaher-Radiometer.)

(1) Track record: Sanghera — built the category and the platform (real); commercial execution and capital-markets credibility (poor — serial guidance disappointments). Van Parys — scaled regulated life-science P&Ls inside the best operational culture in the industry (Danaher) (strong on paper; unproven at ONT). (2) Skin in the game: founder ~1.6%; large strategic holders historically (IP Group ~10.3% post-IPO, Oracle cornerstone $205m, Tencent ~7.9%, plus Temasek/Wellington/M&G/Nikon from the 2021 pre-IPO round) — note these are 2021 figures; current register not freshly sourced — n/a — 2025/26 share register not sourced. (3) Capital allocation: funded almost entirely by equity (IPO raised ~$478m / £3.4bn val; pre-IPO rounds) into R&D, the Harwell factory, and commercial build-out; zero debt, £302.8m cash. No buybacks/dividends (appropriate pre-profit). The £22.6m FY25 restructuring is the new regime starting to prune. (4) Red flags: the pattern of over-promising growth then resetting; heavy SBC (typical of the cohort) flattering adjusted metrics; otherwise no related-party or governance scandals surfaced. (5) Archetype shift: founder-visionary → Danaher-trained operator at exactly the stage (scale-to-profit) where that swap is usually the right call. This is the single most important positive change in the story — and also the biggest execution risk.

Lens 10 · Forensic Red Flags

Grounded in web disclosure (no filings on disk).

• Adjusted vs. statutory gap. FY25 adj. EBITDA loss ~£(89)m vs. statutory loss £(145.2)m — a ~£56m wedge from D&A, stock-based comp, and restructuring. Watch that SBC isn't quietly inflating the "see-through" margin and adjusted-EBITDA improvement narrative; the statutory loss has been flat for two years even as adjusted metrics improve.
• Inventory / margin quality. A £3.3m non-cash inventory charge hit H1-2025 gross margin (−315bps). Recurrent inventory write-downs in a consumables business can signal demand-forecasting or shelf-life issues — one instance isn't a trend, but it's the line to track.
• "See-through" gross margin is a management-defined, non-statutory construct (~60.9% vs. reported 58.6%). Treat the 62% FY26 target as guidance on the adjusted figure, not GAAP.
• Revenue lumpiness / recognition. Heavy reliance on large national-program contracts (Emirati, UK Biobank, Genomics England) makes period-to-period revenue lumpy and timing-sensitive — the EMEAI "projects conclude before new ones begin" air-pocket is a recognition-timing phenomenon, not pure demand loss.
• Cash-flow vs. earnings: healthily, cash burn is falling faster than the P&L loss (burn −~£50m YoY; outflow ~£101m) — a good divergence (the loss is more non-cash than the burn). Runway: £302.8m at ~£100m burn = ~3 years to cash breakeven (2028 target) without raising.

Regulatory findings (required sub-section). Per the pre-fetched regulatory/regulatory-findings.md (generated 2026-06-18, sources SEC EDGAR EFTS LR + AAER): Oxford Nanopore has no SEC CIK — it is not a US filer, so no SEC Litigation Releases or AAERs are searchable; total_sec_findings = 0.

• Item 3 (Legal Proceedings): n/a — no 10-K on disk (non-EDGAR filer). UK equivalent (Annual Report litigation note) not separately fetched.
• Non-SEC / web enforcement search ("Oxford Nanopore" (FTC OR DOJ OR FDA OR... settlement OR fine OR penalty)): no material government enforcement actions, fines, or consent decrees found. The material legal matters are offensive IP litigation initiated by ONT, not enforcement against it: ONT is suing BGI Group / MGI affiliates (England & Wales, plus US discovery subpoenas and Australian proceedings) for breach of contract, breach of confidence, trade-secret misappropriation and patent infringement around nanopore tech. MGI says nanopore products were <3% of its 2025 revenue and aren't sold in the UK; BGI calls the claims baseless. Separately, a 2018 PacBio↔ONT settlement had ONT refrain from "2D" products in UK/Germany through end-2023 (now expired).
• Conclusion: No material adverse regulatory or legal findings against the company — verified via the pre-fetched SEC EFTS check (n/a, no CIK) and web search as of 2026-06-18. The IP litigation is a strategic positive (defending the moat) but carries cost and an adverse-ruling tail risk.

---

Phase D — Project & stress-test

Lens 11 · Forward Projection (FY2026 → FY2028)

ONT is loss-making, so the scoreable metric is not EPS but the path to adjusted-EBITDA breakeven (FY2027 target) and cash breakeven (2028) plus revenue. Built bottom-up from FY2025 actuals (£223.9m rev; ~£(89)m adj-EBITDA loss; £302.8m cash) and FY2026 guidance (+21–25% cc, 62% GM target).

Revenue (£m), constant-currency growth applied:

Adjusted-EBITDA path:

• FY2026 ≈ £(55)–(65)m loss · FY2027 ≈ breakeven to slightly negative (the company's stated target) · FY2028 ≈ positive adj-EBITDA, approaching cash breakeven.

The forecast that matters (and the one I'd score): Does ONT hit adjusted-EBITDA breakeven in FY2027? Base-case probability ~55% — management has already pushed this once (FY26→FY27), the margin/cash trend is genuinely improving, but the EMEAI growth air-pocket and China/NIH-funding overhang make a second slip (to FY2028) a live ~40% risk. (Per --watchlist rules, the Brier forecast.ts create step is intentionally skipped — logging is only for genuinely committed base cases outside the sweep.)

No EPS line logged — pre-profit; EV/Sales + breakeven-timing is the honest valuation frame.

Lens 12 · Bull vs Bear

Bull case. Oxford Nanopore owns a category Illumina cannot enter with its current chemistry — long reads, real-time streaming, native epigenetics — inside a razor-and-blade model with a self-upgrading installed base and ~60% gross margins climbing to 62%+. Growth is durable and high-quality (clinical +59%, applied to ~34% of mix), the balance sheet is fortress-grade (£302.8m cash, zero debt), burn is falling decisively, and a Danaher-trained operator is taking the wheel precisely to convert the platform into profit by 2027. The optional upside — proteomics, then small-molecule/chemical sensing on the same platform — is essentially free in the stock at ~4x EV/Sales near all-time lows. If breakeven lands on time and proteomics shows signal, the multiple re-rates from "broken IPO" to "scaled, profitable, multi-omics platform."

Bear case. Three things could permanently impair or de-rate the equity: (1) the growth ceiling is structurally lower than the IPO thesis — management itself cut medium-term growth from 25–30% to 21–25%, and a sequencing-tools TAM constrained by NIH/academic-funding cycles and China export controls may not support the re-rate bulls need; (2) breakeven slips a second time — it's already been pushed once, and a single soft year (EMEAI air-pocket + China) blows the 2027 credibility that is the thesis, triggering another leg down in a stock that punishes every reset; (3) the long-read niche stays a niche — for the large, routine, high-accuracy clinical market that drives volume, short-read (Illumina) remains the gold standard and PacBio HiFi contests the high-accuracy long-read sub-niche, capping ONT's addressable share.

Pre-mortem (18 months out, thesis broke): It's late 2027. ONT missed adjusted-EBITDA breakeven — the EMEAI strategic-project gap that was supposed to be temporary persisted, China revenue kept eroding under export controls, NIH funding stayed soft, and the new CEO's first full year delivered ~16–18% growth instead of 22%. The "deep-tech optionality" (proteomics) produced press releases but no revenue. The market re-rated it toward PacBio's ~2x EV/Sales. The stock is sub-90p.

Are multiples too high? At ~4.2x EV/Sales for +22% cc growth and 60% GM, it's not demanding for the quality — but it's not cheap for a company that has missed before and is two years from cash breakeven. The multiple is fair-to-slightly-generous; the de-rate risk is timing-of-breakeven, not absolute valuation.

Contrarian view (what the market refuses to see): Consensus treats ONT as a serially-disappointing UK growth-stock husk (down 73%, near lows, low conviction). The thing the tape is ignoring: this is the first time the company has a real operator and an under-promised number — the setup that historically precedes the end of a de-rating. The market is so trained to fade ONT that it may miss the inflection where falling burn + a Danaher operator + a beaten-down ~4x multiple turns a "broken IPO" into a quiet compounder. The risk/reward has quietly improved even as sentiment has worsened.

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

Dismantling the bull case:

• What structurally breaks the money-machine? The razor-and-blade only compounds if the installed base keeps expanding and running. If instrument placements stall (funding-constrained academia, China shut out, clinical adoption slower than hoped), consumable pull-through plateaus and the model's flywheel stops — you're left with a sub-scale tools company burning ~£100m/yr.
• Revenue concentration: large national-genomics programs (Emirati, UK Biobank, Genomics England, the NHS biosurveillance build) are a meaningful chunk. The Emirati renegotiation already proved one contract can break a half-year. A pause/loss/renegotiation of any flagship is a guidance miss — and this management has shown it can't absorb misses without a share-price puncture.
• Why the moat is weaker than bulls think: short-read still wins on routine SNV accuracy and dominates clinical volume; PacBio HiFi contests high-accuracy long-read; and Chinese nanopore entrants (Qitan/QNome, plus BGI/MGI's contested tech) threaten the one region (APAC) ONT needs and the price umbrella globally. ONT is litigating MGI precisely because it fears the cloning.
• Most dangerous underestimated competitor: not Illumina — it's MGI/BGI and Chinese nanopore players. They can undercut on price, they're protected at home, and if ONT loses or only partly wins its IP suits, the long-read "monopoly" framing collapses in the fastest-growing region.
• Worst capital-allocation / incentive signals: serial breakeven pushes while paying out heavy SBC; an "adjusted/see-through" metric vocabulary that flatters a statutory loss that hasn't improved in two years.
• Assumptions that must hold for ~115p: (i) +21–25% cc growth actually sustains through the EMEAI air-pocket; (ii) breakeven lands in 2027 this time; (iii) China/NIH don't deteriorate further; (iv) IP suits don't go badly. If growth disappoints by 20–30% (i.e. ~15% instead of 22%), breakeven slips to 2028+, the cash runway tightens toward a possible dilutive raise, and the stock re-rates toward PacBio's ~2x EV/Sales → a 25–40% downside scenario.
• Single scenario that permanently impairs: a structural China lock-out + a domestic Chinese nanopore champion taking the global price umbrella down, combined with NIH/academic funding staying soft — turning ONT into a perpetually-subscale, perpetually-cash-burning niche tools vendor that must raise equity from a depressed price. Plausibility: moderate — not the base case, but not tail-thin given the macro.

Lens 14 · Management Questions (ordered by information value)

1. Breakeven credibility: You've moved adj-EBITDA breakeven from 2026 to 2027 once. What specifically is different now, and what is the single biggest risk to a second slip — and at what revenue-growth rate does 2027 breakeven fail?
2. EMEAI air-pocket: The 2026 deceleration is framed as strategic projects concluding before new ones begin. Quantify it — how much of EMEAI revenue is "between projects," what's the visibility on the next cohort, and is this timing or structural?
3. China: With export controls and domestic nanopore entrants (Qitan, MGI), what is your realistic 3-year China revenue assumption — managed decline, plateau, or write-off — and what's baked into guidance?
4. The ASIC supply chain: Who fabricates the sensor ASIC, what's the geographic/geopolitical exposure, and what's the second-source plan if that foundry is disrupted?
5. Proteomics — money or marketing? What's the realistic timeline to material proteomics revenue, what does early customer traction actually look like, and how much R&D is it consuming vs. core sequencing?
6. MGI/BGI litigation: What does winning get you commercially, what does losing cost you (in China share and globally), and what's the expected legal spend run-rate?
7. Pricing model: The new pricing model added ~+525bps to margin in H1-2025. Is that a one-time step or a continuing lever, and is there volume-elasticity risk as you raise effective consumable prices?
8. Dilution: At ~£100m burn and £302.8m cash you have ~3 years. Under what scenario would you raise equity, and is there a hard floor below which you would not issue at a depressed price?
9. Van Parys's mandate: What are the 3 operational changes (Danaher-system style) you're making in the first 12 months, and what KPI should investors hold you to?
10. Clinical durability: Clinical grew +59%. How much is repeatable annuity (diagnostic menu, reimbursement-backed) vs. one-off program ramps, and what's the reimbursement path?
11. Installed base economics: What's the active installed base, the consumable pull-through per active device, and the trend in revenue per installed instrument?
12. Inventory charge: What drove the H1-2025 £3.3m inventory write-down — forecasting, shelf-life, or quality — and is it a one-off?
13. SBC: Statutory loss has been ~flat at £(145)m for two years while adjusted metrics improve. Walk us through the SBC trajectory and when statutory and adjusted profitability converge.
14. Capital allocation post-breakeven: Once cash-flow positive (2028), what's the priority — reinvest in proteomics, M&A, or returns?
15. The 100-year framing: Sanghera said ONT will exist in 100 years. For the next 3, what one number would you stake your credibility on?

---