Seeing Machines Limited (AIM: SEE) — Investment Research Note

Executive summary

Seeing Machines is an Australia-headquartered, AIM-listed computer-vision company that designs AI-powered Driver Monitoring Systems (DMS) licensed to automotive OEMs (via Tier-1s) and sold direct into commercial fleets under the Guardian brand. Over the five-year period the group has moved from ~1m installed vehicles and heavy NRE-dependence to an 8.2m installed base with automotive royalties compounding at 100%+, culminating in a first-ever profitable H2 (H2 FY26 adjusted EBITDA of US$10.7–11.7m on US$52.9m revenue) as the July-2026 EU GSR mandate came into force 2026-08-11 trading update. The single most important valuation point today is that a fragile balance sheet (US$4.3m cash, US$54m convertible current, refi under exclusive negotiation) is being resolved just as high-margin royalty revenue inflects — the equity re-rates only if that refinancing lands cleanly.

Fair value estimate

• Fair-value range: 4.5p – 7.5p per share (£216m – £360m market cap); midpoint ~6.0p / £288m
• Methodology: forward EV/EBITDA supported by a check on royalty run-rate economics. Q4 FY26 automotive production of 2.11m units 2026-08-11 annualises to ~8.4m units. Applying an implied royalty of ~US$4/unit at scale (derived from FY26 automotive royalty US$33.9m on 4.5m units) plus Guardian ARR US$15m and residual NRE/hardware yields a plausible FY27 revenue envelope of US$110–140m. With incremental EBITDA margins of ~80% demonstrated in H2 FY26 (a US$29m revenue swing produced a US$24m EBITDA swing) 2026-08-11, structural EBITDA capacity at that revenue is US$20–35m. Central case FY27 Adj EBITDA ~US$22m × ~18× forward multiple = ~US$400m EV, less ~US$50m net debt post-refi ≈ US$350m equity ≈ £275m ≈ 5.7p.
• vs. current £222.7m mcap / 4.64p: absolute upside to midpoint ~29% (range −3% to +62%).
• View: undervalued, but conditionally — the equity is cheap only if the convertible is refinanced on non-dilutive terms and H2 FY26 EBITDA run-rate holds.

Sector context

• ICB Technology classification is fair, though functionally SEE is auto-safety hardware/software with an OEM royalty model. Compared with peers, SEE has above-average growth momentum, above-average operating leverage, comparable moat (patents + 50%+ share of current DMS production volumes per 2026-03-27 H1 report), but below-average balance-sheet quality and inferior earnings quality (heavy R&D capitalisation, minimum-royalty-guarantee accounting).
• Listed comparables: Smart Eye AB (STO: SEYE) — nearest DMS pure-play; Cipia Vision (TASE: CPIA); indirect: Mobileye (MBLY), Ambarella (AMBA), Valeo (which is also a partner).

Investment thesis

1. Regulation-driven royalty inflection is now visible in the numbers, not just guidance. Automotive production volumes rose 195% in FY26 to 4.5m units, automotive royalty +135% to US$33.9m, and Q4 alone hit 2.1m units — a 4× step-up from Q4 FY25. The EU GSR mandate came into force 7 July 2026 2026-08-11. This is a hard regulatory tailwind, not an option.
2. Operating leverage is textbook — and now demonstrated. H2 FY26 adjusted revenue US$52.9m produced Adj EBITDA of ~US$11m; H1 FY26 revenue US$23.4m produced −US$13.7m. Implied incremental EBITDA margin ~80%. Cost base was structurally reset in FY25 (~US$12m annualised OpEx removed, headcount from 455 → 353) 2025-03-27 H1 report, so further revenue growth flows disproportionately to profit.
3. Expanding option value across Cabin AI, Future Mobility and Aviation. 3D Cabin Perception Mapping launched at CES 2026, impairment detection targeting the US regulatory roadmap, US$5.6m robotaxi order, Collins Aerospace collaboration and MELMB Japan channel — each is optionality that isn't in numbers yet 2026-08-11, 2026-03-27.

Key risks

1. Convertible Note refinancing (US$47.5m face, US$54m carrying) matures 4 October 2026. Cash of US$4.3m at 30 June 2026 means the company depends on completing refi "in an exclusive negotiation period" 2026-08-11. Failure or highly dilutive terms would be materially adverse.
2. Working-capital funding of royalty growth. Trade receivables and royalties owed jumped to US$25.3m from US$11.6m in six months as auto customers pay quarterly in arrears 2026-08-11. Cash conversion lag is a real risk in a low-cash business; the A$11m receivables facility 2026-03-27 H1 report is small relative to receivables.
3. Earnings quality / adjusted accounting. Heavy R&D capitalisation (US$2.4m capitalised H1 FY26, US$8.7m H1 FY25), minimum-royalty-guarantee accounting that pulls forward revenue at start of production, and a large gap between statutory and adjusted revenue mean reported figures require careful interpretation 2026-03-27, 2025-08-21. Statutory losses remain large (H1 FY26 net loss US$22.5m).

Operating leverage

This is arguably the strongest single feature of the equity story. Adjusted operating expenses fell to US$27.7m in H1 FY26 from US$32.4m 2026-03-27, and management has explicitly held OpEx flat while revenue inflected. Gross margin is 58% and rising with royalty mix. The H2 FY26 revenue mix shift is the clearest quantification available: a US$29.5m half-on-half revenue increase produced an ~US$24m EBITDA swing — an implied ~80% incremental EBITDA margin. Automotive royalties in particular are near-100% gross margin (embedded software licence per vehicle). If FY27 revenue reaches US$110–140m as production of the 8.2m installed base scales into royalty flow, incremental revenue of US$35–60m over FY26 could generate US$25–45m of additional EBITDA — i.e. a 10–20% revenue beat vs consensus expectations plausibly multiplies EBITDA rather than adds to it. This is the exact "long-tail" profile the investor's operating-leverage pillar seeks.

Value-trap signals

• Persistent accumulated losses (US$264.5m at 31 Dec 2025) with no dividend history and repeated equity raises (2021 £30m; 2024 MELMB £26.2m at 4.09p).
• Repeated pushouts of cash-flow break-even guidance: FY25 → "during 2025" → "by end CY 2025" → now targeting FY27 2025-03-27, 2025-02-26, 2025-08-21.
• Balance-sheet fragility: current liabilities exceed current assets due to the convert now being classified current; cash of US$4.3m is materially below convert face.
• Otherwise no evidence of terminal-industry decline — regulation is a genuine tailwind, so the "cheapness" is refi/execution-driven rather than structural.

Earnings vs. expectations

• FY22 revenue A$54.2m — in line with A$55.6m consensus; cash A$59.3m vs A$46.1m (beat) 2022-08-15.
• FY23 revenue US$57.8m vs US$53.9m consensus — beat 2023-08-22.
• FY24 revenue US$67.6m vs US$65.1m consensus — beat; but Cash EBITDA loss guided down in June 2024 warning 2024-08-28.
• FY25 revenue US$62–63m vs US$58m consensus — beat on revenue; but EBITDA break-even target missed and rolled into FY26 2025-08-21.
• FY26 revenue US$76.3m vs US$79.7m consensus — slight revenue miss but Adj EBITDA loss US$2–3m vs consensus US$3.9m — beat on profitability 2026-08-11.
• Pattern: consistent revenue beats vs consensus over five years, offset by repeated timing slips on profitability guidance. Reliable top-line delivery; less reliable on when operating leverage kicks in.

Conviction — 3 (moderate)

Anchors: (a) H2 FY26 provides a first hard data-point on incremental EBITDA margins; (b) EU GSR is a legislated demand driver, not a speculative TAM; (c) installed-base growth (8.2m vehicles) creates a compounding royalty annuity that's simple to model directionally.

Limits: (a) balance-sheet fragility and pending convert refinancing make the equity binary in the near term — outcome materially changes the fair-value distribution; (b) heavy use of adjusted metrics, minimum-royalty-guarantee accounting and R&D capitalisation make GAAP earnings hard to anchor; (c) royalty per unit and take-rate assumptions in FY27+ are estimates, not disclosures.

Driver scoring (0-100)

• ai_beneficiary: 50 — applied AI (computer-vision DMS), demonstrable revenue growth, but classic "picks-and-shovels" this is not.
• operating_leverage: 85 — H2 FY26 demonstrated ~80% incremental EBITDA margin at scale.
• earnings_surprise_trend: 60 — consistent revenue beats offset by repeated profitability-timing misses.
• cyclicality: 55 — auto-tied but with a hard regulatory demand floor via GSR.
• moat: 55 — 50%+ share of current DMS production, 18+ OEM programmes, patents; but Smart Eye, Cipia and OEM-internal solutions compete.
• leverage: 65 — convertible ~US$54m current, US$4.3m cash; would drop to 35–40 post successful refi.
• earnings_quality: 45 — heavy R&D capitalisation, multiple non-IFRS metrics, minimum-guarantee timing effects.
• management_quality: 55 — restructuring delivered, MELMB anchor investor secured, but repeated slippage on cash-flow targets.
• growth_momentum: 80 — FY26 revenue +45%, auto royalty +135%, production volumes +195%.

Overall score: 555 / 1000

Partial-to-strong fit for the investor profile. Very high operating leverage and a reasonable-to-cheap central-case valuation with a real regulatory tailwind are the positives; the AI angle is "applied AI" not "AI receiver in the picks-and-shovels sense", and downside protection is weak until the October 2026 convertible refi completes. Not a top-band buy, but a genuine candidate for a smaller sizing.