MicroVision

Phase A — Understand the business

Lens 1 · Company Overview

MicroVision designs lidar sensors plus perception software for three end-markets: automotive ADAS/AD, industrial autonomy (AGVs/AMRs, warehouse, mining/agri), and — newly emphasized — security & defense. The historical IP is MEMS laser-beam-scanning (the same scanning-mirror tech that once shipped in a Microsoft HoloLens-class HUD and in U.S. military heads-up displays); the modern product suite spans short/mid/long-range, time-of-flight and FMCW, multiple wavelengths (905/940/1550nm), and both flash and MEMS architectures. Named product lines: MOVIA (short-range flash, the current revenue workhorse, incl. "MOVIA L" for defense/industrial), MAVIN (flagship long-range MEMS for automotive OEMs — the big-prize line), IRIS and HALO (1550nm ToF solid-state long-range, acquired from Luminar), plus the integrated perception SoC.

The business model is the classic merchant-lidar pitch: win an automotive OEM/Tier-1 design slot, then earn high-volume supply revenue (and royalties) at start-of-production years later, while selling industrial units near-term to bridge the gap. The reality is that almost no revenue exists yet. FY2025 total revenue was $1.21M, down from $4.70M (FY2024) and $7.26M (FY2023) — a two-year −83% collapse. Customer concentration is extreme and shifting: FY2025's top four customers were 42% / 19% / 15% / 12% of revenue — i.e., the largest customer (an agricultural-equipment maker) was $0.5M. The 10-K explicitly states revenue "has been negatively affected by the loss of certain of these customers." Q1-2026 revenue was $935K (top-10 auto OEM 54%, a construction/mining equipment maker 22%, an EU defense-autonomy robotics firm 17%). So: a 30-year-old public company (IPO 1996) with a $957M accumulated deficit and roughly one mid-sized SaaS customer's worth of annual revenue.

Lens 2 · Supply Chain

Upstream inputs → MicroVision → end customer, named where the filings/commercial layer name them:

• Inputs: semiconductor foundries produce the custom MEMS scanners and ASICs "to order"; laser diodes (905/940/1550nm), optics/opto-mechanics, detectors. The 1550nm fiber-laser path (inherited from Luminar/IRIS-HALO) implies dependence on specialty fiber-laser and InGaAs detector suppliers — a structurally costlier BOM than 905nm silicon detectors. Specific upstream vendors are not disclosed — n/a — not named in filing.
• MicroVision: fabless. "Our products were manufactured by a contract manufacturer based on our proprietary design" — it outsources assembly to contract manufacturers and is "evaluating long-term Tier-1 relationships". R&D split U.S. (Redmond) + Germany (the Ibeo/Hamburg and now Scantinel/Ulm legacy), ~135 engineers at YE2025.
• Downstream: automotive OEMs and Tier-1 suppliers (unnamed; "an automotive OEM customer" co-developed the perception software); industrial — automated-warehouse operators, materials-handling OEMs, mining/agri (John-Deere-class ag-equipment maker is the historical top customer, unnamed); defense — an EU autonomy-systems firm, plus a new U.S. defense push via a 2025-formed Defense Advisory Board.

Chokepoints / single-source dependencies: (1) custom MEMS+ASIC fabrication at a foundry is a single-source physical dependency — a lidar without its scanner ships nothing; (2) the entire revenue base is single-digit customers, so "supply chain" risk here is really demand concentration — the loss of one ag-equipment customer halved revenue. The robotics commercial-layer files referenced by the briefing (kb/robotics/wiki/supply-chain.md etc.) are topic-level, not company-specific, so the chain above is filing-derived. Names present, but the named end customers are anonymized in the filings — a genuine limitation, not laziness.

Lens 3 · Competitive Advantages (moats)

Claimed moat: a broad, multi-architecture IP portfolio (MEMS scanning + flash + FMCW + ToF across wavelengths), automotive-grade qualification on IRIS/HALO/MOVIA, and an integrated perception-on-sensor SoC that "lowers cost and simplifies the system architecture" by avoiding an external ECU. After the Luminar/Scantinel deals, MicroVision can credibly say it owns one of the widest lidar technology stacks of any Western pure-play.

My assessment — the moat is weak-to-illusory at the unit-economics level. Three reasons:

1. No volume = no scale moat. Lidar moats come from design wins that lock in years of SoP volume and from cost-down at scale. MicroVision has no disclosed automotive production design win for MAVIN; production launches OEMs describe are 2028–2029. Its largest revenue cohort is sub-$1M industrial. Hesai ships hundreds of thousands of units into Chinese AVs; MicroVision ships samples.
2. Negative gross margin is the anti-moat. FY2025 cost of revenue was $18.5M on $1.21M revenue → a −$17.3M gross loss (incl. $9.9M inventory write-downs). A product you sell below cash cost has negative bargaining power — you need the customer far more than they need you.
3. Bargaining power is inverted. Against giant OEMs (who second-source everything) and against a balance sheet screaming distress, MicroVision is a price-taker. Its leverage over the foundry (tiny order volumes) is also weak.

The one defensible asset is the patent estate + automotive-grade pedigree + acquired Ibeo/Luminar/Scantinel IP and talent — real option value if the lidar market consolidates to 2–3 Western survivors and MicroVision is one. But "valuable IP at a going-concern" is an acquisition thesis, not a moat that compounds equity value. Durable moat verdict: low.

Lens 4 · Segments

MicroVision does not report disaggregated product or geographic segments in the income statement — it runs as a single reportable segment, and segments.csv on the shelf is empty (header-only). The only granularity disclosed is the customer-concentration footnote, which doubles as a de-facto end-market mix:

Trend and cause: revenue is decelerating hard — the ag-equipment customer that was 60% of a $4.7M base in FY2024 ($2.8M) shrank to 42% of a $1.2M base in FY2025 ($0.5M). The decline is loss-of-customer, not pricing. The forward story is a mix pivot: away from a single ag customer toward (a) industrial/defense MOVIA L units and (b) the acquired IRIS/HALO automotive-grade long-range line. Geography: U.S. + Germany R&D; revenue is global but unquantified by region. Hard requirement met by exception: no segment table exists to source, so this lens is built from the only disaggregation the filings provide and labeled accordingly.

---

Phase B — Measure performance

Lens 5 · Earnings Result (latest print — Q1 2026, reported 2026-05-13)

The numbers, straight from the 10-Q:

Drivers: the revenue "beat" and gross-margin flip are real but trivial in dollars; the story Q1 actually tells is opex re-inflating as MicroVision absorbs two acquisitions, with R&D nearly doubling. Cash rose to $46.1M (from $32.4M at YE) — but only because they drew $42.5M from selling investment securities and took $20.7M of new note proceeds, while paying $33.2M for the Luminar assets. Guidance/tone: management raised the framing — reiterated FY2026 revenue of $10–15M, raised gross-margin guide to 35–40%, and lowered cash-burn guidance to ~$60M (from $65–70M) on synergy cuts. Balance-sheet flags: $9.9M inventory write-down in FY2025 and a written-to-zero acquired-backlog intangible (the Ibeo order backlog impaired) tell you the prior pipeline did not convert. Market reaction: despite the "beat," the stock fell from ~$0.61 (Jun 1) to $0.29 (Jun 26), a >50% June drawdown — the market is pricing the financing overhang and reverse-split risk, not the revenue line. Unusual vs. own history: the company is now a distressed-asset acquirer, a strategic departure from the prior decade of organic design-win waiting.

Lens 6 · Earnings Calls (sentiment trend)

No transcripts on the research-layer shelf (transcripts/ empty), so this is web-derived and labeled. The arc across the last several calls:

• FY2024 → mid-2025 (Sharma era): tone defensive — "targeting automotive partnerships amid lidar challenges," repeated emphasis on RFQ activity that kept slipping ("quotes could close before end of 2025… may shift to 2026"). The recurring, never-landed phrase was the imminent automotive design win. SoP dates pushed to 2028+.
• Late-2025 → 2026 (DeVos era): a deliberate rhetorical reset — "boldly executing to transform the lidar industry," "Lidar 2.0," pivot language from a single automotive bet to "multi-vertical, non-correlated revenue streams" (industrial + defense + auto). The new framing leans on near-term industrial/defense revenue to fund the automotive marathon.
• What they stopped saying: specific near-term automotive production award timing. What they started saying: "defense," "Lidar 2.0," "reseller/partner program," "synergies."

Sentiment shift: from defensive-waiting to acquisitive-pivoting. Honest read — the pivot is rational (industrial/defense has a real near-term TAM and shorter sales cycles), but the language is doing heavy lifting against a balance sheet that constrains every option. A reset narrative on a melting share price is the oldest tell in micro-cap.

Lens 7 · Comps

Western lidar pure-plays. Multiples are with source/date or marked not-sourced — **no multiple is fabricated.** MVIS market cap is; its EV/Sales is an `` off research-layer revenue.

Read: the whole cohort is unprofitable and richly valued on hope, but MicroVision is the worst-positioned name on revenue scale — Ouster does ~$143M revenue (8x+ Aeva, and ~120x MVIS) at a grounded ~7x sales with <$5M quarterly burn and ~$170M cash; Innoviz is integrated into Mobileye Drive with L3 SoP 2026/27; Hesai has real automotive volume. MVIS trades at a far higher implied EV/Sales than Ouster on a fraction of the revenue and a worse balance sheet. On any revenue-anchored multiple, MVIS is the most expensive lidar name for the least revenue and the most financing risk. The only honest defense of the valuation is option-value-on-the-IP, not multiples.

Lens 8 · Stock-Price Catalysts (what moves the tape)

MVIS is a retail/meme-driven, financing-news-driven stock — its >5% moves cluster on dilution events, squeeze dynamics, and binary "deal" headlines, far more than on fundamentals:

• Short-squeeze mechanics: short interest ~22% of float (~67M shares), ~18 days-to-cover. The 2021 meme era saw 1000%+ moves; the setup has repeatedly drawn squeeze speculation. But in 2026 the squeeze failed to fire: $0.84 (Feb 25) → $0.36 (Jun 11) → $0.29 (Jun 26).
• Dilution / financing prints move it down: the $43M convertible-note raise and the reverse-split proxy were treated as overhangs, not catalysts.
• M&A headlines move it up short-term: the Luminar-bankruptcy "winning bid" (Jan 2026) and Scantinel deal generated pops that faded.
• Earnings: Q4-2025 (Mar 4) and Q1-2026 (May 13) were billed as "binary catalysts"; the stock ultimately bled regardless of the revenue "beats."

What the market actually reacts to for this name: (1) the share count / financing trajectory, (2) squeeze positioning, (3) any credible automotive design-win or large defense contract (none yet at scale). It is not an earnings-fundamentals stock — there are no earnings. This is the single most important behavioral fact for sizing any position.

---

Phase C — Judge people & books

Lens 9 · Management

Leadership just changed at the top. Sumit Sharma (CEO Feb 2020 → Sep 30 2025; with the company since 2015) was replaced by Glen DeVos, who joined as CTO in April 2025 and became CEO effective Sep 30 2025. The FY2025 10-K thus reports a year run mostly by Sharma but signed into a DeVos regime.

1. Track record: Sharma's tenure delivered the Ibeo acquisition (2023) and a string of un-landed automotive RFQs, while revenue fell from $7.3M (FY23) to $1.2M (FY25) and shares roughly doubled — a value-destructive scorecard, however much the lidar market itself disappointed. DeVos brings 30+ years at Aptiv/Delphi (ex-CTO; ran the Advanced Safety & User Experience unit) — genuinely relevant automotive-supplier credibility and OEM relationships, the most legitimate positive in this file. The CFO seat is also unsettled: Stephen Hrynewich was named Interim CFO in Dec 2025 — a non-permanent finance chief during a cash crisis is a yellow flag.
2. Tenure & skin in the game: insider ownership is immaterial relative to the ~345M share count; insider-transactions.csv is empty (not sourced on the shelf). Sharma base comp was ~$530K + 100%-of-salary bonus + 1.125M RSUs (2024 agreement) — RSU-heavy comp that dilutes alongside the equity it rewards.
3. Capital allocation: the defining acts are issuing equity and convertible debt to fund losses (FY2025 financing inflow $60.9M, incl. $77.4M gross stock/warrant issuance) and buying distressed competitor assets cheap (Luminar $33.2M out of Chapter 11; Scantinel as a bargain-purchase gain). The roll-up is opportunistic and arguably smart if the survivor thesis plays out — but it is being financed off a sub-$1 stock, which converts strategic logic into shareholder dilution. ROE/ROIC: persistently and deeply negative.
4. Red flags: serial dilution; a toxic/structured convertible (Lens 10/13); a pending 1-for-5 to 1-for-15 reverse split (vote Jul 10 2026); interim CFO; reliance on an ATM at distressed prices.
5. Archetype: professional managers, not founders, now led by an automotive-industry operator. For this stage (survival + consolidation), DeVos's profile is the right kind of leader — but he inherited a balance sheet that limits what good management can do.

Lens 10 · Forensic Red Flags

Accounting / disclosure risks (all figures ``):

• Negative gross margin + recurring inventory write-downs. FY2025 cost of revenue ($18.5M) was ~15x revenue ($1.21M), driven partly by $9.9M of inventory write-downs (FY2024: $2.0M). Building/buying inventory the market won't absorb at cost — watch for further write-downs as acquired Luminar inventory is revalued.
• Impairments across the board. FY2025 booked $10.1M intangible impairment + $1.2M ROU-asset + $2.2M PP&E impairment; the acquired-backlog intangible was written to zero — the Ibeo "order backlog" did not convert to revenue. This is the clearest forensic signal that prior "pipeline" disclosures were optimistic.
• Earnings quality / non-cash noise. Net loss is heavily distorted by derivative- and warrant-liability mark-to-market swings (FY2025: +$5.7M derivative gain, +$4.4M warrant gain; −$4.7M debt-extinguishment loss; $18.5M interest expense incl. $11.2M debt-discount amortization + $7.3M non-cash interest). The reported net loss (−$95.0M) and the cash burn (−$58.7M operating) diverge materially — always read the cash-flow statement here, not the P&L. The convertible's mark-to-market can flatter or hammer GAAP net income quarter to quarter independent of operations.
• SBC fell sharply ($16.1M FY23 → $11.5M FY24 → $0.7M FY25) — unusual; likely forfeitures/repricing amid the stock collapse. It means non-GAAP "improvement" isn't coming from SBC add-backs being hidden — but it also signals retention stress.
• Going-concern / liquidity covenant. The 10-K and 10-Q carry the standard "sufficient cash for at least the next 12 months" language but conditioned on expected financing activities, and the convertible imposes a minimum-cash covenant of 110% of the outstanding note (~$19.5M at YE2025). The auditor/management framing stops short of a formal going-concern qualification, but the risk factors repeatedly flag "ability to continue as a going concern" and a Nasdaq sub-$1 deficiency (notice Jan 12 2026).

Regulatory findings (required sub-section). Read regulatory/regulatory-findings.md:

• SEC Litigation Releases: none naming MicroVision since 2021 (EDGAR EFTS LR search — total 0).
• AAERs: none (total 0).
• Non-SEC enforcement: web search ("MicroVision" (FTC OR DOJ OR FDA OR CFPB OR consent decree OR settlement OR fine OR penalty) enforcement) surfaced no material regulatory action.
• 10-K Item 3 (Legal Proceedings): the company's own disclosure — "We are not currently party to any legal proceedings that management believes are reasonably possible to have a material adverse effect".
• Conclusion: No material regulatory or legal findings — verified via SEC EDGAR EFTS (LR, AAER), web search, and 10-K Item 3 as of 2026-06-30. (The risks here are financial/dilutive and listing-related, not enforcement.)

---

Phase D — Project & stress-test

Lens 11 · Forward Projection (FY2026–FY2028 — there are no EPS, so model revenue, burn, and runway)

For a company at ~$1M revenue with a negative-margin product, an EPS line is meaningless and would be fabricated precision — so the scoreable variable is revenue vs. guidance and runway. Bottom-up from the latest actuals + guidance +:

Inputs (labeled):

• FY2025 actual revenue $1.21M; Q1-2026 $0.94M. Management FY2026 guide $10–15M, H2-weighted, GM 35–40%, cash burn ~$60M.
• FY2026 burn ~$60M against cash $46.1M + $42M ATM headroom. Runway < 12 months without tapping the ATM → the ATM (sub-$1) or the unissued $30M convertible tranche / new $43M note must be drawn → further dilution is near-certain.

Scenario paths (revenue, `` with arithmetic):

• Bear (most weight): acquired-asset revenue ramps slower than guided; FY2026 ≈ $5–7M, FY2027 ≈ $10–15M, FY2028 ≈ $20M. Burn stays ~$55–60M; equity count balloons past 400M pre-split. Stock is a financing residual.
• Base: management roughly hits the low end — FY2026 ≈ $10M, FY2027 ≈ $20–25M, FY2028 ≈ $35–45M as industrial/defense scales; still deeply FCF-negative through FY2028. Requires ≥1 more capital raise.
• Bull: a real automotive design win (MAVIN/IRIS) lands and a defense program scales — FY2026 ≈ $15M (top of guide), FY2027 ≈ $40M+, FY2028 ≈ $80M+, approaching gross-margin breakeven at the product line. Even here, corporate cash breakeven is beyond the window and the share count is far larger.

No forecast.ts create in the --watchlist loop (per skill rule — only log a Brier forecast on genuine conviction; this is a breadth pass). For the record, the scoreable base claim would be: "MVIS FY2026 revenue < $12M (misses guide midpoint), resolves 2027-03-31" — directionally my lean, but not logged here.

Lens 12 · Bull vs Bear

Bull case. MicroVision now owns one of the widest Western lidar IP/product stacks (MEMS + flash + FMCW + ToF, 905→1550nm) after buying the Ibeo, Luminar (IRIS/HALO), and Scantinel assets — much of it for cents on the R&D dollar, out of bankruptcy. A new CEO with real Aptiv automotive pedigree is executing a coherent "fund the automotive marathon with near-term industrial/defense revenue" pivot. Industrial autonomy and defense lidar are genuine, shorter-cycle, higher-margin TAMs with tailwinds (warehouse automation, autonomous mining/hauling, counter-UAS/ground-robotics). If the Western lidar field consolidates to 2–3 survivors and MicroVision is one — or gets acquired for its IP/automotive-grade qualification — equity could re-rate violently off a ~$96M cap. Optionality + ~22% short interest = explosive upside if a catalyst lands.

Bear case (2–3 permanent-impairment risks).

1. Financing death-spiral. ~$60M annual burn, <12-month cash, a structured convertible with low conversion prices, an active sub-$1 ATM, and a pending 1-for-5/1-for-15 reverse split designed to preserve listing — the equity is being continuously diluted to fund losses, and a reverse split resets the dilution runway rather than fixing it. This alone can impair shareholders to near-zero even if the technology succeeds.
2. Revenue may never reach scale before the cash does. Two years of −83% revenue, automotive SoP not until 2028–2029, no disclosed production design win, and a TAM the whole cohort keeps over-promising. The acquired backlog was literally written to zero.
3. Negative gross margin structurally. Selling sensors below cost with recurring inventory write-downs means revenue growth consumes cash rather than generating it until volumes are an order of magnitude higher.

Pre-mortem (18 months out, thesis broke): the squeeze never came; the reverse split executed and the post-split stock kept sliding as the ATM and converts issued into weakness; FY2026 revenue landed mid-single-digit millions (missed the $10–15M guide); a dilutive raise at a distressed post-split price gutted holders; the IP optionality got realized only via an acqui-hire that paid debt and crumbs to equity.

Are multiples too high? Yes — on any revenue anchor (≈80x trailing / ~50x forward-guide EV/Sales vs. Ouster's ~7x on 120x the revenue), MVIS is the most expensive, least-revenue, worst-balance-sheet lidar name. The valuation is pure option premium.

Contrarian view (what the market refuses to see): the genuinely contrarian take cuts bearish, not bullish — the consensus retail narrative treats MVIS as a coiled short-squeeze + "Lidar 2.0 transformation" story, but the dominant, under-priced reality is that the equity is a financing instrument: every operational "win" is funded by selling more of the company at sub-$1, so good news for the lidar business can still be bad news for the per-share owner. The asymmetric, less-crowded insight is that a reverse split here is a countdown, not a catalyst.

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

Dismantling the bull case. Where revenue is concentrated: in a handful of sub-$1M customers and an aspirational automotive pipeline — if the one ag/construction customer or the one EU defense customer slips, revenue halves again, as it already has. Why the moat is weaker than bulls think: "widest IP stack" assembled from three companies that failed or distressed (Ibeo bankruptcy, Luminar Chapter 11, Scantinel bargain-purchase) — a portfolio of other people's failures is not obviously a winning hand; if the tech were a durable moat, these assets wouldn't have been for sale at fire-sale prices. Most dangerous competitor bulls underestimate: Hesai (Chinese scale + cost) and Ouster (real $143M revenue, $170M cash, near-breakeven) — both can out-spend and out-ship MicroVision into oblivion while it rations cash; on the OEM side, Innoviz (Mobileye Drive, L3 SoP) is already in a production platform MVIS is still quoting for. Worst capital-allocation / structure: the toxic convertible — zero coupon, conversion at $0.7462 / $1.596 (well below where the stock once traded), monthly forced partial repayments at a 10% premium, a 110%-of-balance minimum-cash covenant, and a holder who can dump converted stock into the market; plus a new $43M note at $0.8819 (61.3M shares) and an open ATM. Assumptions that must hold for today's price: continued capital-markets access at sub-$1, listing maintained through the reverse split, and the industrial/defense ramp actually printing in H2-2026. If growth disappoints 20–30%: essentially irrelevant to a DCF (there's no positive cash flow) but highly relevant to financing — a guide miss closes the equity window and forces worse terms, a reflexive down-spiral. Single permanent-impairment scenario (and plausibility): a financing-driven recapitalization or delisting that wipes most equity value — plausible, not tail-risk, given <12-month cash and the existing structured debt.

Lens 14 · Management Questions (ordered by information value)

1. With ~$60M FY2026 burn and <12 months of cash before the ATM, exactly how many more shares do you expect to issue (ATM + the unissued $30M convertible tranche + the $43M note) over the next 12 months, and at what assumed price?
2. What is the purpose of the 1-for-5 to 1-for-15 reverse split beyond Nasdaq compliance — and what prevents the post-split stock from re-breaching $1 within a year?
3. Name a firm automotive production design win (OEM, program, SoP year, lifetime volume) — or state plainly that none exists yet.
4. What are the specific terms and triggers of the senior secured convertible(s), and under what scenario does the holder's conversion/forced-repayment most dilute or destabilize the equity?
5. What is the realistic timeline to corporate cash-flow breakeven, and what revenue level does it require?
6. What is the unit gross margin on MOVIA L and IRIS/HALO at current and at 10x volume — when does product-level gross margin turn durably positive ex-write-downs?
7. Of the FY2026 $10–15M revenue guide, how much is contracted/PO-backed today vs. pipeline, and how much is H2 hockey-stick?
8. What did you actually acquire from Luminar and Scantinel — defensible IP, qualified products, or mainly talent and inventory — and how do you avoid impairing it like the Ibeo backlog?
9. How do you compete with Ouster's revenue scale and Hesai's cost while rationing cash — what is your right to win?
10. What is the defense revenue opportunity concretely (programs, primes, timeline), and is any of it contracted?
11. Why should investors believe the automotive RFQ pipeline now, after multiple prior cycles of slipped timing?
12. What is the permanent CFO plan, and why an interim CFO during a liquidity crisis?
13. How much of reported results is derivative/warrant mark-to-market, and how should investors track the cash trajectory you're managing to?
14. What listing contingency exists if Nasdaq compliance fails post-split (Capital Market, NYSE American, OTC) and what does that do to the convertible covenants?
15. Under what conditions would the board consider selling the company / the IP rather than continuing to fund losses through dilution?

---