{
  "@context": "https://frontierpicks.com/schemas/dossier.v1.json",
  "ticker": "INDI",
  "name": "indie Semiconductor, Inc.",
  "url": "https://frontierpicks.com/dossiers/INDI/",
  "json_url": "https://frontierpicks.com/dossiers/INDI.json",
  "status": "DORMANT",
  "current_conviction": "LOW",
  "graded_conviction": "LOW",
  "archetype": {
    "code": "a2",
    "n": 2
  },
  "current_thesis": "Automotive ADAS-silicon design-win story keeps improving on paper ($7.4B backlog, $25M radar order, ams OSRAM sensor buy) while the tape stays broken below the ~$4.00 50-day MA; TD Cowen's 2026-07-13 lift to a $4.50 Hold is a thaw, not a trigger. No clean setup until Q2 earnings 2026-08-06 — value-trap zone.",
  "invalidation_trigger": "A weekly close below $3.00 confirms the failed breakout is extending toward the $2.32 52-week low and keeps it a no-touch falling knife; the constructive case only re-arms on a reclaim of ~$4.00 then a weekly close above $4.81 on volume with auto-semi peers (MBLY/NXPI) confirming.",
  "catalyst_date": null,
  "outcome": "INVALIDATED",
  "outcome_date": "2026-06-10",
  "invalidation_fired": true,
  "themes": [
    "semiconductors-analog",
    "ai-chips-memory",
    "physical-ai-robotics"
  ],
  "tags": [],
  "sources": [],
  "notes": [
    "Headline revenue consolidates Wuxi indie Micro (~$30M of the $70M Q3 midpoint); core indie is the ~$40M line that grades the design-win story.",
    "Class V stock and TeraXion options sit inside the 227.6M non-GAAP share count; the dilution vectors are structural, not one-off events.",
    "GAAP and non-GAAP diverge sharply (Q2 2026: $35.1M vs $8.9M operating loss) on amortization and stock comp — both lines need reading.",
    "€35M of cash is committed at the ams OSRAM CMOS close against $138.975M of cash and equivalents at 2026-06-30.",
    "High-beta micro-cap: single option-flow prints have produced double-digit intraday moves that subsequently mean-revert.",
    "Primary peer group is automotive semis (MBLY/NXPI/ON); the humanoid-supplier association dates only to 2026-08-19 press coverage."
  ],
  "body_markdown": "## Current Thesis\nTwo things happened since the 2026-08-06 Q2 print. First, a shelf formed: the stock reclaimed the ~$4.00 area it lost on 2026-06-16, closed $4.04 on 2026-08-07, held $3.98 on 2026-08-19 and $4.00 on 2026-08-20, and finished at $4.15 on 2026-08-21. Second, a new narrative leg attached itself to the name. Unitree Robotics listed on Shanghai's STAR Market on 2026-08-19 — RMB 6.1B (~$904M) raised, oversubscribed more than 8,000 times, shares up 460% at the close from the 150.8-yuan offer and as much as 629% intraday to 1,100 yuan — and indie is a named silicon supplier into that platform, flagged on its own 2026-08-06 call alongside Agibot. Benzinga's 2026-08-19 mover piece described indie as supplying Unitree's radar chipsets, vision processors and mixed-signal SoCs, and noted Unitree shipped more than 5,000 robots in 2025 using iND88X-family processors.\n\nThe leg an investor would be buying is therefore two-part: an automotive ADAS design-win pipeline ($7.4B strategic backlog disclosed 2026-05-07) converting into a 30%-handle revenue growth rate at a still-loss-making supplier, plus a free option on humanoid silicon content. What the tape said about part two on the day: INDI fell 8.29% to $3.98 into the Unitree debut session. The next day, Unitree founder Wang Xingxing said humanoid robots remain roughly a decade from their \"ChatGPT moment.\" One session of sympathy trading that went the wrong way supports no conclusion about correlation; it does establish that the humanoid bid has not yet been paid for here.\n\n**The narrative is maturing.** The automotive leg is well known and still working operationally — Q2 revenue $64.009M against a $62.542M consensus, Q3 guided $67–73M — but flow has moderated and the marks are compressing toward the tape rather than away from it. MarketBeat's panel showed a $4.75 consensus target across 4 analysts (1 Sell / 2 Hold / 1 Buy) as of 2026-08-21, versus the high-$5s-to-$6s aggregator averages that stood at the last refresh. The physical-AI attachment dated 2026-08-19 is new attention, but it is one headline cycle old and produced a down session, which is not enough to call the name accelerating.\n\n## Bull Case\n- **Q2 2026 beat on both lines (2026-08-06):** revenue $64.009M vs $62.542M consensus, +24% YoY and above the guidance midpoint; non-GAAP EPS -$0.05, in line.\n- **Q3 guide implies acceleration (2026-08-06):** $67–73M against a $70.401M estimate, roughly +30% YoY at the midpoint, split ~$40M core indie / ~$30M Wuxi indie Micro; non-GAAP opex ~$37M, non-GAAP LPS -$0.04.\n- **Losses narrowing (2026-08-06):** GAAP operating loss $35.1M vs $43.0M a year earlier; non-GAAP operating loss $8.9M vs $14.5M.\n- **The humanoid customer now has a public balance sheet (2026-08-19):** Unitree raised RMB 6.1B (~$904M) on the STAR Market and closed its debut up 460%, with the raise earmarked for manufacturing scale-up. A supplier's end-customer going from private to funded-and-listed is a change in the order-visibility question, whatever the share price did that day.\n- **Backlog and hard orders (2026-05-07):** $7.4B strategic backlog disclosed at the Q1 print, plus a $25M 77GHz radar production order through one Tier-1 spanning an EU and an Asian OEM.\n- **Sensing stack pending (Master Agreement 2026-05-08, announced 2026-05-11):** subsidiary Ay Dee Kay Ltd. To acquire ams OSRAM's CMOS image-sensor business via ams Sensor Belgium B.V. for €40M — €35M cash at close plus a €5M vendor note at 2.5% simple interest payable 24 months after closing.\n- **Liquidity at the print:** cash and equivalents $138.975M at 2026-06-30 ($149M including restricted cash).\n\n## Bear Case\n- **The sympathy bid failed on the day it was offered:** -8.29% to $3.98 on 2026-08-19, the session of the Unitree debut, per Benzinga's own mover coverage.\n- **The customer capped the timeline the next morning (2026-08-20):** Wang Xingxing put humanoids about a decade from their \"ChatGPT moment,\" which argues against near-term volume conversion of the physical-AI wins named on the 2026-08-06 call.\n- **Marks compressing:** Roth Capital cut to $5 from $7 on 2026-08-07, the session after the beat; TD Cowen has carried a Hold at $4.50 since 2026-07-13; Weiss Ratings reiterated a Sell (D-) on 2026-07-17. MarketBeat's 4-analyst consensus stood at $4.75 on 2026-08-21. Other aggregators still publish averages above $5 — that dispersion means no headline consensus figure is a usable anchor for this name.\n- **Leverage rose through the first half:** long-term debt $403.7M at 2026-06-30 versus $339.8M at 2025-12-31; current debt obligations $12.3M.\n- **GAAP economics unchanged in character:** $35.1M operating loss on $64.0M of revenue in Q2; GAAP loss per share -$0.17 against non-GAAP -$0.05.\n- **Consolidation carries the headline:** ~$30M of the $70M Q3 midpoint is Wuxi indie Micro. The organic line is the ~$40M core.\n- **Deal still open:** the Q2 Form 10-Q for the period ended 2026-06-30 records the ams OSRAM acquisition as not yet closed, subject to regulatory approvals, expected in the third quarter of 2026.\n\n## Setup & Price Structure\n- Reference close $4.15 on 2026-08-21; 30.8% below the $6.00 52-week high; a three-month price change of -13.7%; RSI(14) at 61.4.\n- The $4.00 area is doing structural work. It was lost 2026-06-16, reclaimed at $4.04 on 2026-08-07, tested at $3.98 on 2026-08-19 and $4.00 on 2026-08-20, and left behind at $4.15 on 2026-08-21. Four post-print closes sit inside a band of roughly 4% around it. That cluster is the shelf; below it the chart is open toward the $2.32 52-week low.\n- Upside confirmation is unchanged from the prior frame: a weekly close above $4.81 on expanding volume, with automotive-semi peers participating.\n- **Crowding and positioning observables, stated as observables:** retail-facing mover coverage attached a $4 small-cap to a headline mega-IPO within a single session (Benzinga, 2026-08-19); RSI(14) 61.4 sits in the upper half of its range while price remains nearly a third below the 52-week high; there is no earnings date inside 30 days to force a repricing; and the dilution vectors are structural rather than event-driven — Class V stock and TeraXion options sit inside the 227.6M non-GAAP share count, with €35M of cash contractually committed at the ams OSRAM close against $138.975M of cash and equivalents at 2026-06-30. No new insider transactions surfaced in the filings reviewed for this refresh; absence of a filing is weak evidence and is recorded as such.\n\n## Catalyst Calendar (next 30 days)\n- **2026-08-22 to 2026-09-21:** no confirmed company-scheduled event. The window is empty of dated binaries.\n- **~2026-09-30 (est.):** completion of the ams OSRAM CMOS image-sensor acquisition. Guided to close in Q3 2026 and recorded as still pending in the Q2 Form 10-Q. No closing date has been announced, so it could land inside the 30-day window without notice.\n- **~2026-11-05 (est.):** Q3 2026 results. Tests the $67–73M guide and, specifically, whether core indie clears the ~$40M it was guided to versus the ~$30M Wuxi indie Micro consolidation.\n\n## What Would Change Our Mind\nThe load-bearing structure is the four-session shelf around $4.00 built between 2026-08-07 and 2026-08-21. If that gives way — a weekly close below $3.75 — the post-print repair failed and the $3.00 area and the $2.32 52-week low come back into the frame; the 2026-08-19 down session on the Unitree headline would then read as distribution into attention rather than a shakeout.\n\nThree non-price conditions would also change the read. First, the ams OSRAM close: if 2026-09-30 passes with the transaction still listed as pending and no restated timeline, the sensing-stack leg becomes a question about regulatory approval rather than integration. Second, the core/Wuxi split at the Q3 report: core indie printing below the guided ~$40M, or the company ceasing to disclose the split, would mean the growth headline is a consolidation artifact. Third, any ATM, S-3 takedown or convertible announcement, or a non-GAAP share count above 227.6M at Q3, would confirm that the €35M cash outlay and the GAAP burn are being funded by the equity.\n\nOn the other side, a weekly close above $4.81 on volume, with a quantified humanoid or radar order disclosed rather than described, would move this from maturing toward a genuinely accelerating leg.\n\n## Correlation Notes\n- Primary beta is automotive semis — MBLY, NXPI, ON — not the AI-datacenter memory complex. A Tier-1 inventory correction shows up here before it shows up in the backlog number.\n- The humanoid/physical-AI correlation is asserted by press coverage (Benzinga, 2026-08-19) and by the company's own customer list (2026-08-06 call), but has one session of price evidence and it was negative. Treat it as unproven until a second, independent Unitree-driven session moves the stock in the expected direction.\n- China exposure is direct: ~$30M of the $70M Q3 midpoint is Wuxi indie Micro, plus iND880 vision-processor wins at Chinese OEMs. Export-control headlines transmit to this revenue base.\n- The ams OSRAM consideration is euro-denominated (€35M cash, €5M note), so the dollar cost of the close moves with EUR/USD between signing and completion.",
  "first_seen": "2026-05-08",
  "last_analyzed": "2026-08-22T11:02:44+00:00",
  "last_synthesized": "2026-08-22",
  "last_update_source": "watchlist_research",
  "license": "Content © FrontierPicks. Cite the canonical URL."
}