{
  "@context": "https://orbyd.app/schemas/dossier.v1.json",
  "ticker": "TSLA",
  "name": "Tesla, Inc.",
  "url": "https://orbyd.app/dossiers/TSLA/",
  "json_url": "https://orbyd.app/dossiers/TSLA.json",
  "status": "WATCHLIST",
  "current_conviction": "LOW",
  "graded_conviction": null,
  "archetype": {
    "code": "a1",
    "n": 1
  },
  "current_thesis": "Terafab (2026-08-06) put a number on the split: $16.8B first phase, and Musk's approximate guess allocates ~25% of output to Tesla Optimus versus ~75% to SpaceX — the listed entity co-funds compute it mostly does not receive. Energy is the leg still compounding (Q2 storage 13.5 GWh, +41% YoY; Megapack 3 line live 08-06), while Germany's 367 July registrations (-66.9%) put Q3 deliveries back in play. Saturated and unresolved.",
  "invalidation_trigger": "A weekly close below $297 forfeits the July and 52-week low at $297.38, which has no 2026 shelf beneath it; secondary confirmation is the ACEA July aggregate (~2026-08-20) printing an EU-wide Tesla decline consistent with Germany's -66.9%, or August ending with no disclosed Optimus unit count from the Fremont line.",
  "catalyst_date": "2026-08-14",
  "outcome": "OPEN",
  "outcome_date": null,
  "invalidation_fired": null,
  "themes": [
    "megacap-ai-platforms",
    "ev-autonomous-mobility"
  ],
  "tags": [
    "watchlist"
  ],
  "sources": [],
  "notes": [
    "2026-04-19: AI/robotics/FSD narrative — Optimus + Robotaxi binary catalysts",
    "2026-06-05: JPMorgan perma-bear analyst flipped to Neutral, PT $475; TD Cowen reiterated Buy $490 (6/4). Two upgrades in 48h AFTER the 6/2 gap-down. Treat as sentiment-regime datapoint, not a buy trigger on its own.",
    "2026-06-04: Robotaxi removed safety monitors + widened Austin map BUT fleet shrank to 20 cars. The scale tell (fleet count / weekly rides) matters more than the driverless milestone — track whether 20 expands or contracts.",
    "2026-06-05: Roadster demo slipped to August (cold-gas thruster work w/ SpaceX) — third robotics catalyst to slip/disappoint; now outside 30d window.",
    "Optimus has never had an official production unit count disclosed; every unit deployed to date is internal, and none is confirmed shipped to an external customer.",
    "Four NHTSA investigations into FSD/Autopilot remain open. EA26002 is an Engineering Analysis, the last stage before a case closes or a recall is requested.",
    "Single-country European registration prints swing hard on shipment timing and read properly only against the ACEA aggregate published later in the month.",
    "Tesla's China wholesale figure includes exports. Domestic retail deliveries move separately and have declined for five consecutive quarters.",
    "Post-Q2 sell-side price targets ranged from $130 (Wells Fargo) to $600 (Wedbush); any single target is one view rather than a consensus level.",
    "Leveraged inverse TSLA and SpaceX ETFs reset daily, so multi-day returns diverge from the inverse of the underlying and are not a clean sentiment gauge."
  ],
  "body_markdown": "\n_Levels and figures below are stated as of the 2026-08-06 close; nothing is claimed about sessions after that date._\n\n## TSLA — Tesla, Inc.\n\n## Current Thesis\n\nThe 2026-07-22 print broke the AI-and-robotics re-rate on margin — adjusted EPS $0.33 against $0.51 expected, operating income down 57% YoY to roughly $400M, free cash flow **-$1.09B** versus **+$1.44B** in Q1 2026 — and the question since has been where the AI optionality actually sits. On 2026-08-06 part of it got a number. Tesla and SpaceX confirmed **Terafab**, a jointly developed chip complex in Grimes County, Texas: **$16.8B first phase**, more than 100 million square feet of planned manufacturing space housing logic, memory, packaging and test under one roof, at least 3,000 jobs across Grimes and Brazos counties, sited on the Gibbons Creek Reservoir with a commitment to reservoir water instead of local groundwater. Musk described the output split as a \"very approximate guess\": **~25% for Tesla Optimus and ~75% for AI spacecraft.**\n\nThat ratio is the frame an investor is now underwriting. The listed vehicle co-funds a fab whose stated majority output serves the private one, a quarter after printing negative free cash flow. It is the same structural argument the Wall Street Journal raised on 2026-07-31 about a China divestiture clearing a SpaceX merger, restated as committed capex with a disclosed allocation.\n\nThe leg inside TSLA that is measurably compounding is energy. Q2 storage deployments were **13.5 GWh against 9.6 GWh** a year earlier, roughly +41%. On 2026-08-06 Megapack 3 began rolling off the line at the Brookshire, Texas Megafactory, 16 months from groundbreaking, designed for 50 GWh/year, each unit about 5 MWh versus 3.9 MWh on the Megapack 2 XL. Ørsted's Old 300 project in Needville, Texas — 250 MW / 500 MWh of Tesla Megapacks — went operational the same day.\n\nAgainst that, Europe's July prints split hard, and the German number is the one that changed the picture since the last update. Tesla registered **367 vehicles in Germany in July, -66.9% YoY**, a 0.1% share of Europe's largest car market, ending five consecutive months of triple-digit YoY gains (June was 7,768). BYD registered 5,240, +365.4%. France was +86% and Denmark +52%; Norway -97%, Sweden -60%, Portugal -68.7%. One month, six registries, no single direction.\n\n**Life-cycle: SATURATED.** Dating it on 2026-08-06: a $16.8B first-phase capex confirmation and the start of Megapack 3 production landed in the same session, and the stock closed **-0.628% at $319.53**. Alongside that, the purer listed expression of the Musk-AI trade has moved: SPCX notional short interest of roughly $23.6B overtook TSLA's roughly $22B (S3 Partners, reported 2026-08-03), Direxion launched the SpaceX Bear 2X ETF on 2026-08-06 completing a bull/bear pair across both names, and SpaceX's post-IPO lock-up began expiring the same day, freeing 911.5 million shares, about 12% of total. What holds the label off DEAD is that robotaxi throughput and storage deployments are both compounding against disclosed baselines, and the corporate-structure question is genuinely open.\n\n## Bull Case\n\n- **Storage is growing faster than cars and is now capacity-backed (2026-08-06):** Q2 deployments 13.5 GWh, +41% YoY; Megapack 3 production started at Brookshire, Texas, 50 GWh/year design capacity, 5 MWh per unit versus 3.9 MWh on the prior generation; Ørsted's 250 MW / 500 MWh Old 300 project live in Needville, Texas.\n- **Terafab is a real, sited, costed commitment (2026-08-06):** $16.8B first phase in Grimes County, Texas, >100M sq ft planned, ≥3,000 jobs, vertically integrated logic and memory for edge inference in Optimus and Cybercab.\n- **China wholesale hit its 2026 high (CPCA, 2026-08-04):** July 93,579 units, +37.85% YoY, the highest July in Tesla's history and the strongest month since Shanghai's 97,171 in December 2025 — a ninth consecutive month of YoY gains. YTD 561,528, +29.88%.\n- **Autonomy has a disclosed baseline (2026-07-22):** ~700,000 paid robotaxi miles since launch, 500+ vehicles across Austin and the Bay Area, fleet doubling roughly monthly. A Nevada AVNC application filed 2026-06-05 covers Clark County including Harry Reid and Henderson Executive airports, for up to 5,000 vehicles in year one.\n- **Related-party demand shows up in a filing:** SpaceX purchased **$295M** of Tesla Megapack batteries in Q2, disclosed in SpaceX's 2026-08-04 quarterly materials.\n- **Volume was never the broken part (2026-07-22):** Q2 deliveries 480,126, +25% YoY against roughly 406,600 expected, on record revenue of $28.24B, +26% YoY.\n- **US policy leans toward domestic humanoids:** the FCC's import restrictions on new Chinese humanoid and quadruped robots took effect 2026-07-28; the FCC chair confirmed on 2026-08-05 that existing owners are unaffected and framed the rules as building a domestic supply chain.\n\n## Bear Case\n\n- **Germany collapsed in a growing market:** 367 units in July, -66.9% YoY, 0.1% share, versus 9,252 in March, 5,111 in May and 7,768 in June. BYD outsold Tesla more than 14 to one at 5,240 units.\n- **The European BEV market was expanding at the same time:** UK July new car registrations rose 11.7% to 156,571 with BEVs +44.5% to a record July share of 27.5% (SMMT). Portugal's light BEV registrations rose 54.3% YoY in the month Tesla fell 68.7% there.\n- **China volume is export-carried:** Q2 China deliveries -2.05% YoY, a fifth consecutive quarterly retreat; China fell to 26.28% of Tesla's global volume, under 30% for the first time since Q4 2020; Q2 Shanghai exports of 128,394 exceeded domestic deliveries of 126,157 for the first time in any quarter, with exports now 49% of Giga Shanghai output against 28% a year earlier.\n- **Terafab's allocation runs against the listed entity:** ~75% of the guessed compute output points at SpaceX, funded from a balance sheet that produced -$1.09B free cash flow in Q2.\n- **Unit economics have not turned:** Q2 automotive gross margin 16.3%, average revenue per vehicle ~$42,730 versus $45,345 a year earlier, regulatory credit revenue $146M versus $439M, operating expenses +47% to $4.35B.\n- **Washington is hostile to the vehicle business:** at a Las Vegas rally on 2026-08-05, Trump said EV drivers \"have a disease\" and said of the mandate, \"I ended the electric mandate. I don't want to talk about it because Elon wasn't exactly thrilled with me, but I had to do it.\" The $7,500 federal credit is already gone.\n- **A long-standing bull moved the robotaxi crown elsewhere:** Gary Black said on 2026-08-06 that Uber, not Tesla or Waymo, has the best shot at bringing robotaxis to the masses.\n\n- **Optimus remains unmeasured:** no official production unit count has ever been disclosed, and the Fremont line converted from Model S/X in 46 days was guided only to a late-July/August start.\n\n## Setup & Price Structure\n\n- July closed **-26.01%**, from a $432.86 high to a **$297.38** low that is also the 52-week low, with no 2026 shelf beneath it.\n- Closes since: **$321.27 on 2026-08-05**, **$319.53 on 2026-08-06** (-0.628%, session range $315.52–$322.99). That leaves the stock roughly **7.4% above the July low** and still without a base — three weeks of repair against a 26% month.\n- As of the 2026-07-31 close, the 50-day sat near $384 and the 200-day in the $400s, both falling, with the 50-day beneath the 200-day. Neither has been tested since; the reclaim so far is off the low, not through resistance.\n- The two largest announcements of the week — $16.8B of committed fab capex and the start of a new 50 GWh/year product line — produced a red close on 2026-08-06. That is the crowding observable worth weighing: incremental headline supply is not buying incremental price.\n- Positioning around the complex: TSLA notional short interest roughly $22B (S3 Partners, reported 2026-08-03), now below SPCX's roughly $23.6B, with 206 million SPCX shares short, 32.2% of that float. Direxion's SpaceX Bear 2X ETF listed 2026-08-06, the same day SpaceX's lock-up began releasing 911.5 million shares.\n- No Tesla earnings date falls inside the next 30 days. Third-party calendars place the Q3 print in late October; Tesla has not confirmed it.\n\n## Catalyst Calendar (next 30 days)\n\n- **2026-08-14** — Q2 2026 13F filing deadline. Institutional positioning as of 06-30, before the 07-22 print, establishing the pre-break baseline that Q3 filings get measured against.\n- **~2026-08-20 (est.)** — ACEA July EU new-car registrations. The aggregate that decides whether July's European split reads like Germany's -66.9% or France's +86%, and therefore whether Q3 deliveries carry a European hole.\n- **~2026-08-26 (est.)** — NVIDIA quarterly print. Sets the AI-capex regime the megacap complex trades inside; TSLA's daily direction has tracked that tape more closely than its own news flow since 07-22.\n- **~2026-08-31 (est.)** — Optimus production cell at Fremont, guided to a late-July/August start. A disclosed unit count would be the first hard Optimus production number ever released.\n- **~2026-09-01 (est.)** — August European national registrations begin printing (KBA, Nordics, France). Shows whether Germany's 367 was allocation or demand.\n- **~2026-09-02 (est.)** — CPCA China August wholesale volumes. Whether July's 93,579 holds, and whether the export share keeps rising while domestic keeps falling.\n\n## Elapsed catalysts\n\n- **2026-08-07 onward (no deadline)** — NHTSA Engineering Analysis EA26002 on FSD visibility. An Engineering Analysis is the last step before a case closes or a recall is requested; escalation reprices the autonomy leg on its own. *(passed 2d ago)*\n\n## What Would Change Our Mind\n\nEurope breaks it first. If the ACEA July aggregate around 2026-08-20 shows an EU-wide Tesla decline consistent with Germany's 367 units rather than France's +86%, July stops being a shipment-calendar artifact and becomes a Q3 delivery problem — and the German number would then be a five-month growth streak ending in a market where BEV registrations were still rising.\n\nTerafab breaks it second. The $16.8B first phase was committed one quarter after -$1.09B free cash flow. A Q3 disclosure showing a Tesla funding share without a matching Tesla-allocated output commitment would confirm that the listed entity is paying for compute the private entity mostly consumes; a disclosed offtake or wafer allocation to Tesla would argue the other way.\n\nPrice carries the third condition. $297.38 is the July and 52-week low and there is nothing beneath it on the 2026 chart: **a weekly close below $297** forfeits that reference and takes the read from saturated to structurally broken.\n\nWhat would argue the other direction, concretely: a disclosed Optimus unit count out of Fremont before month-end; a Q3 robotaxi disclosure roughly doubling the ~700,000 paid miles and 500+ vehicles reported 2026-07-22; storage deployments above 13.5 GWh with Brookshire contributing; and a reclaim of the falling 50-day, which sat near $384 on 2026-07-31 and has not been approached since.\n\n## Correlation Notes\n\n- **TSLA and SPCX are now financially entangled, not merely thematically.** SpaceX bought $295M of Tesla Megapacks in Q2, Terafab is jointly funded, and Musk's guessed compute split is ~25% Tesla Optimus / ~75% AI spacecraft. SpaceX's first public print (2026-08-04) beat on revenue at $7.81B, +92% YoY from $4.1B, with EPS of -$0.09 against -$0.26 expected, net loss narrowed to $541M, adjusted EBITDA $3.5B, and capex of $18.37B roughly 39% above estimates — the stock rose 9.4% in the session and gave back about 7% after hours on the capex line.\n- **The AI-capex tape is the dominant driver.** SpaceX said on 2026-08-05 it will use Nvidia GPUs exclusively, and JPMorgan sees it taking a significant share of Nvidia's 2027 shipments. NVDA's late-August print sets the regime for the whole complex.\n- **BYD is the direct European counterparty.** Germany July: BYD 5,240, +365.4%; Tesla 367, -66.9%. Same market, same month, opposite signs.\n- **Policy correlation is two-sided.** FCC restrictions on imported Chinese humanoids (effective 2026-07-28) support the Optimus leg, while the removal of the $7,500 federal EV credit and Trump's 2026-08-05 remarks weigh on the vehicle leg. The two do not net out inside one reported segment.\n- **Leveraged inverse products on TSLA and SPCX reset daily**, so their multi-day returns diverge from the inverse of the underlying and do not read as a clean sentiment gauge.",
  "first_seen": "2026-04-19",
  "last_analyzed": "2026-08-07T06:10:36+00:00",
  "last_synthesized": "2026-08-07",
  "last_update_source": "theme_discovery",
  "license": "Content © orbyd. Cite the canonical URL."
}