Robotics
A profitable EV maker priced as a solved-autonomy robotics company — the car business is shrinking, the GAAP profit prop (reg credits) is going to zero, and the entire ~190x multiple now rents on robotaxi + Optimus execution that is real but years behind the price.
Research
The verdict
A profitable EV maker priced as a solved-autonomy robotics company — the car business is shrinking, the GAAP profit prop (reg credits) is going to zero, and the entire ~190x multiple now rents on robotaxi + Optimus execution that is real but years behind the price.
Tesla makes money today in three reported segments, but management is mid-pivot from "EV maker" to "real-world AI & robotics company" (Musk's own framing on the Q4 2025 call).
FY2025 consolidated: total revenue $94.83B (−$2.86B / ~−2.9% YoY — Tesla's first-ever annual revenue decline), GAAP net income to common $3.79B (−$3.30B YoY), GAAP operating income $4.4B.
Contract structure / payment terms. Almost entirely transactional (vehicle sale at delivery; storage at commissioning). Recurring revenue is still thin: FSD (~1.3M paid subscribers globally as of Q1 2026 ), Supercharging, insurance, and a nascent robotaxi per-ride model. No take-or-pay backlog cushioning the auto business — demand is spot and price-elastic, which is why ASP cuts flow straight to margin.
Customers = retail vehicle buyers (no single-customer concentration on autos) + utilities/IPPs for Megapack (more concentrated, project-lumpy). Suppliers — battery cells from Panasonic, LG Energy Solution, CATL, plus in-house 4680. Competitors — BYD and the Chinese OEM bloc, legacy OEMs, and in autonomy Waymo (ahead).
Map: raw materials → cells → packs/powertrain → vehicle/Megapack → end customer, with named stakeholders.
Names present → lens passes. The structural point: Tesla is the most vertically integrated battery supply chain in the West, which is a genuine cost/control moat for both autos and energy — but it is capital-hungry (>$20B FY2026 capex guide, see Lens 5).
Real, durable moats:
Contested / weakening moats:
Bargaining power. Over suppliers: high (scale + in-house alternative cell). Over customers: falling — ASP cuts and rising inventory in Q1 2026 say Tesla is now a price-taker at the margin in autos, the opposite of the 2021–22 pricing-power era.
No segments.csv data on the shelf — all ``, sourced to SEC 8-Ks / earnings coverage.
| Segment | Q1 2026 rev | YoY | FY2025 rev | FY2025 YoY | Margin signal |
|---|---|---|---|---|---|
| Automotive | $16.2B | +16% | declined (vol + ASP) | down | Auto GM ex-credits 19.2% Q1'26 (up seq from 17.9%), aided ~$230M warranty true-down + tariff relief; FY2025 9-mo auto GM 16.9% vs 19.0% PY |
| Energy gen & storage | $2.41B | −12% (lumpy QoQ) | $12.77B | +27% | ~29.8% FY GM, Q4'25 GM 28.7%, record gross profit $1.1B; 46.7 GWh (+49%) |
| Services & other | $3.7B | growth | — | — | Margin 9.2% Q1'26 (up from 8.8% seq) |
Geography — China is the #2 market; BYD overtook even without selling in the US, and China-market share pressure is the single biggest geographic risk. (No `` geographic split available — n/a at the country-revenue level.)
Trend read. Auto = decelerating/declining volume, margin recovering off a low on cost-out not pricing. Energy = the growth + margin engine, but quarter-lumpy. The mix is shifting toward energy and (prospectively) services/robotaxi — exactly the pivot management is selling.
Source: 8-K exhibit 99.1 via CNBC/TIKR (SEC direct fetch returned 403 — figures cross-checked across CNBC + TIKR + StockTitan).
n/a at GAAP-EPS granularity.n/a); FY2025 ran 46.7 GWh.The catch beneath the beat. The EPS beat is real but low-quality: it leans on a $230M warranty true-down, a $250M one-time energy tariff benefit, and cost-out — not volume or pricing. Volume fell, FCF went negative, and the margin "recovery" is off a depressed base. The headline ("revenue +16%, EPS +52%") flatters a quarter where the core unit economics and cash generation deteriorated. Market reaction: TSLA is down ~20% YTD in 2026 and slid on the delivery shortfall.
Tone has shifted decisively from "we sell cars" to "we are an AI/robotics company" across the last several calls:
What they keep saying: autonomy "this year," Optimus as the trillion-dollar product, energy as the steady compounder. What they've stopped saying: firm volume-growth targets for the car business and concrete affordable-model demand detail. Sentiment read: management is deliberately redirecting investor attention away from the car P&L toward optionality — a tell that the near-term auto numbers are not the story they want owned. Repeated timeline slippage on FSD/robotaxi scale is the credibility tax.
| Company | Ticker | Mkt cap | Fwd P/E | EV/Sales | Notes |
|---|---|---|---|---|---|
| Tesla | TSLA | ~$1.52T | ~190–200x | n/a | Priced as robotics + autonomy, not autos |
| BYD | 1211.HK | n/a | n/a | n/a | #1 BEV maker (2.26M BEV 2025); trades at a fraction of TSLA's multiple |
| Waymo (Alphabet) | GOOGL | n/a (subsidiary) | n/a | n/a | Autonomy leader (L4 live multi-city); valued inside GOOGL |
| Toyota | TM | n/a | n/a | n/a | Volume leader, ~single-digit P/E typical for autos |
| Figure AI | private | n/a — private | n/a | n/a | Humanoid pure-play; Figure 03 on BMW lines |
Read. On any auto-industry yardstick (Toyota/GM/Ford trade high-single to low-double-digit P/E; BYD modestly higher for growth), TSLA at ~190x forward earnings is 2026 EPS consensus only ~$2.08–$2.25 against a ~$405 price. The multiple is not an auto multiple — it is a call option on autonomy + Optimus. Anyone underwriting TSLA on car earnings is structurally short the stock; anyone long is long the option. There is no defensible "fair multiple" without a probability-weighted robotaxi/Optimus TAM — which is exactly where the disagreement lives.
Annual prints frame the regime:
What the market actually reacts to for this name: (1) the Musk/narrative regime — Twitter, politics, and macro risk-on/off move it more than fundamentals (2022 vs 2023–24 are almost entirely narrative/macro, not earnings); (2) deliveries (volume prints move it hard, both ways); (3) autonomy milestones (robotaxi launches, FSD versions); (4) the election/policy axis (EV credits, AV regulation). Earnings beats per se are a weak catalyst — Q1 2026 beat on EPS and the stock is still down YTD. TSLA trades on story and Musk, then deliveries, then everything else.
CEO Elon Musk — founder-archetype (functionally; chair/CEO/largest shareholder), simultaneously running SpaceX, xAI, X, Neuralink, Boring.
Income statement / balance sheet / cash flow risks (all ; no filing-parsed numbers available):
n/a) — a flag in itself.Regulatory findings (required sub-section). From regulatory/regulatory-findings.md (generated 2026-06-17, SEC EDGAR EFTS LR + AAER): 0 SEC Litigation Releases and 0 AAERs naming Tesla in the 2021-06-17→2026-06-17 window (AAER search hit an EFTS HTTP 500, so AAER coverage is partial). Non-SEC enforcement (web):
n/a this pass).Net: clean on SEC accounting-enforcement record; materially exposed on safety/autonomy regulation and product-liability litigation, which is the live regulatory front for the robotaxi thesis.
Built bottom-up off the latest actuals + guidance. All inputs ; outputs with arithmetic. No forecast.ts create run — this is the unattended breadth loop.
Anchors: FY2025 GAAP net income $3.79B; FY2025 revenue $94.83B; 2026 consensus revenue ~$108.9B (+15%) and 2026 EPS consensus ~$2.08–$2.25; reg credits ~$595M 2026 → ~$0 2027; FY2026 capex >$20B.
| Scenario | FY2026 EPS | FY2027 EPS | FY2028 EPS | Logic |
|---|---|---|---|---|
| Bear | ~$1.30 | ~$1.10 | ~$1.40 | Volume flat-to-down vs BYD/price war; reg credits → $0 by '27 removes ~$0.15–0.20 of EPS; robotaxi/Optimus stay pre-revenue; >$20B capex + rising SBC compress margins. EPS falls in '27 as credits vanish before robotaxi scales. |
| Base | ~$2.10 (≈ consensus) | ~$2.40 | ~$3.20 | Auto stabilizes on cost-out + an affordable model; energy compounds ~25–30%; robotaxi contributes a small high-margin services line by '27–'28; credit loss offset by energy + cost-out. Tracks the ~$2.08–2.25 2026 consensus, modest reacceleration after. |
| Bull | ~$2.40 | ~$3.80 | ~$6.50+ | Robotaxi scales to thousands of cars across multiple metros at high software margin; Optimus reaches paid internal/external deployment; energy keeps +40%-ish. The "robotics company" P&L starts to appear — this is the path that retroactively justifies ~190x. |
Arithmetic note (base FY2028): energy gross profit compounding ~27%/yr off ~$3.8B FY2025 energy GP ≈ ~$7.8B by FY2028; auto GP roughly flat ~$10–11B on cost-out vs ASP pressure; a nascent robotaxi services line adds ~$1–2B GP; net of >$20B/yr capex-driven D&A and rising SBC → ~$3.20 EPS. The dispersion is the point — bear-to-bull FY2028 EPS spans ~$1.40 to ~$6.50+, i.e., the fundamental value is undefined without assigning a probability to autonomy/Optimus working. At ~$405, the market is implicitly underwriting bull-ish outcomes.
Forecast to log (not logged here per loop rules): "TSLA FY2026 non-GAAP EPS ≥ $2.10, p≈0.50, resolves 2027-01-31, tags tesla,deep-dive" — flagged for Connor to log via forecast.ts if promoted.
Bull case. Tesla owns the deepest real-world driving dataset and the most integrated Western battery supply chain; the energy business (+49% GWh, ~30% margins) is a best-in-class compounder already in the financials; and the company has a credible (if late) path to (a) a robotaxi network with software-like margins and (b) Optimus as a multi-trillion-dollar product category. Cost leadership on the 4680 plus an affordable model could re-accelerate volume. Capital allocation is aggressive but pointed at the two biggest TAMs of the decade (autonomy, humanoids). If even the base robotics case lands, today's price is cheap.
Bear case (2–3 permanent-impairment risks). (1) Autos are structurally losing — BYD has overtaken on volume, China share is under sustained attack, ASPs and pricing power are gone, and the reg-credit profit prop hits $0 by 2027, which can turn the core business GAAP-unprofitable. (2) Autonomy may be a Level-2→4 chasm Tesla can't cross on vision-only while Waymo scales L4 — endless FSD-version slippage (V15 now late-2026/early-2027) and teleoperator-assisted "robotaxis" suggest the hard part is unsolved. (3) Key-man + valuation — ~190x forward earnings prices near-certain success on two unproven products run by a serially-distracted, serially-late CEO.
Pre-mortem (18 months out, thesis broke). It's late 2027. FSD V15 slipped again; robotaxi is still a few hundred geofenced, teleoperator-shadowed cars; a high-profile unsupervised-mode fatality triggered a NHTSA grounding; reg credits hit $0 and the core auto business printed a GAAP loss in a quarter; BYD's cheaper models took more global share; the $1T pay package passed and was seen as the top tick. The multiple re-rated from 190x toward an auto-plus-energy SOTP ($120–180 range), a ~55–70% drawdown.
Are multiples too high? On current earnings, unambiguously — ~190x discounts flawless execution on two pre-revenue bets. The only way the multiple is "right" is if you assign high probability to the bull autonomy/Optimus path.
Contrarian view (what the market refuses to see). The market treats Tesla as one story (autonomy will work, so the car decline doesn't matter). The contrarian frame: it's two companies glued together — a declining, soon-credit-less, GAAP-marginal automaker and an excellent energy-storage business — with a giant lottery ticket stapled on. The energy business + cost-leadership are underappreciated; the autonomy timeline is over-believed. The risk isn't that the robots never come — it's the duration: the market is paying a 2027 price for a ~2030 payoff while the auto P&L erodes underneath it.
Dismantling the bull case:
The only profitable humanoid maker and #1 by volume — but the float is a Shanghai-only embodied-AI bet whose revenue is 74% lab demos, not labor replacement, and whose Western TAM is being legislated to zero.
A genuinely differentiated case-handling robotics moat wrapped around an ungovernable accounting-and-concentration core — adverse ICFR opinion + active SEC whistleblower probe + >90% Walmart revenue means the multiple is pricing a clean compounder that the filings say does not yet exist.
World-class robotic hands and the #3 humanoid patent estate on Earth, attached to a balance sheet that already had to sell its best asset (the Apptronik stake) to make payroll — this is an IP carcass walking, far likelier to be acqui-hired or stripped for patents than to IPO independently.