Business

Know the Business

Tesla is no longer one business. It is a hardware-thin, capital-heavy car company funding an AI/robotics platform option, and the market is paying $1.47T (357× trailing earnings) almost entirely for the option, not the cars. The auto economic engine that everyone benchmarks — Model Y, Model 3, Cybertruck — is now a mid-teens-gross-margin, single-digit-operating-margin business in structural decline; the energy storage segment is quietly becoming a higher-margin, faster-growing pool than autos. The biggest analytical mistake here is valuing the auto P&L on car-company multiples, then bolting on a generic "AI premium" — the right frame is to separate what is earning today from what the price requires you to underwrite.

1. How This Business Actually Works

Tesla makes money in three economically distinct ways, glued together by one balance sheet.

Auto is still 81% of the revenue line, but a shrinking share of the gross-profit line. Storage segment gross margin (29.8%) is now nearly double auto's (16.2%) and is growing — the inverse of where the segments were four years ago.

The economic engine in plain terms. Build a vehicle for ~$36k of variable cost, sell it for ~$43k, and recover the gigafactory (each $4–8B fixed) by amortising it across high volume. Above some breakeven utilisation per line, every incremental car drops most of its contribution to operating profit; below it, fixed costs dominate and margins collapse. That is exactly what the last four years show: utilisation rose with volume to FY22, then peer-driven price cuts and a stale lineup pulled per-unit prices down faster than COGS, and operating margin halved.

Where the bottlenecks are. Battery cell supply (now mostly resolved via in-house 4680 + LFP partners), Cybertruck and refreshed-Model-Y ramp execution, the cost of the next-generation $25k vehicle platform, and the regulatory + safety bar for FSD/robotaxi commercialisation. Capex is the swing variable — guided to >$25B in 2026 versus $8.5B in 2025 — because the AI/robotaxi/Optimus push requires hardware spend that auto's current cash generation does not cover.

Bargaining power. Tesla still has more than most legacy OEMs (no dealer network, direct-to-consumer pricing, software-OTA updates, Supercharger network monetised by peers via NACS). It has materially less than it did three years ago: BYD has matched range, beaten cost, and started taking European share at 2× Tesla's pace; Chinese FSD-equivalents are bundled at lower price points; brand value is down 36% in 2025 and Europe registrations have fallen for 13 consecutive months.

2. The Playing Field

Tesla sits at one end of a barbell: the only Western EV-pure-play with positive operating margin, and the only auto company in the world with a trillion-dollar plus valuation premium.

The chart tells you everything about how the market is pricing this group. Toyota earns a real margin (10%) and trades at 0.9× sales. The Detroit OEMs are barely profitable to losing money and trade at 0.2–0.3× sales. The two pure-play EV startups are deeply unprofitable and trade at 3–7× sales on hope. Tesla is alone in the upper-middle: a 4.6% operating margin earning a 15× sales multiple — i.e., the market is paying Tesla a 17× higher revenue multiple than Toyota for half the operating margin. That gap is not an auto multiple. It is an option premium on Optimus, robotaxi, FSD, and energy storage at a much larger scale than today.

What the peer set actually tells you: best-in-class for an auto manufacturer is Toyota's 10% operating margin and double-digit ROE through cycles. Tesla used to clear that bar (16.8% op margin in FY22). It does not now. Whether it gets back there in autos depends mostly on the $25k vehicle ramp; the market has stopped waiting and is paying for something else.

3. Is This Business Cyclical?

Two cycles run on top of each other, and they are out of phase.

Cycle 1 — automotive demand and pricing. Tesla rode the EV adoption S-curve from 2018–2022. Revenue 4×, operating margin -1.8% → 16.8%. The downturn that started in 2023 is the first time Tesla has had to behave like a car company in a competitive market: price cuts to defend volume, used-vehicle deflation, and 13 straight months of declining EU registrations. Revenue actually declined in FY25 for the first time in over a decade. That is what an EV cycle looks like when it stops being a growth story and starts being an industry.

Cycle 2 — capex and reinvestment cycle. Tesla entered a heavy capex phase in 2020–2024 (gigafactories), partially harvested FCF in 2021–2022, and is now re-entering an even heavier capex phase ($25B+ guided 2026, ~3× last year) to fund Cybercab, Optimus, Dojo/AI compute, and Megapack expansion. Each capex cycle has historically preceded several years of margin pressure before the new fixed cost base is absorbed.

The cyclicality lands in three places: price (the contribution margin per car has lost ~1,000 bps in three years), utilisation (FY24 deliveries declined Y/Y for the first time), and working capital + capex (FCF has whipsawed from $7.6B in FY22 to $3.6B in FY24, back to $6.2B in FY25, and is guided negative for the rest of FY26). Working capital is well-managed (large customer deposits, low DSO), so the cash hit will come from capex, not receivables.

4. The Metrics That Actually Matter

Forget the headline P/E. The five things that decide whether the bull or bear case wins:

This is the chart that should change how you think about Tesla. Five years ago, the auto business was the entire economic engine and energy was a money-loser. Today the energy segment earns a higher gross margin than the auto segment, and the gap is widening. If you keep modelling Tesla as "an EV company with a small energy attachment," you are valuing the wrong dollar.

Why these five and not the usual ratios. P/E (357×) is uninformative because earnings are in transition and SBC is depressing the comparable base. EV/EBITDA flatters Tesla by capitalising the same R&D legacy automakers expense. Revenue growth is moot — flat-to-down for two years. The five metrics above are the ones that change the thesis when they move; everything else is downstream.

5. What Is This Business Worth?

This is the rare auto company where sum-of-the-parts is genuinely the right lens, because the consolidated numbers blend an established (decelerating) car business, a high-margin compounding storage business, and three pre-revenue platform options. Underwrite each on its own terms.

What would justify a premium: (1) Megapack scaling past $25B revenue at 30%+ gross margin without margin sacrifice, (2) FSD take-rate above 25% on the installed base, (3) any commercial robotaxi unit-economics demonstration credible enough to underwrite a Waymo-comparable business, (4) Optimus shipped to a third party at a price someone willingly paid.

What would force a discount: (1) auto operating margin staying below 6% for another year, (2) BYD reaching North America via Mexico, (3) FSD regulatory setback in any major market, (4) capex remaining above $20B/yr without commensurate revenue growth in the new businesses, (5) any meaningful Musk distraction or departure.

6. What I'd Tell a Young Analyst

Stop trying to be precise. The interesting question is not "what is Tesla worth?" — at $1.47T it is worth whatever you think the option on Optimus and robotaxi is worth, because the current operating business does not pay for the rest. The interesting question is: what evidence in the next four quarters would make the option more or less likely to pay off?

Three things to actually watch, in order of decision-relevance:

1. Auto gross margin ex-credits, every quarter. If it stabilises in the 14–16% band, the auto business is a viable cash funder of the option. If it breaks below 12% on the next $25k vehicle ramp, the auto business is no longer subsidising the AI bet — it is competing with it for capital, and that breaks the thesis.

2. Free cash flow after SBC, every quarter. Headline FCF was $6.2B in FY25; SBC was $2.8B; the real number is $3.4B against a $1.47T market cap (0.23% yield). With FY26 guided FCF negative, the company will be issuing equity or debt to fund the build-out — dilution is the silent cost.

3. Megapack revenue and margin trajectory. This is the only segment with both growth and improving margins. If the energy business reaches $25B revenue at sustained 25%+ gross margin within two years, it stops being a footnote and becomes a real second engine.

What the market is most likely underestimating: how good the energy storage business has quietly become. What the market is most likely overestimating: that Optimus and robotaxi will commercialise on the timeline implied by the multiple. Both can be true simultaneously, and that is exactly what makes this stock harder to short than to admire.

The one trap to avoid: do not benchmark Tesla against Toyota, GM, and Ford on auto margins to "prove" it is overvalued. The market knows the auto math. It is paying for what is not yet earning a dollar. Whether that bet pays off is a question about Optimus, FSD, and Cybercab — not about Model Y pricing in Q3.