Tesla

IThe business

Tesla sells cars, energy storage and services. In the twelve months to September 2026 its revenue was $103.6B, up 8.3% on the twelve months before, after $94.8B in 2025, which was down 2.9%.⁴⁹¹⁰²³ From 2019 to 2025 revenue grew 25.2% a year.⁹ Operating income was $3.5B, a margin of 3.3%, against 16.8% in 2022.⁴⁹ Without regulatory credits and tariff refunds it was $1.84B on $102.7B, or 1.8%; before the $4.3B rise in R&D that I charge to autonomy it was $6.16B, or 6.0%.C (calculated)³⁴

The business has three lines, and they earn different money. Automotive, which holds vehicle sales, leasing and regulatory credits, brought in $75.0B and $14.3B of gross profit. Services and other, which holds used cars, service, paid Supercharging, insurance and merchandise, brought in $16.3B, up 36%, and $2.0B of gross profit. Energy generation and storage brought in $12.3B, up 2%, and $3.6B of gross profit.C (calculated)⁴⁵¹⁰²³ Tesla reports two segments and no segment operating income, so gross profit is the only profit by line. Every vehicle-versus-energy margin on this page, including the energy margins in my cases, is my allocation of costs by revenue.⁵

Per unit, a delivered vehicle brought in about $42,400 of revenue and $7,500 of gross profit without credits, and a kilowatt-hour of storage brought in $244 and $73.C (calculated)⁴⁵

What drives each line

• Cars. Deliveries peaked at 1,808,581 in 2023 and were 1,742,908 in the last twelve months. Production was 1.76M, at most 74% of installed capacity of more than 2.38M for Model 3, Model Y and Cybertruck.C (calculated) In Q3 2026 production of 464,391 was below deliveries of 486,532, and days of supply fell from 15 to 11. That fits Tesla’s statement that battery supply, not demand, limited volume, alongside its largest passenger-vehicle backlog in four years. The 74% says factory capacity was not the limit, and I cannot test the battery claim.^4111819 Model S and Model X ended in mid-2026.³
• Energy. Storage deployments grew 49% to 46.7 GWh in 2025 and were 50.2 GWh in the last twelve months.C (calculated) Energy revenue fell 6% in the first nine months of 2026, or 2% after adding back $339M of tariff refunds owed to customers, on a lower price per Megapack, fewer Powerwall deployments and the refunds, while deployments rose 11% to 36.0 GWh.C (calculated) Megafactory Texas, planned at 50 GWh, began production in Q3 2026, and Tesla reports its highest order rate and backlog. My base case needs energy revenue to grow 15% in year one, and I cannot verify that.³⁴¹⁸
• Services and other. Gross profit reached a record $734M in Q3 2026, and the line grew 36% over the year.⁴
• Autonomy and robotics. Full Self-Driving (Supervised) had 1.74M active subscriptions, up from 1.04M. Tesla’s Q2 2026 Update named seven Robotaxi metros: the Bay Area with a safety driver, and Austin, Dallas, Houston, Miami, Orlando and Tampa without one. Tesla began Robotaxi rides with Cybercab in Austin on 3 September 2026 and reports nearly 2M cumulative unsupervised miles. Optimus production began in Fremont. Tesla gives no revenue, ride, fleet or unit figure for any of them.³⁴⁴⁵

The United States was 50.2% of 2025 revenue and China 22.1%; in the first nine months of 2026 the shares were 46.2% and 17.9%.C (calculated) China revenue was $14.07B in the nine months, 1.4% below a year earlier and 7.9% lower in Q3, while revenue outside the United States and China rose 49%.C (calculated)³⁵

What changed

Research and development was 4.1% of revenue in 2023 and is 8.6% in the last twelve months, $8.9B.C (calculated) Capital spending was $8.5B in 2025, $15.4B in the last twelve months, and is guided above $20B for 2026 in the 10-Q; on the Q3 call Taneja said about $20B.⁵⁴ Investment in what Tesla calls Super Intelligence (SI) infrastructure, which includes its Cortex training clusters, went from $6.8B to $11.0B at cost in nine months.³⁴⁵ The 10-Q says periods of heavy spending:

will necessitate additional funding beyond our operating cash flow

It signed $30B of undrawn credit lines on 29 September 2026.³²¹

Cars and energy pay for a business that has no revenue line.

IIThe moat

I think Tesla had a manufacturing cost advantage that its margins since 2022 do not show, and has a set of assets whose worth depends on autonomy working. It has no switching costs. A Tesla owner can choose any car at the next purchase, and the charging network now serves other makers.

The cost advantage showed in 2022, when automotive gross margin without regulatory credits was 26.7%. It is not visible now: the margin was 15.4% in 2024 and 2025 and is 17.8% in the last twelve months, or 17.4% without the Q3 tariff refund.C (calculated)³⁴⁵⁹¹² Against BYD the ratios do not show a cost lead either, as the next section sets out. In 2025 automotive sales fell 9%, on a decrease of about 8% in cash deliveries and a lower average price from mix and incentives, and 2023 is still the volume peak.⁵

Vertical integration is the other lever. Tesla makes 4680 cells at over 40 GWh a year in Texas, has begun lithium iron phosphate cells, cathode and lithium refining, and says battery availability still limits vehicle volume.⁴ The spending widens the asset base. Whether it widens the moat depends on cost per kilowatt-hour, which Tesla does not report.

Tesla’s charging network gives it coverage, and other makers can use it, so I do not count it as a moat. It had 8,904 Supercharger stations and 84,465 connectors at Q3 2026, up 14.8% and 14.4% in a year, and the 10-K says all major automakers have announced adoption of Tesla’s connector standard in certain markets.⁴⁵ Customer deposits rose from $1.31B to $1.88B in nine months, and Tesla held more than half of U.S. electric vehicles in Q3 2026 in a market that fell 39.5%.³²⁶

The asset that could become a wide moat is the one the price rests on. Tesla has a lead in supervised driving data and in U.S. share, and none in unsupervised miles: it reports nearly 2M cumulative unsupervised Robotaxi miles, and Waymo reports 271.3M rider-only miles through June 2026.⁴²⁷ Cumulative deliveries of 10.2M, 1.74M paying FSD subscribers and over 160M kilometres logged in nine European countries are a large driving-data position, and whether it becomes a lead in unsupervised driving is the open question.⁴ In energy storage Tesla is among the top two on Wood Mackenzie’s 2025 ranking, with Sungrow, which does not give the order, and second on Benchmark’s, behind BYD, in a market where Chinese integrators hold 76%.²⁸²⁹

Automotive gross margin without credits was 26.7% in 2022 and is 17.8% today, and the moat the price needs is not yet in the numbers.

Regulatory credits are payments from rivals that miss emission rules. They were $1,993M in 2025 and $1,181M in the last twelve months, 34% of operating income, and Q3 2026 credit revenue was $113M after a $475M reversal. The contracted backlog fell from $4.68B at the end of 2024 to $384M at 30 September 2026.³⁴⁵¹⁰ The 10-K says credits carry negligible incremental cost, so I subtract their full revenue.⁵

IIIThe competitive advantage

Tesla’s advantage over named rivals is largest in the United States and in supervised driving data, and smallest in global volume, in Europe, in cost and in storage. Every rival figure here is third party or from an industry body.

Where Tesla leads

In the United States Tesla held more than half of the 262,649 electric vehicles sold in Q3 2026, in a market that fell 39.5% after federal incentives ended, though its own sales fell 20%.²⁶ Its cars sit at 11 days of supply.⁴ The lead I would defend is the one in the United States, the charging coverage and the supervised driving data. I cannot show a cost lead over BYD. Without regulatory credits Tesla’s gross margin was 16.3% in 2025 against BYD’s 17.74%, and 15.5% in 2024 against 19.44%.C (calculated)⁹³⁷³⁸ Both are company-wide figures, and BYD’s includes a handset assembly business that was 19% of its 2025 revenue, so the comparison is between companies, not cars.³⁷

Where it trails

BYD sold more battery-electric cars than Tesla in 2025 and in Q2 2026, and sells 2,288,709 plug-in hybrids that Tesla has no match for.²⁴²⁵ In Europe BYD registered 234,099 vehicles in January to August 2026, 22% more than Tesla’s 191,787, and Tesla’s growth slowed to 4.0% in August from 43.3% for the eight months. ACEA counts every powertrain, and BYD sells plug-in hybrids.C (calculated)²⁵³² Waymo reports more than 500,000 rides a week, against Tesla’s none, and lists 15 service areas. Tesla’s equivalent is the seven metros in its own Update and no reported ride count.³⁰³⁹⁴⁵ In storage BYD led Benchmark’s 2025 ranking with 13% against Tesla’s 10%, and Tesla deployed 46.7 GWh against over 60 GWh for BYD.¹⁸²⁹ China was 17.9% of nine-month revenue and I found no primary share series for Tesla there from CPCA or CAAM; Tesla’s own China revenue was lower, as the business section shows.³ Tesla itself names competition in AI and robotics as a risk.⁵

IVManagement

Elon Musk has been chief executive since October 2008 and a director since 2004. He is also chief executive of SpaceX, runs Neuralink and The Boring Company, and led xAI Holdings, which became a SpaceX subsidiary on 2 February 2026. Tesla’s 10-K says it depends heavily on him, that he does not devote his full time to Tesla, and that none of its key employees has a fixed-term contract.⁵⁸ Vaibhav Taneja has been chief financial officer since August 2023 and chief accounting officer since March 2019. Tom Zhu, who led the construction and operations of Gigafactory Shanghai, runs Asia-Pacific and global vehicle manufacturing.⁸

The proxy names three executive officers: Musk, Taneja and Zhu. None runs autonomy or Optimus, and no Tesla filing I read names who does, so I cannot give a reader the tenure of the people who lead the two businesses the price rests on.⁸ The only senior departure the 8-Ks record since January 2024 is Andrew Baglino, senior vice president for powertrain and energy engineering since October 2019, who resigned on 14 April 2024.²²

From 2019 to 2025 revenue grew 25.2% a year, from $24.6B to $94.8B, and gross profit went from $4.1B to $17.1B.⁹ The capital spending record is mixed. Tesla spent above the range it gave in 2022 ($7.2B against $5.0B to $7.0B) and 2023 ($8.9B against $6.0B to $8.0B), met its outlook of over $10.0B for 2024 with $11.3B, and fell $2.5B short of over $11.0B in 2025, when it spent $8.5B.^510111213 Its Q3 2026 update gives no number for deliveries, margin or earnings.⁴

The record that matters more is the one on autonomy, because the price rests on it. Tesla was on time on three milestones it set for itself: Austin in June 2025, Cybercab production in the first half of 2026 and the start of Optimus production in 2026. It was late on five of the seven cities it planned for the first half of 2026. It has not delivered the 2019 aim of customer cars that drive unsupervised, and it has not reported a Cybercab cost.^44041434445

Who holds the votes

Musk beneficially owns 699.6M shares, 19.8%, of which 413.2M are in his trust and 286.4M are restricted shares from exercising the 2018 award, which vest on service to 19 January 2028. He has pledged 207.5M shares, worth $84B at the closeC (calculated), 29.7% of his stake. The board caps his pledge-backed borrowing at the lesser of $3.5B and 25% of the pledged value, 4.2% of that value at the closeC (calculated), but exempts pledges for taxes and the $334.09 award offset. The 10-K says that if he borrows against his shares and the price falls substantially, lenders could force sales.⁴⁹⁵⁸

The proxy computes Musk’s 19.8% on a base that leaves out the 423.7M unearned award shares, as its footnote 2 says, and every other holder’s percentage on 3,952.6M shares. I use 3,529M for market value because the unearned shares vest only on milestones.⁸ The board has nine directors in three staggered classes and judges seven independent; the table gives the chair, Musk’s brother, a supplier’s chief executive and the 2025 say-on-pay vote.⁸

Pay and incentives

Musk takes no salary or bonus. His pay is two stock awards. The 2018 award, reinstated by the Delaware Supreme Court on 19 December 2025, was exercised in the second quarter of 2026. The 2025 award is 423.7M restricted shares in twelve tranches. Each tranche needs a market value and a number of operating milestones: the first needs $2.0T and any one milestone, and the twelfth needs $8.5T, all twelve milestones and a board-approved succession framework. Vested shares carry an offset of $334.09 each, which works like a strike price.³⁸²⁰

The milestones are Tesla’s own chosen yardsticks: 20M vehicles delivered, 10M FSD subscriptions, 1M bots, 1M Robotaxis, and Adjusted EBITDA from $50B to $400B over four quarters.⁸ Tesla has reached 10.2M vehicles and 1.74M subscriptions and has $15.0B of EBITDA.⁴ Its accounting treats only the 20M-vehicle milestone as probable: $802M of expense was recognized in nine months, and $105.8B to $120.4B is unrecognized for the rest.³ The proxy shows 2025 compensation of $158.4B on the summary table, of which $132.3B is the maximum value of the 2025 award and $26.1B an interim award forfeited on 21 April 2026. Realized pay was zero.⁸

The award ties Musk’s pay to market value and operating milestones, not to return on capital or free cash flow, and free cash flow after stock pay is negative. Its first step is a market value of $2.0T. On the full 3,953M shares the close gave $1,594B, 80% of the step; the 2025 proxy does not say whether unearned shares are in the count the award uses.⁴⁹³⁸²⁰ If every tranche were earned, Musk would hold 28.4% before the offset.⁸

Capital allocation

Tesla has returned no cash to owners. Over 2016 to 2025 it generated $79.4B of operating cash flow, spent $55.0B on capital and kept $24.5B of free cash flow.C (calculated) It sold $15.2B of stock, the last $12.3B in 2020, and diluted shares rose 5.6% a year, from 2.16B to 3.53B.C (calculated) The 10-K says Tesla has never paid a cash dividend; I found no repurchase programme.⁵⁹

Three uses sit outside the car and energy business. Tesla bought $1.5B of bitcoin in the first quarter of 2021 and holds 11,509 coins worth $962M against a $386M cost.³¹³ It invested $2.0B in SpaceX common stock in March 2026, after a related-party review; the stake is carried at $2.865B, and at the Q4 2026 SpaceX close it would be $2.56B. It paid $1.95B in stock and awards for a Super Intelligence (SI) hardware company in the second quarter of 2026; the 10-Q does not name the seller.³⁶⁴⁷⁵¹ SpaceX bought $626M of Megapacks in nine months, and Tesla’s receivable from it is $729M, mostly for products not yet delivered.³

Tesla’s 10-Q also says a committee of disinterested directors is exploring a deeper relationship with SpaceX, including a business combination, and that there is no assurance of a transaction. SpaceX’s board has formed a committee too. Musk would sit on both sides of any deal, and my values assume none.³⁴⁸

The pay shows what he is paid to do. The 2025 award vests on market value and Adjusted EBITDA, not on cash left after stock pay, and its first step is $2.0T.⁸ His 19.8% is large, but 286.4M of those shares came through the 2018 award and 207.5M are pledged, so I cannot say how much is his own cash. Over 2016 to 2025 Tesla kept $24.5B of free cash flow, paid none out, sold $15.2B of stock and added 5.6% a year to the share count.⁵⁹ The pay rewards size: revenue grew 25.2% a year from 2019. Cash after stock pay was $3.4B in 2025 and −$1.0B in the last twelve months.⁴

Management has built a large autonomy bet on cash from cars and energy, and has not yet shown what it returns.

VFinancials

Tesla earned $3.5B of operating income, and $1.6B of it is regulatory credits and tariff refunds. Three items sit inside or below it.³⁴

1. Regulatory credits were $1,181M, 34% of operating income. They are running off.³⁴
2. Tariff refunds added a net $447M to Q3 2026 gross profit. Without them the Q3 gross margin was 17.8%, not 19.6%, and the energy margin 23.8%, not 33.1%.C (calculated)³
3. Interest income of $1.75B sits below operating income and equals 50.5% of it. Net income of $3.2B also holds an $863M SpaceX mark, a $353M bitcoin loss and a net $230M benefit from tax items.³⁴

After all R&D, vehicles, services and energy earned $1.84B on $102.7B of revenue, a 1.8% margin. Before the $4.3B rise in R&D that I charge to autonomy they earned $6.16B, 6.0%.C (calculated)³⁴ Stock pay was $4.3B, 1.34 times net income.C (calculated)⁴

Cash

Operating cash flow was $18.7B and capital spending $15.4B, so free cash flow was $3.3B. After $4.3B of stock pay it was negative $1.0B.C (calculated)⁴ Four free cash flow figures appear on this page, and the table bridges them. I start the reverse DCF from 2025 because the twelve-month figure is negative, and no growth rate turns a negative start into a positive cash flow.C (calculated) Tesla counts energy-system purchases as capital spending from 2025; I apply that back to 2016, which lowers earlier free cash flow by up to $0.7B a year.⁴⁵ Nine months of 2026 capital spending were $13.0B, so the guide of over $20B implies $7.0B or more in Q4, against $4.7B in Q3.C (calculated)³⁴

Revenue was $22.4B in Q1 2026, the weakest quarter of the year, on 358,023 deliveries, and $28.2B in Q2.⁴⁷

The balance sheet

Cash and short-term investments were $45.4B at 30 September 2026 against $9.5B of debt and finance leases, almost all non-recourse, and $0.6B of minority interests.³ That is net cash of $35.3B, or $39.2B with the SpaceX stake and bitcoin. Operating leases of $6.9B and a warranty reserve of $10.1B are not deducted; their cost runs through gross margin.³ The $30B of new credit lines include a $20B delayed-draw loan, and the 10-Q says spending may need more funding.³²¹

Accounting to watch

1. Capitalized compute. SI infrastructure at cost rose from $6.8B to $11.0B in nine months, with $11.2B more in construction in progress. Construction in progress is not depreciated until it is in service.³
2. Taxes. The effective rate was 27% in 2025 and 16% in Q3 2026, and 2023 carried a $5.0B net benefit. The CEO award expense is not tax-deductible.³⁵¹¹
3. Related-party revenue. SpaceX and xAI are customers and Musk’s companies; $648M of deferred revenue and a $729M receivable sit against SpaceX.³
4. Marks. The SpaceX stake is carried at its 30 September closing price, a quoted Level 1 price, with sale restrictions that end in December 2026. At the Q4 2026 close of $134.72 it would be worth $2.56B, $307M less than carried, or $0.09 a Tesla share.C (calculated)³⁴⁷⁵¹
5. Warranty and litigation. In Benavides the jury awarded $129M of compensatory damages and found Tesla 33% at fault, plus $200M of punitive damages. That makes Tesla’s share about $243M.C (calculated) Tesla has appealed to the Eleventh Circuit and recorded an immaterial accrual. The Ninth Circuit vacated the FSD class certification on 16 October 2026.³

VIValuation

Before the numbers

I start with how this goes wrong, in three ways. Autonomy comes late or small while the build is paid for: with no autonomy revenue and the spending kept, my base is $21 a shareE (estimate).⁴ The cars stay at today’s margin: the core earns 1.8% after all R&D, and credits are 34% of operating income and running off.³⁴ Or control goes wrong: one person runs the company, has pledged 29.7% of his stake and would sit on both sides of a SpaceX deal.³⁸ I think the first is the most likely.

Next, what an owner can take out. Net income was $3.2B, with $1.75B of interest on cash in it. Free cash flow was $3.3B, and after $4.3B of stock pay it was −$1.0B.³⁴ That is harsh, as most of the $15.4B of capital spending is for growth. If standing still cost only the $6.7B of depreciation, owner cash would be $7.7B, with interest and $1.18B of fading credits still in itC (calculated). Tesla does not say what standing still costs, so I show the range: −$1.0B to $7.7B, or −0.1% to 0.5% of the $1,423B market valueE (estimate).

I cannot know how large autonomy gets or at what margin, because Tesla reports no fleet, ride count or Robotaxi revenue. So I give a range: $15 to $303 a share, with $65 in the middleE (estimate).

What the price implies

At $403.30 the market value is $1,423B on 3,529M shares, and the enterprise value $1,387B.⁴⁹⁵⁰³⁴ The share count matters: the 10-Q cover shows 3,953M shares, which includes 423.7M unearned award shares, and that count gives $1,594B. The proxy computes Musk’s percentage without those shares, so I use 3,529M for market value and 3,541M, the Q3 diluted average, for value per share.³⁸

Tesla’s free cash flow after stock pay was negative in the last twelve months,⁴ so no constant growth rate gets from today’s cash flow to the price. Starting from 2025’s $2.05B, the price needs free cash flow to grow 39.4% a year for fifteen years, to $300B in year 15, 147 times the start.C (calculated) A constant rate is a crude shape, so the next table turns the price into autonomy revenue.

What the price assumes

Value depends on one thing the filings do not show: how large autonomy and robotics become, and at what margin.³⁴ My base case for the other lines is vehicles and services growing 5.4% a year to $152B in year ten at a 9% operating margin, and energy growing 11.4% a year to $37B at 18%.E (estimate) Everything else is autonomy and robotics. Holding all of that, the price needs $600B of autonomy and robotics revenue in year ten at my 40% margin after depreciation. Total revenue would be $790B, operating income $253B and Adjusted EBITDA $337B.C (calculated)

The price needs year-ten Adjusted EBITDA near $337B, more than 22 times the last twelve months and above Tesla’s fifth award step of $300B.

If all of that revenue were rides at $70K of revenue per vehicle-year, a figure I chose and Tesla does not report, the fleet would be 8.6 million robotaxis. Tesla’s own award asks for 1M. Waymo’s fleet is over 4,000 on a third-party report.E (estimate)³⁸³¹ At about 125 rides a vehicle a week, Waymo’s approximate pace from an April ride count and a September fleet count, $70K is a fare of $10.77 a ride.C (calculated)³⁰³¹

Even a software-like 60% margin leaves the price needing $368B of autonomy revenue.E (estimate)⁴⁹⁴ My bull case has $265B at 50%, and it earns 7.8% a year at the price.

The discount rate

I use 10%, the rate on my NVIDIA, Palantir and TSMC pages, so that the pages can be compared. For Tesla it is the favourable end of the range. The capital asset pricing model gives 12.7% on Damodaran’s own basis: his 4.20% equity risk premium of 1 October is measured against a 5.29% Treasury yield, and Tesla’s beta against the S&P 500 is 1.75 over 60 months. Pairing the premium with the 5.00% yield of Q4 2026 mixes dates and gives 12.4%. The 36-month beta of 1.68 gives 12.4% and the ten-year beta of 1.92 gives 13.4% on his basis.C (calculated)³³³⁴³⁵ Beta measures how the stock moves with the market and nothing else, and my case weights are probabilities, not a risk premium. At 12.7% my base case is worth $41, my bull case $179 and my probability-weighted value $67, against $110 at 10%.E (estimate)

Three cases

Each case runs three businesses for fifteen years and then a terminal value: vehicles and services, energy, and autonomy and robotics. Vehicles and services start from $90.1B of revenue and energy from $12.6B, with the margins in the table below, which remove credits and refunds and charge the $4.3B rise in R&D since 2024 to autonomy.C (calculated)³⁴⁹ Autonomy and robotics start at zero revenue and carry that cost.

I count stock pay as a cost, as under GAAP, which includes about $1.1B a year of expense on the 2025 award, and tax at 20%, against 27% in 2025 and 16% in Q3 2026. Each case also spends on compute and the fleet before revenue arrives.E (estimate)³⁵

The vehicle margin is the assumption that moves the core most, so the table sets it against the past. Without credits and refunds, company operating margin was 14.9% in 2022, 7.5% in 2023, 4.5% in 2024 and 2.5% in 2025. In the last twelve months it was 1.8%, or 6.0% before the R&D rise. My base case reaches 10.8% for vehicles, services and energy by year ten, between 2023 and 2022, and it needs volume back toward installed capacity, because factories made at most 74% of what they can. The bear case holds today’s 5% on vehicles, so it is not a shock to the core.C (calculated)³⁴⁹¹²

The bear case is that autonomy stays supervised and Tesla stops the build by year three. Price competition holds vehicles to a 5% margin, about today’s level, and energy to 10%. Free cash flow is negative for two years and reaches $2.4B in year ten. The value is $15: the core at today’s margins with no recovery is worth $23 with cash, the autonomy spending takes off $7, and cash is $11 of the total.E (estimate)³⁴

The base case is that autonomy works in U.S. cities at 0.5 million robotaxis, which is at most 125 times Waymo’s fleet today and 50,000 added a year for ten years, with $35B of software and robots on top, and the business earns a 40% margin before its R&D. Vehicles earn 9% and energy 18% once volume recovers. Free cash flow is negative for four years, $29B in total, which cash covers, and reaches $23.2B in year ten.E (estimate)³⁴³¹ Year-ten Adjusted EBITDA is $69B, above the award’s first step and below its second.⁸

The bull case is that the plan works: 2.5 million robotaxis, $90B of software and robots, a 50% autonomy margin, 13% on vehicles and 22% on energy. Year-ten revenue is $522B and Adjusted EBITDA $214B, past four of the award’s six EBITDA steps. The year-ten value of $689 a share would vest two tranches, which costs holders $3 a share and overlaps, by at most that, the award expense already in my margins.E (estimate)⁸²⁰ Growth after year 15 is 2.5% in the bear case, 3.0% in the base and 3.5% in the bull, each continuing its last forecast year for vehicles and services; at 3.0% the bull value is $288.

I weight the cases 30%, 45% and 25%, which gives $110. The weights are my judgement. No weighting reaches the price: 100% on the bull case is $303, and the table shows two others.E (estimate)³⁴

What closes the gap

My base value for vehicles, services, energy and cash is about $44 a share if margins recover to 9% and 18%: vehicles $22, energy $12 and cash $11.E (estimate)⁴⁹³ That is 11% of the price. At today’s margins it is $28 before the $4.3B rise in R&D and $9 after it, and the table shows the range, including a floor for a shock to the core. The base case adds $21 for autonomy after its cost, and the remaining $338 is not in my base case. In my base case autonomy is worth −$3 a share over the first fifteen years and $24 after them, so the value after year 15 is 113% of its $21.E (estimate)

One input does most of the work. Swapping in the bull case’s autonomy revenue alone takes the base from $65 to $206; its margin adds $60, vehicles $20 and energy $8. The bull case’s R&D and capital path lowers the value by $7, and its 3.5% growth after year 15 adds $15.E (estimate)³⁴

Other single changes to the base: autonomy revenue of zero, with the spending kept, gives $21; a vehicle margin 2 points lower gives $58 and 2 points higher $73; an energy margin 4 points lower gives $62 and 4 points higher $69; energy growth of 5% in years one and two gives $63; a 25% tax rate gives $60; $50K of revenue per robotaxi in place of $70K gives $59; a 30% autonomy margin gives $51.E (estimate)³⁵ The value after year 15 is 60% of my base, 20% of the bear and 66% of the bull.

The margin of safety

• Price$403.30Close, Q4 2026 [49]
• Base value$65My estimate E (estimate)
• Margin of safety below$46my base value less 30%, to the nearest dollar C (calculated)

My rule for the discount to base value is 25% for a business with net cash and no customer above a tenth of revenue, 30% when one cycle or two customers drive results, and 35% when fixed charges such as leases sit ahead of the shares. Tesla’s vehicle earnings follow the car cycle, so it takes the 30% tier. Two facts argue against going lower: one person controls both the company and the outcome the price rests on, and the board’s cap on his pledge-backed borrowing exempts pledges for taxes and the award offset.³⁵⁸ That gives $46. The price is 777% above that line, so the stock would have to fall 89% to reach it.C (calculated)

I did not set the line from the core value. They land close, $46 against $44, only if margins recover; at today’s margins vehicles, services, energy and cash are worth $28, below the line. At the line my base case returns 13.2% a year and my bear case loses 4.4%, holding fifteen years and selling at my year-15 value at a 10% rate.E (estimate) At the CAPM rate of 12.7% my base is $41, so the $46 line sits above it, and the same discount would put the line at $29. The lowest close of the past year was $298.32, on 29 July 2026, still 6.5 times the line.C (calculated)³⁶⁵²

Where this could be in 5 and 10 years

I value each case at a date the way I value it today. It is the present value of the free cash flow after that date, per share, in that year's dollars.E (estimate)

For my base to be right at five years, Tesla has to reach $163B of revenue and a 9% operating margin by September 2031. Free cash flow has to turn positive, at $1.0B, after $29B of build. For ten years it has to reach $260B of revenue and $23B of free cash flow. That needs 0.5 million robotaxis and $35B of software and robots at a 40% margin.E (estimate) The base value goes from $65 to $109 to $145 mostly because time passes at a 10% rate. Owners get $13 a share in cash over the ten years.

The cases differ on three numbers. One is autonomy revenue in year ten: $3B, $70B or $265B. One is the margin on vehicles, 5%, 9% or 13%. One is the capital the growth needs. The bull's $689 comes from 2.5 million robotaxis. Even there, today's price earns 6.8% a year over ten years, under my 10% rate.E (estimate)

The most likely way my base is wrong is autonomy. It supplies $70B of year-ten revenue from zero, and Tesla reports no robotaxi revenue, fleet or ride figure. At a 30% margin instead of 40%, the base is $51 today, not $65. At a 12.7% discount rate it is $41.E (estimate)

The next best use of the same dollar is another page here. Alphabet owns Waymo, the robotaxi business Tesla is chasing, and my base case returns 5.4% a year over ten years on it, against −8.9% for Tesla.³⁰⁵⁸ Quality and price are separate questions. Tesla looks like a business I cannot yet judge, at a price that does not wait: 6.2 times my base value. The price needs several things at once: $600B of autonomy revenue, a 40% margin and vehicles back at 9%. At even odds on each, all three come to one in eight, and $600B is more than twice my bull case’s $265BE (estimate).

These are estimates of value, not forecasts of the share price. The market price can sit above or below value for years.

VIIThe case for the price

What most people miss

Tesla’s Q3 Update lists a Cybercab line of more than 125,000 a year in Texas. That is 2.5 times the 50,000 a year my base case adds, and it sits outside the 2.38M of car capacity I used earlier.C (calculated)⁴ Tesla’s Lars Moravy said on the call that the listed figure is the rate Tesla can run today, and that “the actual capacity of that vehicle line is significantly higher”.⁵⁴

Energy customers are paying ahead. Their prepayments rose from $2.04B to $3.30B in nine months, while energy revenue fell 6%. SpaceX accounts for $648M of deferred revenue, and contract revenue due within a year slipped from $4.96B to $4.80B.³⁵ Musk said Tesla is “almost sold out of all Megapacks for 2027”.⁵⁴

Half of FSD’s more than 1.7M paying customers now pay by subscription, Vaibhav Taneja said. The filings do not give that split.⁵⁴

The bull case

The fullest case is ARK’s. On 19 October 2026 Tasha Keeney and Daniel Maguire published an open model that runs 32 inputs through a simulation. It gives an expected value of about $4,400 a share in 2031, with about $3,300 and $5,000 at the 25th and 75th percentiles. Its base has 8.1M robotaxis, $753B of net ride-hail revenue, $212B from Optimus and $902B of EBITDA, and it puts about 80% of enterprise value on robotaxis.⁵⁷ The workbook on GitHub shows a saved run that averages $4,342, and ARK notes that each run differs.⁵⁵

Ron Baron puts it per car. His letter of 4 August 2026 says a robotaxi will cost Tesla $30,000 and that robotaxis “could each earn Tesla $50,000 to $60,000 cash flow every year”. A car sold once earns $7,000 to $10,000.⁵⁶ On the Q3 call Taneja said that “using exponential math” Robotaxi would become a significant part of revenue and profit in 2027. Musk said it starts to show around Q2 2027 and is “probably a big deal by Q3 or Q4”.⁵⁴

Against reality

To reach the price, my model needs $600B of autonomy and robot revenue by September 2036 at a 40% margin, with everything else on my base case. That is $337B of Adjusted EBITDA, profit before depreciation and stock pay. Today it is $15.0B.E (estimate)⁴

If Tesla has about 257,000 robotaxis and $34B of autonomy and robot revenue by September 2029, then 857,000 and $120B by September 2031, reaches $600B by September 2036, and vehicles and energy hit my base margins, the price is possible.E (estimate) That is my base path, scaled up 8.6 times.

Today’s numbers are a long way off. Tesla reports nearly 2M unsupervised miles, but no ride count, fleet count or Robotaxi revenue.⁴ The revenue needed is 8.6 times my base and 2.3 times my bull.C (calculated) My base gives one robotaxi $28.0K of operating profit a year, about half of Baron’s figure, on a stricter measure.E (estimate) ARK’s reported 2031 EBITDA of $902B is 2.7 times the $337B, about five years sooner, on a different definition.C (calculated)⁵⁷

The Cybercab line at its listed rate makes 375,000 by September 2029 and 625,000 by September 2031.C (calculated) That clears the first date. The second needs more lines or customers’ cars. Taneja said the key is regulatory approval to deploy them as fast as Tesla can build them.⁵⁴

What they have to hit

Tesla gives no fleet count, so these are six lines it does publish. The FSD, margin and energy thresholds match my signposts in the next section.

Where I come down

The Cybercab line moved me more than the prepayments, because the factory does not stop the 2029 path. ARK’s fleet did not, at 16 times my base fleet and five years sooner. A reported autonomy revenue line tracking $4B in year three and $14B in year five would raise my values. So would the listed line run full for ten years: 1.25M robotaxis against my 0.5M lifts my base to about $98, still under a quarter of the price.E (estimate)

VIIIRisks, and what would change my view

1. Autonomy does not arrive, or arrives slowly. This is the risk the price carries. With autonomy revenue at zero and the spending kept, my base value is $21. A two-year delay to the whole ramp, with the spending kept, lowers the base to $50.E (estimate) The 10-Q says scaling autonomy “presents significant challenges due to evolving regulations”.³
2. Key person and control. Tesla depends on one person who runs several companies, holds 19.8% and has pledged 29.7% of it. The award can add up to 423.7M gross shares, and unearned shares vote in proportion to other votes.⁵⁸
3. Thin core profit. Core operating margin is 1.8% after all R&D and 6.0% before the rise; credits, refunds, interest and marks flatter reported profit, and credits are running off.C (calculated)³⁴
4. Capital intensity. Free cash flow is $3.3B on $15.4B of capital spending, and spending is guided above $20B.³⁴
5. Competition and trade. BYD leads global battery-electric volume and, so far in 2026, European registrations; Chinese integrators hold 76% of storage; the Supreme Court voided the IEEPA tariffs in February 2026, which created the Q3 refunds, and Tesla says the tariff regime hits energy storage more than automotive; and U.S. electric vehicle sales fell 39.5% after incentives ended.^324262832
6. Product liability and regulation. The Benavides verdict, the FSD class action, a securities suit over Autopilot, FSD and Robotaxi statements, and requests from NHTSA, the NTSB, the SEC and the Department of Justice.³
7. Related parties and a possible combination. Megapack sales to SpaceX and xAI, a $2.0B SpaceX stake and a $1.95B acquisition from an unnamed seller. A committee of Tesla’s disinterested directors is exploring a business combination with SpaceX, no terms exist, and my values assume none.³⁸⁴⁸
8. Gaps in what I can verify. Tesla gives no Robotaxi fleet or ride count, Optimus units or FSD revenue, and tesla.com did not load, so I took no FSD price. Waymo’s own site gives no fleet size, so the fleet figure is a news report; U.S. share and storage figures are also third party. I found no primary source for Tesla’s share of the Chinese market and no revenue consensus; Nasdaq’s earnings consensus does not state its basis.³¹⁴⁶

What would move me

What would raise my base value: a reported autonomy revenue line that tracks my path of $4B in year three and $14B in year five, vehicle gross margin without credits above 21%, deliveries above 2.0M, and free cash flow above zero with capital spending at or under $20B. What would lower it: no ride or fleet figure by Q3 2028, vehicle margin below 15%, capital spending above $25B in 2027, or new pledges of Musk’s shares.E (estimate)⁴

The next tests come in December 2026, when the annual meeting is held on 3 December and the SpaceX sale restrictions end, and in early January 2027, when I expect Tesla to report Q4 deliveries; it reported Q4 2025 deliveries on 2 January 2026.³⁸¹⁸