LARAY.AI

TSLA · Hypothetical workbench

Everything short of a valuation

Third-party models reproduced in their own units and horizons, an explicit record of what each cannot establish, and Laray’s own scenario assembly where one exists. This is not a Laray valuation and carries no decision.

Two things on this page are Laray’s own work rather than someone else’s: the scenario assembly and the equity bridge. Both are proposals awaiting owner decisions, and neither produces a fair value, a probability, or a per-share figure.

The canonical record lives at the valuation record.

Standing

Draft

This is a workbench, not a valuation. It records what third-party models say in their own terms. It carries no Laray fair value, no scenario probabilities, and no decision. It is blocked from becoming a canonical valuation until the gaps below are resolved by the owner.

Decision eligible
No
Probability eligible
No
Canonical promotion
No

Horizon treatment: preserve source native no silent normalization · comparison target · recorded

Source reproductions

3 independent of 4

Each model below is reproduced in its own units and its own horizon. Nothing here has been rescaled to a common date, because doing so silently would misrepresent every source at once.

ARK Tesla 2029 open model, version 8

Output Bridge reproduction · independent source model · native horizon 2029

How this source models uncertainty: normal downside upside one sigma with bucketed tails

Source artifact · version 8 · retrieved 2026-08-15 · sha256 dd1c0f025f76

Source-native inputs

Autonomous EBITDA margin
20% 80% (inner 40% 60%)
Downside and upside are ARK's minus-one-sigma and plus-one-sigma inputs, not triangular low and high values.
Tesla Valuation Inputs!B12:E12
Maximum annual vehicle production increase
10%/yr 100%/yr (inner 20%/yr 60%/yr)
Downside and upside are ARK's minus-one-sigma and plus-one-sigma inputs, with tails bucketed at minimum and maximum.
Tesla Valuation Inputs!B13:E13
Annual miles per robotaxi
30,000 miles/vehicle/year 130,000 miles/vehicle/year (inner 70,000 miles/vehicle/year 120,000 miles/vehicle/year)
Downside and upside are ARK's minus-one-sigma and plus-one-sigma inputs, with tails bucketed at minimum and maximum.
Tesla Valuation Inputs!B14:E14
Robotaxi launch year
2,025 calendar year 2,030 calendar year (inner 2,026.5 calendar year 2,025.25 calendar year)
The upside year is earlier than the downside year because earlier launch is economically favorable; the columns are not ordered quantiles.
Tesla Valuation Inputs!B15:E15
Autonomous adoption takeover time
5 years 14 years (inner 5 years 8 years)
These values retain ARK's native column labels and must not be relabeled as Laray low, mode, and high inputs.
Tesla Valuation Inputs!B16:E16

What this reproduction cannot establish

  • This is an exact output-bridge reproduction, not a reimplementation of all forty-five operating inputs and recursive spreadsheet formulas.
  • The cached 5,000-simulation outputs are one stochastic workbook state and differ slightly from the article's rounded one-million-simulation outputs.
  • ARK samples normal distributions with downside and upside as one-sigma inputs and buckets tails; the current Laray triangular sampler is not distribution-equivalent.
  • ARK's Optimus module models internal Tesla factory labor savings rather than external humanoid-robot sales.

Cern Basher Tesla valuation deep dive

Bounded Public Equation reproduction · independent source model · native horizon 2030 and 2035

How this source models uncertainty: creator reported no public distribution

What this reproduction cannot establish

  • The creator's underlying spreadsheet is unavailable, so the $150 trillion full-model output and $3,322 present value cannot be reverse-engineered without inventing missing inputs.
  • Only timestamped public equations and claims are reproduced; automotive, energy, Optimus, dilution, and discounting remain incomplete as a unified model.
  • No probability is inferred from the magnitude or specificity of the creator's bull case.

Steven Mark Ryan ARK 2025 model walkthrough

Lineage Only reproduction · derived from another source in this set, not an independent view · native horizon 2025

How this source models uncertainty: not applicable

What this reproduction cannot establish

  • The selected video explains ARK's 2025 model and is not a second independent valuation source.
  • Exact assumptions must come from the upstream ARK workbook, not from a secondary transcript.
  • A direct Steven Mark Ryan model would require separate source selection and ingestion.

InvestAnswers public Cybercab illustration

Bounded Public Equation reproduction · independent source model · native horizon ten-week illustration, 2027-2028, 2030, and 2032

How this source models uncertainty: creator reported no public distribution

What this reproduction cannot establish

  • The creator references a proprietary twelve-spreadsheet model that is not publicly available; only the bounded public illustration is reproduced.
  • The ten-week illustration does not tie exactly, and no input is altered to force a tie.
  • The $3,000 and at-least-$8,000 per-share outputs lack a complete public market-cap, dilution, and date bridge.

Recorded gaps

6

Horizon Normalization

Blocks promotion

The selected source models use 2029, 2030, 2032, and 2035 outputs while the canonical comparison date is 2036-08-15.

Directly ranking or averaging the native outputs would mix different compounding periods and operating states.

Owner decision required: Approve an explicit source-by-source roll-forward or select which native assumptions should be rebuilt directly at the 2036 horizon.

Distribution Translation

Blocks promotion

ARK uses clipped normal sampling with downside and upside as one-sigma inputs; Laray HVE currently samples triangular low, mode, and high ranges.

Copying the four ARK columns into triangular ranges would falsely claim distribution equivalence.

Owner decision required: Choose between preserving an ARK-native simulation adapter and approving a separately labeled Laray triangular translation.

Source Inconsistency

Blocks promotion

The loaded deterministic example produces about 1,842 GWh of 2029 storage deployment while the article discusses an approximately 850 GWh estimate.

Silently choosing either value would merge distinct ARK outputs and distort the energy segment.

Owner decision required: Select the relevant ARK state or direct Laray to rebuild energy from approved source-native drivers.

Source Availability

Blocks promotion

Cern's full spreadsheet and all bridge assumptions are unavailable.

The $150 trillion 2035 market cap and $3,322 present value per share cannot be fully reproduced from public inputs.

Owner decision required: Use only captured equations as modular assumptions, obtain the workbook, or retain the full-model outputs as non-promotable benchmarks.

Source Availability

Blocks promotion

The proprietary twelve-spreadsheet model is unavailable and the public ten-week illustration computes to $118 billion versus the stated approximately $100 billion.

The public segment is insufficient for a complete creator-model reproduction or per-share bridge.

Owner decision required: Approve the public equations as bounded inputs, obtain the private model, or keep InvestAnswers as benchmark-only evidence.

Owner Input

Blocks promotion

No owner-approved 2036 ranges, balance-sheet adjustments, dilution, or canonical scenario probabilities exist.

TSLA-HVE-001 cannot produce a decision, add zone, or canonical valuation run.

Owner decision required: Approve normalized operating ranges and bridges first; approve four world-state probabilities only at the later promotion gate.

Owner decisions required

5
  1. Choose whether ARK remains in its native clipped-normal simulation or receives a separately labeled Laray triangular translation.
  2. Approve how each 2029-2035 source-native assumption is rebuilt or rolled forward to 2036-08-15.
  3. Decide whether Cern and InvestAnswers missing spreadsheets limit them to modular equations and comparison benchmarks.
  4. Approve low, mode, and high operating ranges plus net cash, capital requirements, corporate costs, contingent liabilities, and dilution by category.
  5. Approve four canonical world-state mappings and probabilities only after evidence and risk review.

What still blocks promotion

6

Every item must clear before this workbench can become a canonical valuation run. None is cleared by argument alone.

  • Source-native horizons have not been normalized to the canonical ten-year date.
  • ARK's native distribution has not been preserved or explicitly translated.
  • Missing creator workbooks prevent full-model reproduction for Cern and InvestAnswers.
  • Owner-approved operating ranges, equity bridge, and dilution assumptions do not exist.
  • Canonical bear, base, bull, and exceptional probabilities do not exist.
  • Evidence, risk, human-adjustment, and owner-approval gates remain incomplete.

Scenario assembly

Proposals, not approvals

Two records assemble Tesla’s 2036 operating worlds. TSLA-HVE-003 models the engines that have filed operating anchors. TSLA-HVE-004 accounts for the rest, all at a stated zero. Together they cover every engine exactly once.

Horizon · evidence cutoff · every driver value and every threshold remains unapproved

Engines carrying a proposed figure

2

Automotive manufacturing and sales

annual vehicle deliveries × automotive revenue per delivered vehicle × automotive operating margin × enterprise multiple

Bear
2.5M × 35K × 6% × 10×
$52.5B of proposed enterprise value, awaiting owner approval
Base
4M × 40K × 10% × 15×
$240B of proposed enterprise value, awaiting owner approval
Bull
7.5M × 45K × 15% × 20×
$1.01T of proposed enterprise value, awaiting owner approval
Exceptional
12M × 50K × 20% × 25×
$3T of proposed enterprise value, awaiting owner approval

What these drivers are not

  • Automotive revenue per delivered vehicle is a scenario driver, not a literal vehicle sticker price.
  • The operating margin is not Tesla's reported gross margin and requires owner approval.
  • Robotaxi, standalone autonomy subscriptions, Energy, services, net cash, capital needs, liabilities, and dilution are excluded from this engine value.
  • The terminal enterprise multiple is a Laray proposal rather than a creator or company forecast.

Energy generation and storage

annual GWh deployed × energy revenue per GWh × energy operating margin × enterprise multiple

Bear
200 × 130M × 8% × 10×
$20.8B of proposed enterprise value, awaiting owner approval
Base
500 × 150M × 15% × 15×
$168.75B of proposed enterprise value, awaiting owner approval
Bull
1.5K × 170M × 22% × 20×
$1.12T of proposed enterprise value, awaiting owner approval
Exceptional
3K × 180M × 28% × 25×
$3.78T of proposed enterprise value, awaiting owner approval

What these drivers are not

  • Energy revenue includes generation and other mix, so the revenue-per-GWh driver is only a normalized scenario variable.
  • The operating margin is distinct from Tesla's reported segment gross margin and requires owner approval.
  • Required factories, cells, working capital, tariffs, warranty costs, net cash, liabilities, and dilution are not bridged in this engine value.
  • The terminal enterprise multiple is a Laray proposal rather than a creator or company forecast.

Engines carried at zero

4

An engine here is accounted for, not forgotten. Each carries an explicit zero with the reason behind it. Three are zero because no operating anchor exists to set a driver range; one is zero because its value already sits inside the engines above and counting it again would double-count.

Autonomy / robotaxi

Zero for want of evidence

This repository holds no operating anchor for paid autonomous transport: no paid fleet size, no utilization, no price per mile, no take rate, and no contribution margin. Every driver in the robotaxi equation would therefore have to be invented rather than sourced. Tesla's own filings describe regulatory approval as uncertain, and the one recorded outside view argues that common robotaxi valuations overstate value by sizing capacity instead of demand. Zero is what the evidence supports today, and it is a floor rather than a judgement that the engine is worthless.

What would move it off zero

  • Tesla reports paid, unsupervised robotaxi revenue as a separately identifiable figure.
  • Tesla discloses paid fleet size and utilization for a commercially operating market.
  • A regulator grants unsupervised commercial operation in a named jurisdiction without a safety driver.

Needed first: Four-world paid-fleet, utilization, price, take-rate, and contribution-margin proposals, each bound to a disclosed operating figure rather than to a capacity estimate.

Optimus

Zero for want of evidence

Tesla states in its own filing that it has yet to commercialize Bots and cannot predict demand. That is the company's position, not an outside estimate, and it forecloses every driver the Optimus equation needs: external units, price or profit per robot, operating margin, and commercialization timing. Carrying the engine at zero records the company's disclosure instead of contradicting it.

What would move it off zero

  • Tesla reports external Optimus revenue from customers outside the company.
  • Tesla discloses units produced for external sale together with a price or margin.
  • Tesla withdraws or narrows the disclosure that it cannot predict Bots demand.

Needed first: Four-world external unit, price or profit per robot, operating margin, and commercialization timing proposals, each anchored to a disclosed external sale rather than to an internal deployment target.

AI / compute / software / data

Zero to avoid double counting

This engine is carried at zero for a different reason from the others. Tesla's AI, compute, and data capability is real, but its value already sits inside the Automotive economics that this assembly models and inside the Robotaxi and Optimus economics it does not. Assigning it a separate number would count the same value a second time, which is precisely what the sum-of-the-parts boundary exists to prevent. Zero here is a double-counting control, not a statement that the capability is worthless.

What would move it off zero

  • Tesla reports external monetization of compute, models, or data to third-party customers.
  • An owner decision assigns the shared capability to exactly one standalone-value owner and removes it from the others.

Needed first: An owner decision either to keep zero standalone value or to define evidenced external revenue, its costs, its single owning engine, and the non-overlap rule that keeps it out of the others.

Services / other

Zero for want of evidence

Insurance, charging, service, and used-vehicle economics have never been separated into non-overlapping operating drivers in this repository. There is no anchor for services revenue net of what already sits inside Automotive, so any figure here would double-count the vehicle business or invent a margin. This is the weakest of the four zeros: unlike Robotaxi and Optimus, the underlying activity is already generating revenue today, so the omission understates rather than merely defers.

What would move it off zero

  • Services and other revenue is separated into components that do not overlap Automotive.
  • An operating margin is recorded for the separated services components.

Needed first: Four-world external revenue, operating margin, capital intensity, and enterprise-multiple proposals, with an explicit overlap control against Automotive.

What the two assemblies add up to

2 of 6 engines carry a figure

These are floors, not valuations. 4 of Tesla’s 6 engines contribute zero to every figure below, so each one is the least the modelled businesses could be worth on the proposed drivers — not an estimate of what the company is worth. Nothing here carries a probability, and a large magnitude is not evidence that it is likely.

Bear
$73.3B floor on 2036 enterprise value, from automotive and energy alone
Base
$408.75B floor on 2036 enterprise value, from automotive and energy alone
Bull
$2.13T floor on 2036 enterprise value, from automotive and energy alone
Exceptional
$6.78T floor on 2036 enterprise value, from automotive and energy alone

Proposed falsification gates

4 awaiting approval

Each gate names an observation that would force the drivers above to be re-underwritten. None is approved, and none is applied mechanically to the equations.

2030 annual vehicle deliveries

Warning below 2,649,054, falsified below 2,415,969 vehicles/year

Observed over: Tesla-reported full-year 2030 deliveries

Re-open or reduce the bull and exceptional 2036 Automotive delivery assumptions before any canonical promotion.

Total automotive gross margin

Warning below 0.15, falsified below 0.12 fraction

Observed over: Trailing four reported quarters at the review date

Re-underwrite Automotive operating margins and product economics; gross margin does not substitute mechanically for operating margin.

2030 annual energy storage deployments

Warning below 150.1, falsified below 120 GWh/year

Observed over: Tesla-reported full-year 2030 storage deployments

Re-open or reduce the bull and exceptional 2036 Energy deployment assumptions before any canonical promotion.

Energy generation and storage gross margin

Warning below 0.25, falsified below 0.2 fraction

Observed over: Trailing four reported quarters at the review date

Re-underwrite Energy operating margins and required capital; segment gross margin does not substitute mechanically for operating margin.

Enterprise-to-equity bridge

11 components awaiting approval

This bridge does not resolve. It converts enterprise value into equity value and then into a value per share, and it will produce none of those until every component below and a horizon share count carry an owner decision. The blocked state is the honest output, not a rendering failure.

The enterprise value it would convert is itself a floor, with four engines at zero. Even a fully approved bridge would therefore yield a floor on per-share value rather than an estimate of it.

Components read from the filing

8 cite a filed fact

Each figure below is the number as filed, with the tag it was read from, so it can be checked against the filing rather than taken on trust. Direction is stated in words because a stored magnitude does not carry its own sign.

Cash and cash equivalents

Awaiting approval

adds $16.51B

Cash held beyond the operating businesses that the engine assembly values, so it is added to enterprise value rather than counted inside any engine.

us-gaap:CashAndCashEquivalentsAtCarryingValue · 0001628280-26-003952 · as of

Short-term investments

Awaiting approval

adds $27.55B

Marketable securities are non-operating assets on the same footing as cash for this bridge. Whether the owner treats any part of them as operating working capital is a decision, not a fact.

us-gaap:ShortTermInvestments · 0001628280-26-003952 · as of

Long-term debt

Verified · awaiting approval

subtracts $6.58B

A claim ahead of equity, so it is subtracted from enterprise value.

It cannot be established from the tag alone whether LongTermDebt already includes current maturities. DebtCurrent is separately tagged at $1.569B, and DebtInstrumentCarryingAmount is $8.177B, which is $24M above the sum of the two — a gap consistent with unamortized discount or issuance costs but not proven to be. The owner must decide which figures enter the bridge and whether they overlap, rather than this record guessing at a reading.

Checked and settled. Settled against the filing on 2026-08-18. The debt note (Schedule of Debt and Finance Leases) reports Current 1,569 and Long-Term 6,584 as separate carrying amounts, and the balance sheet rolls them into two lines: current portion of debt and finance leases 1,640 (= 1,569 + 71 finance lease current) and long-term portion 6,736 (= 6,584 + 152 finance lease noncurrent). LongTermDebt is therefore the noncurrent portion only and does not include current maturities. This component and the current-debt component do not overlap.

us-gaap:LongTermDebt · 0001628280-26-003952 · as of

Current portion of debt

Verified · awaiting approval

subtracts $1.57B

Recorded separately from long-term debt because the filing reports them separately. Verified not to overlap with it.

May already be contained in the LongTermDebt figure above. Approving both without checking would double-count roughly $1.6B against equity.

Checked and settled. Settled against the filing on 2026-08-18. No overlap exists: this is the current portion and LongTermDebt is the noncurrent portion of the same debt. Together with finance leases of 223, the three components sum to 8,376, exactly the 1,640 + 6,736 the balance sheet presents. The earlier concern that DebtInstrumentCarryingAmount at 8,177 sat above their sum is also resolved: 8,177 is the Unpaid Principal Balance despite the tag name, and exceeds the 8,153 carrying amount by 24 of unamortized discount and issuance costs.

us-gaap:DebtCurrent · 0001628280-26-003952 · as of

Finance lease liabilities

Awaiting approval

subtracts $223M

Debt-like obligations, current and non-current combined, subtracted on the same basis as debt.

us-gaap:FinanceLeaseLiability · 0001628280-26-003952 · as of

Noncontrolling interests in subsidiaries

Awaiting approval

subtracts $670M

Value inside the consolidated businesses that belongs to other parties, so it is removed before arriving at value attributable to common shareholders.

us-gaap:MinorityInterest · 0001628280-26-003952 · as of

Redeemable noncontrolling interests

Awaiting approval

subtracts $58M

Carried between liabilities and equity on the balance sheet and separately tagged. Included so the bridge accounts for the full gap between enterprise value and common equity.

us-gaap:RedeemableNoncontrollingInterestEquityCarryingAmount · 0001628280-26-003952 · as of

Operating lease liabilities

Awaiting approval

subtracts $6.34B

Proposed for an explicit owner decision rather than assumed. Whether operating leases are debt-like or already reflected in the operating margins the engine assembly uses is a modelling choice; subtracting them while the margins already bear the rent would double-count them.

Whether this belongs in the bridge at all depends on how the assembly's operating margins treat lease expense, which has not been established.

us-gaap:OperatingLeaseLiability · 0001628280-26-003952 · as of

Cumulative capital required to reach the horizon

Held at zero · anchored

subtracts $0

Held at zero pending an owner projection, with the filed base rate attached. Tesla spent $8.527B on property and equipment in 2025 and guides to roughly $20B in 2026. A ten-year cumulative figure depends on the growth assumed in each world, so it is a judgement — but it is now a judgement anchored to what the company actually spends.

Filed anchors

2025 capital expenditure
$8.53B
us-gaap:PaymentsToAcquirePropertyPlantAndEquipment · as of

Corporate costs not assigned to an engine

Held at zero · anchored

subtracts $0

Held at zero pending an owner projection, with the filed base rate attached. Tesla reported $5.834B of selling, general and administrative expense and $6.411B of research and development in 2025 — $12.245B of cost sitting above the segments the assembly values. Projecting that to 2036 is a judgement; finding it is not, and it is no longer an open research task.

Whether these costs are already inside the assembly's per-engine operating margins is unestablished. If those margins are segment-level, corporate cost sits above them and belongs here; if they are company-level, subtracting it again would double-count. The same question as operating leases, and it is the owner's to settle.

Filed anchors

R&D
$6.41B
us-gaap:ResearchAndDevelopmentExpense · as of
Total operating expenses
$12.74B
us-gaap:OperatingExpenses · as of

Contingent liabilities

Held at zero · anchored

subtracts $0

Held at zero pending an owner decision, with the filed exposures attached. Tesla carries an $8.607B standard product warranty accrual and discloses $3.450B of maximum guarantee exposure. These are real and sized in the filing; what is unresolved is whether either belongs in this bridge at all.

Warranty cost is an operating expense already borne by the margins the assembly uses, so subtracting the accrual as well would count it twice. Maximum guarantee exposure is a ceiling, not an expected loss, and subtracting a ceiling would overstate. Both need an owner ruling rather than a mechanical subtraction.

Filed anchors

Warranty accrual
$8.61B
us-gaap:StandardProductWarrantyAccrual · as of
Guarantee ceiling
$3.45B
us-gaap:GuaranteeObligationsMaximumExposure · as of

Share basis

No horizon count approved

3,752,431,984 shares outstanding as of . That is today’s count, not the horizon count. Using one as the other would silently assume no dilution across ten years.

Dilution to be decided, by kind

Ordinary issuance
Routine equity compensation and ordinary-course issuance over ten years, with the filed base rate attached. Tesla's 2025 weighted-average share count was 3.225B basic against 3.528B diluted — a 303M share overhang, 9.4% of basic, already outstanding as awards. Share-based compensation ran $2.825B for the year. Assuming zero ordinary dilution would contradict the filing.

Filed anchors

Basic weighted shares
3,225,000,000 shares
us-gaap:WeightedAverageNumberOfSharesOutstandingBasic · as of
Diluted weighted shares
3,528,000,000 shares
us-gaap:WeightedAverageNumberOfDilutedSharesOutstanding · as of
Share-based compensation
$2.83B
us-gaap:ShareBasedCompensation · as of
Success-contingent issuance
Awards that vest only in the worlds where the operating outcomes are achieved. This dilution is correlated with the bull and exceptional worlds and must not be applied flat across all four.
Distressed issuance
Equity raised on unfavourable terms, which is a bear-world event. Applying it evenly would penalise the worlds where it would not happen.
Acquisition issuance
Shares issued as acquisition consideration, which also brings in acquired earnings. Counting the dilution without the acquired economics would be one-sided.

Decisions this is waiting on

7
  1. Approve or replace each filed component and its sign.
  2. Decide whether operating lease liabilities are debt-like here or already inside operating margins.
  3. Decide whether corporate costs sit above the assembly's per-engine margins or inside them, then project the $12.245B 2025 base rate to the horizon.
  4. Project cumulative capital requirements to 2036 from the $8.527B 2025 base and the ~$20B 2026 guidance, per world.
  5. Rule on whether the $8.607B warranty accrual and $3.450B guarantee ceiling belong in this bridge, given that warranty cost already sits in operating margins.
  6. Approve a horizon diluted share count by dilution category, applied per world rather than flat, starting from the filed 9.4% existing overhang.
  7. Decide whether the 2025-12-31 balance-sheet position may stand in for a 2036 position, or must be projected.

What still blocks a canonical valuation

5
  • Every bridge component is a proposal; no owner-approved signed value exists.
  • No horizon diluted share count has been approved, so no per-share value can be derived.
  • Whether operating lease liabilities, corporate costs, and the warranty accrual belong in the bridge depends on what the assembly's operating margins already bear, which is unestablished.
  • Cumulative capital requirements, corporate costs, and contingent liabilities are recorded at zero and are known to be understated; each now carries its filed base rate, but none has been projected to the horizon.
  • Balance-sheet components are dated 2025-12-31 while the assembly horizon is 2036-08-15; using a current position as a horizon position has not been approved.