Risk / Bear Report — Tesla

Provenance. Written on 2026-07-30 by the seat holder against the Q2 2026 record — not imported from an earlier session. This slot previously read "awaiting import of the original Claude report"; there was no original. Every figure below is either recorded in financials.json or is arithmetic on those figures, shown so it can be checked. No new facts about Tesla are introduced.

The strongest case against

Not "Tesla is overvalued" — this seat has no valuation model and will not pretend to one. The strongest available case is narrower and worse:

The thesis rests most heavily on the engine with the least evidence, and the gap is not closing on any disclosed measure.

Everything the institution can verify sits in the businesses the thesis treats as the funding source. Everything it treats as the payoff is undisclosed. That asymmetry is the whole bear case.

What the income statement actually says

Operating income was $398M on $28.236B of revenue — a 1.4% operating margin, at a company whose thesis is a high-margin autonomy platform.

Three pieces of arithmetic on recorded figures:

• Capex is 14.5× operating income. $5.789B against $398M in the same quarter. The company spends more than fourteen times what it earns from operations, every quarter.
• The Q2 run rate does not reach guidance. $5.789B × 4 = $23.2B against full-year guidance above $25B, so the back half is heavier. The spending is still accelerating.
• Free cash flow is negative $1.092B while that acceleration is guided rather than merely observed.

Where I will not overstate the case

The balance sheet is genuinely strong, and a bear report that skipped this would be the weak-objection failure this seat exists to avoid.

$43.524B in cash and investments against $5.789B of quarterly capex is roughly seven and a half quarters of the entire capital programme funded from the balance sheet alone, before any operating contribution. Tesla is not running out of money. Any bear case built on imminent liquidity stress is wrong on the recorded numbers.

The real risk is not insolvency. It is spending this much, for this long, with nothing disclosed to show for it — and the six kill criteria are correctly aimed at exactly that.

The four things that would make this seat wrong

Stated in advance, so the record shows what would change my view:

1. Robotaxi paid miles or revenue disclosed as a separate line.
2. Energy segment gross margin disclosed, sizing the second funding engine.
3. Operating margin recovering above 3% while capex guidance holds.
4. FSD subscriptions reported against eligible fleet size, making the attach trend checkable rather than inferred.

Each is a disclosure event, not a product event. Worth noticing on its own: this thesis turns on what gets reported, and none of the four requires Tesla to do anything it is not already claiming to do.

What this seat could not assess, and why

• Whether the capital programme earns a return. No return on incremental capital has been derived. Consumption is recorded; productivity is not. Until the outcome arrives, a good investment cycle and a bad one look identical from here.
• Whether margins are cyclical or structural. One quarter is recorded. No trend can be established from this repository in either direction — which cuts against the bear case as much as for it.
• Governance and key-person risk. Named in the load-bearing assumptions, carried on judgement, no structured evidence recorded. Flagged as unassessed rather than asserted.

Recommendation

Consistent with Hold, and for the bear reason rather than the bull one: not enough disclosed economics on the load-bearing engine to support increased conviction, and not enough deterioration in the funding engines to support exit.

What this seat argues for is not patience. It is waiting on four specific disclosures, against deadlines already recorded in the kill criteria.

Corrections to this seat's earlier reasoning

Recorded rather than silently edited, because a risk report that quietly revises itself is not a record.

1. A Tesla–SpaceX transaction would more likely involve SpaceX as acquirer, not Tesla. The earlier framing had the direction backwards. It also remains speculative: no filing supports any combination, and the Evidence Officer classes it tier 3.

2. Recourse debt was effectively zero at Q2 2026 — $2M, against $9.059B non-recourse. Leverage is not a near-term risk, and any bear case leaning on it is wrong on the filing.

3. Success-contingent compensation dilution must be separated from destructive distressed financing. Ordinary employee compensation and success-contingent executive awards are costs of a business that is working; a distressed raise is the opposite. Collapsing all three into one dilution figure hides which is actually happening.

A correction to this seat's own flag

I previously recorded the share counts as a probable transcription error because 3.540B diluted sat below 3.9495B outstanding. That was my mistake, not the filing's: the two are different measures. Diluted weighted average is a period average over April to June; shares outstanding is a point-in-time count on July 16. A snapshot after quarter end exceeding an average across it is ordinary.

The gap is therefore a finding, not an error — roughly 410M shares issued late in Q2 or early in July. That is the only directly measurable dilution signal in the record, and it is now carried as a risk at medium likelihood rather than unknown.

What remains open is attribution. Ordinary compensation, success- contingent executive awards and acquisition consideration would each imply a different thesis, which is exactly why this seat's earlier correction insisted they not be collapsed into one number.

What H1 changes about the cash argument

The first version of this report was written before H1 figures were recorded, and they soften one claim. H1 operating cash flow of $8.634B against $8.282B of capex is positive for the half, while Q2 alone is negative. Free cash flow turned within H1 rather than having been negative throughout.

The argument survives — capex is still accelerating toward guidance above $25B, and the load-bearing engine still discloses nothing — but the starting point is a business that was covering its investment programme from operations two quarters ago and has just stopped. That is a different, and more informative, fact than persistent cash burn.