TSLA delivered a record 480,126 vehicles last quarter and crossed $100B in trailing revenue for the first time in its history. Revenue of $28.2B beat consensus by about 7%, the stock still fell roughly 4% through the evening, and it kept fading during the earnings call rather than stabilizing on it. This was the quarter where the two most concrete legs of the bull case, auto margins recovering, energy margins carrying the mix, both moved the wrong way at the same time, and management quietly told you the funding picture is changing too. Start with the headline, revenue beat by ~$1.8B, and non-GAAP EPS of $0.33 missed the $0.53 consensus by 38%. Operating income fell 57% YoY to $398M. That’s a 1.4% operating margin at a company trading around 86x trailing EBITDA. GAAP net income of $1.1B looks respectable until you read the reconciliation. Roughly $1.0B of it is a mark-to-market gain on Tesla’s SpaceX stake, plus a $274M one-time tax benefit. Strip those out and the quarter’s actual operating economics deteriorated on both of the revenue lines that matter.The auto side is the familiar problem, now worse. Auto gross margin ex-credits fell from 19.2% to 16.3%, below the 17% level that I’d flagged as the pain threshold going into the print. The CFO’s explanation was that adjusting for a ~$230M one-time benefit in Q1, margins were approximately flat. Except $230M against Q1 auto revenue is about 140bps, and the decline was 290bps. Roughly half the move is unexplained, and the most likely culprit, rising interest-rate subvention costs, which management flagged but didn’t quantify, is a recurring headwind, not a comp effect. Regulatory credits, the old margin cushion, fell 67% YoY to $146M and are no longer big enough to matter.The energy story took the bigger hitThe auto margin miss stings, but the energy guide-down is the more important development, because energy was the bull case’s most defensible pillar. This was the segment with 30%+ gross margins, a real backlog, and management’s best guidance track record — the tangible bridge that made the stock less binary than a pure robotaxi bet. Energy gross margin printed 20.4%, down from 39.5% a quarter ago. Volume was fine (13.5 GWh deployed, +53% QoQ). But the CFO then said the segment should “normalize at a gross margin rate in the mid to low 20% range” long-term. That’s a structural reset, delivered in one sentence, on the segment most models had penciled in at 30%. If energy settles at 22–24% while growing toward a fifth of revenue, the blended margin recovery story that carried a lot of 2025-vintage bull math simply doesn’t work as written.A $30B credit line is a tellTucked into prepared remarks, Tesla has arranged debt facilities giving it capacity to borrow up to $30B. This from a company that has funded itself from operating cash flow and a fortress balance sheet for years. Read it alongside the rest of the quarter with FCF negative $1.1B, Capex more than doubled sequentially to $5.8B, the full-year guide is north of $25B, and management says capex grows for another two to three years. Musk, asked about pacing, said it’s “okay to be a little less capital efficient if we get things done sooner.” You don’t pre-position $30B of borrowing capacity if you expect internal cash flow to cover the plan. Management is telling you, fairly plainly, that the capital-consumption phase is bigger and longer than the spring framing implied. Nobody on the call pushed back. Which brings up the funiest detail of the night, across six sell-side analysts, there wasn’t a single question about auto margins, energy margins, or the P&L at all. Every question was Optimus, chips, robotaxi regulation, or SpaceX. A 38% EPS miss went unexamined on its own earnings call but that’s the audience Tesla has and it’s also why the stock can stay disconnected from its financials longer than any short would like.What the price still assumesHere’s the math that keeps me on the bear side, at ~$350 after hours, Tesla trades at roughly 86x trailing EBITDA, about 3x GOOGL at 27x. Alphabet is the comp that matters, because it owns Waymo, the actual scale leader in commercial autonomy: roughly 500K paid rides per week across 10 metros. Tesla’s disclosure tonight was 380K unsupervised miles, cumulative, lifetime. The company asking you to pay the autonomy premium is one to two orders of magnitude behind the company that doesn’t. Run the hold-for-three-years math and the problem gets specific. On a reasonable FY28 EBITDA path (~$26B, which already gives credit for decent growth), earning 10% a year from here requires the market to still pay ~67x at exit. A defensible multiple for a business with this margin profile and this growth, even granting a premium to Alphabet and Nvidia for optionality, is more like 30–45x. That range implies a stock between roughly $217 and $320. You’re not only underwriting execution, you’re underwriting the market’s willingness to keep paying a multiple no comparable business receives, for years, while the disclosed fundamentals go the other way. Robotaxi economics could change all of this, which is exactly the problem, disclosure is a corporate decision, not a data series. Tesla will show those numbers when they’re favorable and not before. Until then, there’s nothing to converge on.Where I landBearish, with honest caveats about sizing. This is a stock with a long history of punishing valuation shorts, a heavy retail and index bid, and sentiment that swings on Musk commentary independent of the numbers. The expected value here is more poor risk/reward for longs than screaming short. But the direction of the evidence this quarter was one-way. Both profit engines got structurally less profitable, the company started lining up outside financing for a capex cycle it describes as accelerating, Optimus slipped from a dated commitment (”late July/August”) to “later this year” — announced, without apparent irony, on July 22, and the one disclosure that could justify the premium stayed undisclosed. Key factors to watch going forward are whether Q3 auto gross margin ex-credits stabilizes above 17% clean, whether energy lands in the new “mid-to-low 20s” or undershoots it, whether FY26 consensus (still around $2.00 in EPS, versus $0.74 earned in the first half) gets cut 10–20% as I expect over the next couple of prints, and any robotaxi number with a dollar sign attached.