Robotaxi Launched, But Musk Just Warned of 'A Few Rough Quarters' as the EV-Credit Cliff Looms
Key Takeaways
- Second consecutive YoY revenue decline — total $22.50B (-12%), auto $16.7B (-16%) — and a clean miss vs. consensus on both lines (revenue $22.50B vs. $22.74B; adj. EPS $0.40 vs. $0.43). Auto regulatory credits halved YoY ($439M from $890M) and the "Big Beautiful Bill" zeros out the emission penalties that drive that revenue stream entirely going forward.
- The $7,500 federal EV tax credit dies September 30, 2025 — and CFO Taneja said Tesla "may not be able to guarantee delivery for orders placed in the later part of August and beyond." That is a Q3 demand pull-forward immediately followed by a Q4 air pocket the company has not quantified.
- Robotaxi went commercial in Austin: ~7,000 miles driven, "no notable safety-critical interventions" per Ashok Elluswamy. But the fleet is still "a handful of vehicles," human safety riders are on board, and the Bay Area expansion will launch with a person in the driver's seat. The technology is real; the commercial business is still vapor.
- Musk himself: "We probably could have a few rough quarters. I'm not saying that we will, but we could… Q4, Q1, maybe Q2." Combined with Adam Jonas's pointed exchange about Musk's 13% control stake (Musk: "that is a major concern for me"), this was the most defensive Tesla earnings call we've seen in years.
- Rating: Maintaining Underperform. Q1's auto-loss-ex-credits read is now confirmed by a second consecutive YoY revenue decline, and the EV-credit cliff guarantees a worse Q4 setup. The Robotaxi launch is a real milestone but is not yet a financial story; the multiple still depends on optionality the company itself just told us would take "until end of next year" to matter.
Results vs. Consensus
| Metric | Actual | Consensus | Beat/Miss | Magnitude |
|---|---|---|---|---|
| Revenue | $22.50B | $22.74B | Miss | -1.1% |
| Adjusted EPS | $0.40 | $0.43 | Miss | -7% |
| GAAP EPS | $0.33 | — | Miss | -18% YoY |
| Operating Margin | 4.1% | — | Miss | -220bps YoY |
| Net Income | $1.17B | — | Miss | -16% YoY |
| Free Cash Flow | $0.15B | — | Miss | ~85% below mid-cycle |
| Auto Revenue | $16.7B | — | Miss | -16% YoY |
| Energy Revenue | ~$2.8B | — | Beat | Record GP |
| Reg. Credit Revenue | $0.44B | — | Miss | -51% YoY |
Quality of Beat/Miss
- Revenue: Headline miss is shallow (-1.1%), but the composition is bad. Auto -16% YoY on a 14% delivery decline (deliveries 384,122). The 19% sequential auto revenue uplift is the new Model Y rampraising ASPs from a depressed Q1 baseline — a recovery, not a return to growth.
- Margins: 4.1% operating margin is a sequential improvement from Q1's 2.1% on better fixed-cost absorption and ASP mix, but tariff costs already added ~$300M of incremental burden in Q2 (two-thirds in auto), with management explicitly flagging that "the full impacts will come through in the following quarters." Reg credits collapsed 51% YoY and will collapse further as the BBB zeros out emission penalties.
- EPS: Adjusted EPS includes a $284M Bitcoin (Redcom Holdings) mark-to-market gain in other income. Strip that out and underlying operating EPS is materially lower — the auto P&L is structurally weaker than the headline implies.
Segment Performance
| Segment | Revenue | YoY Growth | Notable |
|---|---|---|---|
| Automotive | $16.7B | -16% | Q2 deliveries 384,122 (-14% YoY); reg credits halved to $439M |
| Energy Generation & Storage | ~$2.8B | +~3% (sequential power deployment record) | Highest-ever segment gross profit; tariffs absorbed by industrial customers |
| Services & Other | ~$3.0B | +17% segment GP | Supercharging led — 7,377 stations, +18% YoY stalls (~2,900 new) |
| Reg. Credits (within auto) | $0.44B | -51% | BBB zeros emission penalties → this stream goes to ~zero over time |
Automotive — Sequential Recovery, Structural Decline
The new Model Y rampprovided a real sequential uplift (+19% in auto revenue on +14% in deliveries), proving that the changeover excuse from Q1 was at least partially valid. But year-on-year, deliveries are still down 14% and auto revenue is down 16%. Tesla is now a company whose best-selling product is in late-cycle decline globally — Model Y was best-selling in Turkey, Netherlands, Switzerland, and Austria in June, but EU and China share dynamics tell a different story than that selective list. The cheaper model began first builds in June; CFO conceded the Q3 ramp will be "slower than initially expected" because Tesla is prioritizing existing-model deliveries to capture the expiring tax credit.
"If you are in the US and looking to buy a car, let's roll now as we may not be able to guarantee delivery for orders placed in the later part of August and beyond." — Vaibhav Taneja, CFO
Assessment: The CFO is openly telling shoppers to order now or miss out. That's an unusual tactic for a company supposedly seeing strong organic demand — it's a tacit admission that the only thing supporting Q3 is the credit cliff itself.
Energy & Storage — Carrying the Story
Energy posted record power deployments and the highest gross profit in segment history. The AI-data-center storage demand thesis is becoming more concrete: hyperscalers buying Megapack at scale to firm grid output. Industrial customers are absorbing tariffs; residential storage faces an early-2025 consumer-credit expiration that will pressure that subsegment specifically. Tesla is opening its first US LFP cell facility by year-end and a third Megafactory near Houston in 2026.
Assessment: Energy is the cleanest part of this story and is actually accelerating into the AI capex cycle. It is still ~13% of revenue — too small to carry consolidated growth, but a real call option on data-center capex.
Key Topics & Management Commentary
Overall Management Tone: A different call than Q1's defensive posture — Musk is back to selling the autonomy vision, but the CFO is pre-warning Q3/Q4 demand mechanics with unusual specificity. Notably, Musk made an extraordinary admission about his own control stake in response to Adam Jonas's question. The contrast between Musk's "most valuable company in the world" rhetoric and Taneja's quarter-by-quarter caution is the widest we've seen.
Robotaxi — Real Miles, Tiny Fleet
Tesla launched paid driverless rides in Austin during the quarter and has logged over 7,000 autonomous miles with no notable safety-critical interventions. The service area has expanded once and is targeting 10x the original geofence. The Bay Area, Nevada, Arizona, and Florida are next, but the Bay Area launch will operate with a human driver while Tesla awaits regulatory approval — a meaningful walk-back from the "no one in the driver's seat" framing.
"We have more than seven thousand miles operating in the Austin area… So far, there are no notable safety-critical interventions." — Ashok Elluswamy, VP Vehicle Software
"I think we'll probably have autonomous ride-hailing in probably half the population of the US by the end of the year. That's at least our goal, subject to regulatory approvals." — Elon Musk, CEO
Assessment: 7,000 miles with zero safety-critical interventions is a real number. But "half the US population by year-end" is the kind of aspirational target Tesla has consistently missed. Musk's own framing — "subject to regulatory approvals" — is the operative qualifier. We model a US Robotaxi presence in 5–10 metros by end-2025, mostly with safety drivers, contributing immaterially to revenue.
The EV Tax Credit Cliff — Q4 Air Pocket Pre-Wired
The most important number the company didn't put in the deck: how big is the Q3 pull-forward, and how big is the Q4 hole? Taneja flagged the supply constraint already; demand was strong enough to warrant pulling back incentives mid-quarter. Q3 deliveries will likely surge; Q4 will mechanically suffer. Tesla has not given a Q4 unit guide, but our base case is Q4 deliveries down 15–25% sequentially.
"A Few Rough Quarters" — Musk's Tone Pivot
"We're in this weird transition period where we will lose a lot of incentives in the US… Does that mean we could have a few rough quarters? Yeah. We probably could have a few rough quarters. I'm not saying that we will, but we could. Q4, Q1, maybe Q2." — Elon Musk, CEO
Assessment: This is the most explicit downside warning Musk has given on a Tesla earnings call in years. Naming three specific quarters of weakness is a meaningful guide-down by tone if not by number. Buy-side will read this as "Q4 25 → Q2 26 are essentially rebuilding quarters."
Musk's Control Stake — The 13% Question
Adam Jonas asked directly whether Musk is comfortable taking Tesla into "physical AI" with only a 13% stake. Musk's response was unusually candid:
"That is a major concern for me, as I've mentioned in the past. I hope that is addressed at the upcoming shareholders meeting… I want to find that I've got so little control that I can easily be ousted by activist shareholders after having built this army of humanoid robots." — Elon Musk, CEO
Assessment: This previews a renewed CEO compensation/control proposal at the next shareholder meeting. Investors should expect a request for super-voting shares or a dilutive equity grant — both of which carry governance and dilution risk. The framing also raises the probability of Musk threatening to take humanoid-robot R&D outside Tesla if the meeting doesn't deliver him more control.
Optimus 3 + Dojo 2 — Pushed Out, Not Down
Optimus 3 prototypes are now "in three months" with production ramp early next year and a five-year aspiration of 100,000 units/month (vs. earlier "millions per year by 2029" framing — quietly a more conservative shape). Dojo 2 is now "next year at scale," and the AI factory is "end of next year." These are all multi-year, not modelable for 2025–2026.
Hardware 3 — Still Punted
The HW3 retrofit question came up again. Taneja's answer:
"What we want to do is get unsupervised done on hardware four first. Once it's done, then we'll go back and look at what we need to do with the hardware three cars." — Vaibhav Taneja, CFO
Assessment: Translation: HW3 customers — millions of them — will likely receive some form of upgrade, refund, or feature accommodation, and Tesla has not yet quantified or accrued for it. This remains a deferred multi-billion-dollar liability.
Guidance & Outlook
| Metric | Prior (Q1 Call) | Updated (Q2 Call) | Change |
|---|---|---|---|
| FY25 Vehicle Volumes | "Will revisit in Q2" | Still no formal volume guide; Q3 supply-constrained, Q4–Q2 26 "rough" | Effectively guided down |
| FY25 Capex | ">$10B" | ">$9B" | Slight reduction (timing) |
| Robotaxi scope | 10–20 cars Austin June | "Half US population by year-end" (subject to regulators) | Aspirationally raised; concretely tiny |
| Cheaper model SOP | "June, slower ramp" | First builds June, ramp Q3 "slower than initially expected" | Slipped further |
| Optimus 100K/mo target | "5 years" | "5 years" (~Q3 2030); prototypes in 3 months | Maintained (timeline preserved) |
| Reg credits trajectory | Steady | "Will impact total revenues quite fast" (BBB zero penalties) | Materially worse |
Implied Q-over-Q ramp: If Q3 is a US-supply-constrained surge, expect deliveries roughly flat to up modestly sequentially. Q4 then steps down 15–25% sequentially as US demand resets post-credit. FY25 deliveries now look like ~1.6–1.7M, down 5–10% YoY — well below where consensus stood entering 2025.
Street at: Pre-print FY25 EPS consensus was around $1.85–$2.00. We expect estimates to come down another 10–15% to reflect the Q4 pull-forward dynamic and the reg-credit rolloff.
Guidance style: Tesla is now guiding by tone rather than by number. CFO Taneja issuing a delivery-guarantee disclaimer and Musk verbally listing Q4/Q1/Q2 as "rough" is functionally a soft pre-announce.
Analyst Q&A Highlights
Robotaxi Economics & Scale
- Emmanuel Rosner, Wolfe: Asked for Robotaxi KPIs. Tesla disclosed 7,000+ Austin miles, no notable safety-critical interventions, "a handful" of vehicles. Musk added that CyberCab could deliver sub-30¢/mile economics over time and pegged "material impact on financials around the end of next year."
Assessment: 7,000 miles is a real first datapoint, but is six orders of magnitude below the miles needed to underwrite the autonomy thesis. End of 2026 is the earliest Tesla itself thinks Robotaxi moves the P&L. - Dan Levy, Barclays: Asked when third-party-owned Teslas can join the network. Musk: "Confidently next year." Taneja added Tesla will impose "Airbnb-like" criteria for vehicle eligibility.
Assessment: A 2026 third-party fleet is plausible only if Tesla's first-party fleet is operational at scale by then — itself unproven.
Pricing & Affordability
- Mark Delaney, Goldman Sachs: Asked whether Tesla will use price cuts to drive volume now that the IRA credit is going away. Musk gave the "few rough quarters" answer.
Assessment: Translation — Tesla doesn't think price cuts will solve this; the company is bracing for unit decline rather than margin destruction. - Will Stein, Truist: Asked for specifics on the lower-cost model architecture. Tesla refused to disclose the configuration, only confirmed it is built on existing Model Y/3 lines. The "true sub-$30K Tesla on a new platform" thesis is dead.
Assessment: Confirms the read from Q1 — this is a stripped Model Y, not a new vehicle.
Governance & Capital Structure
- Adam Jonas, Morgan Stanley: Asked about Musk's 13% control stake. Musk conceded it's a "major concern." This was the most consequential exchange of the call for shareholders, foreshadowing a meeting-time dilutive grant or structural change.
Assessment: The market has not yet priced in a renewed Musk compensation package. Watch the upcoming shareholder meeting closely.
FSD Adoption
- Mark Delaney, Goldman Sachs: Asked about FSD subscription trends. Taneja disclosed FSD subscription rates up 45% since v12, overall penetration up 25% since v12 + B38. Useful first quantitative read on FSD ARR — but no absolute revenue number.
Assessment: 45% increase off an undisclosed base is good directionally but tells us nothing about absolute scale of the recurring revenue stream that's supposed to underwrite the autonomy multiple.
What They're NOT Saying
- FY25 delivery number, again: Tesla still has not provided a 2025 vehicle volume guide. Two consecutive quarters of refusal to anchor the year.
- Q4 demand modeling: No quantification of the post-credit demand cliff. Investors are left to model it themselves.
- Robotaxi unit economics: No $/mile figure for current Austin operations. Musk gave aspirational CyberCab numbers but nothing on actual current-fleet economics.
- FSD absolute ARR: 45% growth is great; growth off what? Tesla still won't disclose dollar revenue from FSD subscriptions.
- HW3 retrofit accrual: Still no balance-sheet recognition of the future obligation to HW3 customers.
- XAI investment color: Taneja explicitly refused to discuss Tesla's stake/role in XAI. "This is not the forum to discuss this topic." That refusal alone is informative.
- FY26 capex: No early framing for 2026 capital intensity, despite confirmed Optimus production start, Dojo 2 ramp, AI factory build, and third Megafactory in Houston. The 2026 capex number could be eye-watering.
Market Reaction
- YTD context: TSLA down ~18% YTD into the print, worst-performing megacap.
- After-hours move: Initially flat, then -4%+ during the call as the CFO described the EV-credit-cliff supply constraint and Musk delivered the "few rough quarters" line.
- Implied next-day open: Down 4–6% from the closing print.
- Volume: Elevated, consistent with a guidance-shock event.
The reaction is rational. Tesla beat-or-missed by a hair on EPS and revenue but verbally guided Q4 25 → Q2 26 to "rough." That sequence — Q3 surge, Q4 cliff, Q1/Q2 26 rebuild — implies a 12-month NTM EPS that is materially below where consensus sat. The Robotaxi launch is the bull anchor, but with 7,000 miles and a handful of vehicles, it cannot offset the near-term P&L deterioration in the eyes of the marginal holder.
Street Perspective
Debate: Does the Q3 Pull-Forward Help or Hurt the Stock?
Bull view: Q3 will print a strong delivery surge, possibly back to ~470K+ units, providing positive headlines into earnings season and a temporary multiple support.
Bear view: The market will trade the cliff, not the surge. By the time Q3 prints in late October, Q4 demand math will be the dominant narrative.
Our take: Bear view wins. Tesla bulls have pre-announced the Q3 pull-forward thesis already; the marginal incremental information will be the post-credit demand response, not the pre-credit surge. The market trades 6 months forward.
Debate: Is Robotaxi the Real Bull Case Now?
Bull view: 7,000 miles with zero safety-critical interventions, scaling to "half the US population by year-end" → this is the inflection. The auto P&L is irrelevant; just the optionality matters.
Bear view: A handful of vehicles in Austin with safety drivers in the Bay Area is a pilot program, not a business. The "half the US population" framing was qualified with "subject to regulatory approvals" — i.e., we don't know.
Our take: Robotaxi is real and is no longer a pure pitch deck. But the gap between "real but tiny" and "moves the financial needle" is what Musk himself called "end of next year." That is too long to underwrite at current multiples without seeing scaling data first.
Debate: Will the Shareholder Meeting Approve Major Compensation Changes for Musk?
Bull view: Aligning Musk with super-voting control or a fresh equity package locks in the AI/robotics roadmap. Investors should pay up for that certainty.
Bear view: A renewed mega-grant is straight dilution and validates a governance pattern shareholders have already pushed back on (see the Delaware Chancery decision on the prior package). It also signals Tesla cannot retain Musk's full attention without paying him again.
Our take: Whatever the meeting produces will likely involve dilution. Position size should reflect that overhang.
Model Update Needed
| Item | Prior (Post-Q1) | Suggested Change | Reason |
|---|---|---|---|
| FY25 Deliveries | 1.65–1.75M | 1.60–1.70M | Q3 pull-forward but Q4 air pocket cancels |
| FY25 Auto Gross Margin (ex-credits) | ~12–14% | ~13–15% | New Model Y mix lifts ASPs; partially offset by tariff $300M+/qtr drag |
| FY25 Reg Credit Revenue | ~$2.0B | ~$1.7B | Q2 came in $439M (annualized ~$1.7B); BBB accelerates declines into 2026 |
| FY25 Adj. EPS | $1.80–$2.20 | $1.65–$1.95 | Reg-credit rolloff + tariff drag + Q4 demand reset |
| FY26 Adj. EPS | ~$2.50–$3.00 | $2.10–$2.50 | Reg credits trending toward zero; "few rough quarters" extends into Q2 26 |
| FY25 FCF | $0–$1B | $0–$0.5B | $146M Q2 + lower Q4 = barely positive year |
| FY25 Robotaxi revenue | ~$0–$50M | ~$25M (still trivial) | Austin paid rides + tiny SF Bay launch |
Valuation impact: At ~$305 pre-print, TSLA is being asked to underwrite ~120x our $2.50 FY26 midpoint. The 12-month forward picture has gotten worse, not better, since Q1. Robotaxi optionality has gone from "vapor" to "pilot," but the auto P&L has gone from "Q1 pothole" to "structurally declining with reg-credit tailwind disappearing." The risk/reward still skews negative.
Thesis Scorecard Post-Earnings
| Thesis Point | Status | Notes |
|---|---|---|
| Bull #1: Auto volume returns to growth in 2025 | Challenged | Two YoY decline quarters; Q4 cliff incoming |
| Bull #2: Robotaxi launches and scales | Confirmed (launched), Neutral (scale) | Real launch but 7,000 miles and a handful of vehicles |
| Bull #3: Energy storage scales | Confirmed | Record GP, AI demand thesis intact |
| Bull #4: FSD ARR materially scales | Confirmed (rate), Neutral (size) | Subscriptions +45% since v12 — but base undisclosed |
| Bear #1: Brand damage hurts demand | Confirmed | Auto -16% YoY; Musk "rough quarters" admission |
| Bear #2: Margin compression sustained | Confirmed | Reg credits halved; tariffs +$300M; BBB removes future credit revenue |
| Bear #3: Robotaxi slips again | Neutral | Launched on time; commercial scale TBD |
| Bear #4 (NEW): Musk control / dilution overhang | Confirmed | Musk himself flagged "major concern"; meeting-time governance event likely |
Overall: Thesis remains tilted bearish. Three of four bull pillars are partial confirmations at best (Robotaxi launched but tiny; FSD growing but undisclosed in absolute terms). Three of four bear pillars are now confirmed by management.
Action: Maintain Underperform. The Q3 print will deliver a delivery surge that should not be confused with a thesis change — it's a credit-driven pull-forward. The clearer test is Q4: post-credit demand response, Robotaxi miles trajectory, shareholder-meeting outcome. Until at least two of those three break clearly bullish, the risk/reward continues to skew negative.