BUSINESS
Goldman Keeps Tesla at $360 as AI Outruns the Cars
Goldman Sachs kept Tesla at Neutral and $360, below the share price, saying robotaxis and Optimus will move the stock more than Q3 earnings.
Goldman Sachs kept Tesla at Neutral with a $360 target, about 5% below the $378.73 share price, and said robotaxis will move the stock more than Q3 earnings. Analyst Mark Delaney held both the rating and the number on October 6, 2026, four days after Tesla beat every delivery estimate on its own Street list.
The note reads like a vote for physical AI. The $360 line reads like a refusal to pay for it yet.
Goldman Kept Neutral After the Beat
Delaney still looks for better vehicle revenue from the Model Y L ramp. He also warned that higher costs and heavier capital spending could cap 2026 earnings per share. In the same note he argued that Tesla’s physical AI work, robotaxis, Full Self-Driving, and humanoid robots, will weigh more on the shares than the October 21 print.
That is a split personality in one research note. The car business is, in Goldman’s own words, getting back on track. The target did not move. A Neutral stamp on a stock already above $360 is a hold, and the AI language does not change the arithmetic.
One widely cited compilation heading into the report looks for $0.45 of earnings per share, down 10% from a year earlier, and for revenue of $27.78 billion, down 1.1%. Delaney is telling clients those figures are not the main event. He is also telling them not to pay above $360 until the other event, the robots, shows up in the model.
HOW GOLDMAN GOT TO $360 AND STAYED THERE
- September 16, 2026: Delaney cuts his Q3 delivery forecast to 435,000 from 490,000 and keeps Neutral with a $360 target, saying sales in China, the United States, and Europe are tracking below consensus.
- September 29, 2026: Tesla posts a 24-analyst delivery compilation with a mean of 461,974 and a median of 463,406. Goldman Sachs is on the contributor list.
- October 2, 2026: Tesla reports 486,532 deliveries, above every estimate in that table.
- October 6, 2026: Goldman reiterates Neutral and $360 and says robotaxis, FSD, and humanoids will matter more than Q3 earnings per share.
- October 21, 2026: Tesla posts Q3 financial results after the close and hosts a webcast at 4:30 p.m. Central Time.
The sequence is the tell. Goldman lowered the car forecast, watched Tesla clear the bar, and still would not lift the target. Calling that bullish skips the rating.
The Delivery Print Goldman Undercounted
Tesla’s Austin release on October 2 put the quarter in a short table. The company produced 464,391 vehicles and delivered 486,532 vehicles in the third quarter, with 13.7 GWh of energy storage deployed. Model 3 and Model Y accounted for 478,237 of the deliveries. The other-models line, Cybertruck, Semi, and leftover Model S and Model X stock, added 8,295. One percent of deliveries sat under operating-lease accounting.
Tesla had published the Street’s homework three days earlier. The company-compiled average of 461,974 came from 24 firms, including Goldman, Deutsche Bank, Baird, UBS, JPMorgan, HSBC, and Cantor Fitzgerald. Actual deliveries beat that mean by 24,558 vehicles, or 5.3%. Tesla also delivered 22,141 more cars than it built, the second straight quarter of working down stock.
TESLA Q3 2026 DELIVERIES VERSUS THE FORECASTS
| Measure | Vehicles |
|---|---|
| Goldman Sachs Q3 forecast (Sept. 16) | 435,000 |
| Tesla 24-analyst average (Sept. 29) | 461,974 |
| Q3 2026 actual deliveries | 486,532 |
| Q2 2026 actual deliveries | 480,126 |
| Q3 2025 actual deliveries | 497,099 |
Against Goldman’s 435,000, the print was 51,532 vehicles higher. Against the prior quarter it was 6,406 higher, or 1.3%. Against the year-ago record of 497,099, inflated by the last rush for the U.S. federal EV tax credit, it was 10,567 lower, or 2.1%. Storage missed as well. The same Tesla compilation had looked for 15.9 GWh; Tesla deployed 13.7 GWh.
In mid-September Delaney also cut his Q4 delivery forecast to 475,000 from 515,000. The October 6 note still talks up the Model Y L as a sales prop. It does not treat the Q3 beat as a reason to move $360.
Physical AI Still Has to Print Cash
Goldman’s AI list is specific, and none of it is in the Q3 delivery table. Robotaxis, FSD, and Optimus are the three lines Delaney wants the stock to trade on. Tesla has been building the hardware stack that those products need, custom inference chips, autonomy software, foundation models for robotics, and a digital system for computer-based work, and it says the plan covers vehicles, robots, and applications.
That is also why $360 is sticky. If the multiple already prices a robotaxi network, a supervised FSD rollout, and a humanoid line, a Neutral target below the tape is a demand for evidence those projects throw off cash, not another slide deck. A $360 freeze after a 51,532-car miss on Goldman’s own delivery call is a polite way of saying the robot book still cannot be modeled.
THE AI CHECKPOINTS GOLDMAN IS WAITING ON
- Robotaxi cities: Tesla said the network was running in seven major U.S. metros by the end of the second quarter.
- Cybercab service: The purpose-built cab, no steering wheel or pedals, is in limited service in Austin.
- FSD in Europe: Croatia approved FSD (Supervised), after the Netherlands, Belgium, and Slovenia; an EU-wide vote may not come before December.
- Optimus at Fremont: Tesla is installing first-generation humanoid lines in the space freed by ending Model S and Model X production there.
In a September note after Tesla’s Austin Cybercab event, Delaney kept the same Hold-equivalent rating and the same $360 target even while he sketched a cost case. If Tesla hits a $20,000 to $30,000 Cybercab at scale, Goldman put a possible $0.05 to $0.30 per-mile cost edge against robotaxis whose vehicles cost $50,000 to $100,000. Software performance, not the factory, is still the scaling risk in that note. The October 6 reprint of $360 says that math has not been booked.
Wall Street’s Tesla Targets Run From $157 to $480
The rest of the Street is not on Goldman’s number. The spread is the valuation argument in one grid: some desks are underwriting robotaxis and Optimus as if they were already a business, and some are still underwriting a car company with a rich multiple.
SELECTED TESLA TARGETS AROUND THE Q3 PRINT
| Firm | Rating | Price target |
|---|---|---|
| HSBC | Reduce | $157 |
| Goldman Sachs | Neutral | $360 |
| Truist | Hold | $370 |
| Baird | Outperform | $475 |
| Ladenburg Thalmann | Buy | $475 |
| RBC Capital | Buy | $480 |
HSBC raised its target on October 5, to $157 from $125, and kept Reduce. Tesla at $378.73 is more than double that bear case. Baird, Ladenburg Thalmann’s Michael Legg, and RBC Capital’s Tom Narayan sit at $475 to $480 on the other side. Deutsche Bank’s Edison Yu lifted his target to $420 from $370.59 after the delivery beat, citing stronger September sales in Europe and China, while cutting full-year earnings per share to $1.40 from $1.48 and raising revenue to $106.1 billion from $105.4 billion. UBS left its target at $385, almost on top of the stock.
Truist’s William Stein, at Hold and $370, made the same pivot Goldman did. Deliveries beat, he said, but Tesla gave no new read on AI projects or new vehicles, and FSD and Optimus matter more for long-term cash than the auto count. Heading into earnings, one compilation showed 12 Buys, 11 Holds, and two Sells, a Moderate Buy, with an average target of $391.40, about 3% above $378.73. Goldman’s $360 is the lower-conviction version of that crowded middle.
What the $25 Billion Capex Bill Buys
Tesla told investors in the second quarter that 2026 capital spending should exceed $25 billion. The list attached to that figure is the same list Goldman is now using to hold the stock: AI compute, robotaxis, Optimus, semiconductor manufacturing, and factory expansion. Second-quarter revenue was $28.24 billion, up 26% from a year earlier. Adjusted earnings were $0.33 a share, a miss against the Street. Volume went up. Profit did not follow in a straight line.
Tesla vehicle deliveries and storage deployments represent only two measures of the Company’s financial performance and should not be relied on as an indicator of quarterly financial results, which depend on a variety of factors, including average selling price, cost of sales, foreign exchange movements and others.
Tesla, Inc., Q3 2026 production and delivery release, October 2, 2026
Goldman is, in effect, taking Tesla at its word. A 486,532 delivery quarter can still produce a dull earnings night if prices, mix, and the $25 billion spend eat the beat. Custom inference chips, the AI5 and AI6 designs Tesla has been pushing toward 2027 and 2028 production, plus a planned Terafab Tesla has described as necessary to feed Optimus, sit inside that spend. They are also why Delaney can say AI matters more than Q3 earnings and still cap the 2026 EPS upside in the same paragraph.
Croatia, Austin, and Fremont Are the Checkpoints
Europe is the near-term FSD test Goldman flagged. Croatia’s approval of FSD (Supervised) adds a small market to a list that already includes the Netherlands, Belgium, and Slovenia. The larger prize is an EU-wide vote, and that vote has slipped. A decision may not arrive before December, which means the October 21 call can talk about Europe without closing it.
Austin is the robotaxi test. Cybercab is in limited service there, and Tesla is shopping for operators who would own fleets, run hubs, and host events. Seven metros by the end of Q2 is a network, not a national roll-out. Night running has been a separate constraint; Tesla’s robotaxi hours have not been 24/7, and the software stack still has to earn that extension.
Fremont is the humanoid test. Tesla is converting the old Model S and Model X space into Optimus production, with early builds earmarked for an “Optimus Academy” to gather training data. Elon Musk has called the robot a huge business and, more recently, one of the hardest products Tesla has tried to make at scale, because so much of the bill of materials is new. Goldman can put humanoids in the stock-driver sentence and still need a factory that ships more than prototypes.
October 21 Will Not Settle the Argument
Tesla will release third-quarter results after the market close on Wednesday, October 21, then take questions at 4:30 p.m. Central Time. The delivery beat is already in the price. The open questions are average selling price, automotive margin, and how much of the $25 billion capex plan showed up in cash in the quarter. Those are car-company questions. Goldman has already said they are not the ones that should re-rate the stock.
If Tesla posts a clean earnings beat and says little new on unsupervised miles, FSD take rates, Cybercab utilization, or Optimus output, Delaney’s framework says the shares should not chase the print. If Tesla posts messy margins and a sharper robotaxi or Optimus update, the same framework says the print is still not the point. Either way, $360 is the line Goldman refuses to cross until those updates turn into a financial run-rate.
HSBC’s $157 and RBC’s $480 cannot both be right about the next twelve months. Goldman has chosen a third seat: Tesla is becoming harder to value as an automaker, and the bank will not underwrite the AI premium at $378.73. The October 21 webcast is where Tesla has to put numbers on the projects Delaney named. Until then, Neutral and $360 are the same sentence.
Disclaimer: This article is news reporting and analysis of company statements and analyst research notes. It is for information only and is not a recommendation to buy, sell, or hold Tesla stock, and it is not a forecast of future prices or earnings. Price targets, ratings, and estimates are opinions from the named firms, not facts, and they can move when Tesla reports or when those firms update their models. Readers should consult a licensed financial adviser or other qualified investment professional before making any investment decision, and they should treat every figure here as of the dated source attached to it.
-
NEWS6 months agoThe Orchard Bug Forced Zcash to Open an Empty Pool
-
GAMING6 months agoThe $5.99 Game That Won Steam’s 2026 Summer Sale
-
NEWS6 months agoAtlantic City Turned WWE’s Return Into a Civic Night
-
ENTERTAINMENT7 months agoDon Lee Joins Hemsworth as Extraction 3 Starts Shooting
-
NEWS5 months agoYouTube Shorts Retires Dislikes and Swaps Likes for Hearts
-
NEWS5 months agoNEURA Robotics Is Spending Its $1.4B on a Machine Stack
-
BUSINESS6 months agoNorway’s Export Bank Backs Nscale’s $790 Million Narvik Loan
-
GAMING6 months agoGame of Thrones Dragonfire Follows Boston’s Conquest Playbook
