Tesla Beats Earnings but Shares Dip After Hours: Optimus Timelines, FSD Subscriptions and a Possible SpaceX Deal

The presenter opens by saying that Elon Musk and Tesla "have done it again": the company beat earnings expectations, and then the stock fell after hours. He qualifies the word "crushed," noting the decline was only 2.4%.

He attributes the dip to what he calls sandbagging around the Optimus robot and some other updates. Even so, he says the earnings call and report were broadly bullish for many reasons, and he speculates that any temporary selloff might reflect Elon Musk wanting to buy Tesla shares a little cheaper.

The presenter says that a figure named Brad, who he says is apparently on the special committee deciding the price Tesla would be acquired by SpaceX for, suggested the acquisition price might be between $420 and $550 per share. The presenter adds that such a price would give Tesla shareholders a premium over the current price.

He also notes that Tesla's valuation already carries a massive premium, which he says is great for Tesla shareholders and perhaps less wonderful for SpaceX shareholders. He says he is pretty excited about Tesla's long-term prospects.

Musk suggested the Optimus ramp will be slow at first and then start scaling. The presenter explains that the company plans to build about 10,000 to 20,000 Optimus robots and let them play with each other in a self-play environment to learn, comparing the process to practicing a sport.

Musk thinks this process will take a year to a year and a half. The presenter says he thinks that is still somewhat optimistic and still sees robots as a 2030s play, though he says the pace of improvement may be arriving sooner than he originally expected.

Tesla stated a goal of 20,000 units per week, which works out to about 1 million per year. The presenter says he would double Musk's time frame and put himself in the 2030 camp for reaching a million units. He adds that Musk said the first tasks would be dangerous jobs and work that needs to run 24/7, probably factory-style work.

The presenter notes that Tesla did not mention one approach that could be used to train robots: having people wear glasses to collect vision data. Instead, Tesla said the downside is that it does not have the training data, so it wants a playground of operating robots to collect that data, which it expects will take time.

Tesla expects to eventually reach 20,000 units per week, which the presenter notes would work out to a one-year production rate of 1 million units.

On the earnings call, Tesla discussed rising demand and a ramping take rate for Full Self-Driving (FSD). The presenter says he saw this in the report too: active FSD subscriptions increased by about 170,000, which he calls pretty good and higher than expectations. About 75% of new vehicle deliveries captured FSD.

During the call, an analyst asked whether FSD revenue should be assumed to be approaching roughly $1 billion in annual recurring revenue, which the presenter describes as high margin. The presenter says the IR team did not confirm that figure. He notes that FSD is now $99 per month and the lifetime purchase option is no longer offered.

The presenter also comments on the subscription model, noting that the company responded to a question about competitors' L2 solutions by saying nobody is close to Tesla. He says he agrees that no one has what Tesla has for FSD.

The presenter says gas prices have helped accelerate Tesla's largest vehicle backlog in four years. He adds that the company cited battery capacity as the limiting factor for manufacturing more vehicles right now.

Tesla is starting to manufacture lithium iron phosphate (LFP) batteries in Austin for its Megapacks and for home storage. The presenter ties this to reducing dependence on China.

Describing his own experience, the presenter says he usually finds that a household needs an electricity bill exceeding $400 to $500 a month to justify solar plus a battery pack. The battery can offset usage during the most expensive kilowatt-hour time frame, usually 4 to 8 p.m., producing a strong return on investment, especially in California.

He says that is generally where he draws the line before even looking at solar.

The presenter says that increased battery availability would give Tesla more batteries to sell, which have historically been high margin. He suggests the company wants to share that business with China less and rely less on China, as Elon Musk has historically said.

The earnings call also said data centers are clamoring for Megapacks, which the presenter describes as bullish for the broader green energy resurgence. He says he expects green energy to come back strongly and that Tesla is a component of that, through batteries or solar.

The presenter reviews the numbers. Automotive revenue was down 3%, while gross profit rose 9%. Overall gross margin, revenue and EPS all came in better than expected. But operating margin was less than half of last year's level, at 2.6%. EPS was 22 cents, and the presenter estimates the net income margin at about 3.8%.

He says most of the pain comes from higher costs in research and development, which rose 69.3%. He reads that as suggesting heavy hiring for the AI team, the Optimus team, and work on robots and the semi-truck, including the Supercharger network for the semi-truck.

Energy and storage sales fell 15.6%. The presenter says some of that may be because of tariff refunds Tesla had to provide on energy, and possibly because of higher interest rates.

The presenter reads from the earnings call notes. Tesla said that in Q2 people would be able to add their own cars to the robotaxi fleet. The presenter says he doubts this will happen. Tesla also said it could launch Cybercabs to customers by the end of the year. The presenter says he noted 2027 but recalls the company said end of 2026, and he does not expect that to happen either.

He says Tesla mostly sees the Cybercab as a transport-as-a-service product rather than something people own.

The presenter reports that the Cybercab has 50% fewer parts than other Teslas, which he calls really impressive from a margin standpoint. He says that is probably why Tesla suggests owners would not want to own it: it lacks seat heaters and many of the bells and whistles of a normal car, which he compares to a no-frills shell.

Tesla said it has a path to 30 cents per mile in costs but is not there yet. On Optimus, the company said its cost would be close to the Cybercab's. The presenter estimates a Cybercab build cost of about $30,000 to $35,000, and with a 20% to 30% margin, a sale price of roughly $35,000 to $40,000, which would put Optimus at about $40,000 per unit. He says he finds it interesting that Optimus is similar in cost, since he had expected it to be less expensive.

The presenter says the company said the cost split between SpaceX and Tesla for Terafab is still to be determined. He suggests the reason may be that the two companies will build it together after they merge.

The presenter notes that storage deployed rose about 10%, but storage revenue did not grow anywhere near that rate, which he says was probably hit again by tariff refunds.

On vehicles, inventory declined, and Tesla blamed battery capacity as the limit. The presenter says some people claim demand is going through the roof, and he notes Tesla's deliveries in Q3 were almost on par with the prior year even though the $7,500 tax credit was removed. Other countries helped, but he says U.S. sales did not make up all of the lost credit.

Tesla plans to start commissioning equipment at its experimentation fab in Austin. The presenter says the company was blunter on the call, saying it still has to figure out how to use this equipment. He reads that as Tesla having spent heavily on expensive chip fabrication equipment without yet knowing how to use it fully.

Musk, according to the presenter, does not expect viable chips to be printed until the end of 2028 or 2029, and the presenter thinks it may take longer. Tesla currently has 400 megawatts of AI training compute and expects to reach that level by year end. SpaceX currently has 1.7 gigawatts, about four times as much, owing to the xAI acquisition, and expects up to 3 gigawatts by year end, about eight times Tesla's current compute.

The presenter reports that Tesla shipped its first Megapacks made entirely from its own LFP cells. He says the company previously bought them from China, which makes this development notable. Automotive gross margin excluding regulatory credits rose to 18.1%, below the high 20s seen in the past but better than last quarter and last year.

In the prior quarter, Tesla had to record a loss on its SpaceX shares. This quarter it recognized a gain of about $100 million. The presenter says the company took on about $1.1 billion in debt, roughly the same as what it paid down.

Capex was expected to go negative, but free cash flow came in positive: a cash flow of $6.2 billion against $4.7 billion spent. The presenter says part of that was driven by timing, with payables up $2 billion and deferred revenue up $1 billion, which he says amounts to about $3 billion that flattered the number. Without it, he estimates free cash flow would be about negative $1.5 billion, as it was last quarter.

He notes Tesla has about $50 billion in cash and receivables but $36 billion in payables, leaving about $14 billion in free cash, which he says is far below what would be needed to fund the Terafab, Optimus scale-out, Cybercab scale-out and its chip ambitions.

The presenter says Tesla stated it should be able to produce 5 million vehicles by the end of 2030, while his own projection has been 4 million. He says he got a little excited when he raised his spreadsheet from 4 million to 5 million vehicles. He notes that on a PEG ratio basis Tesla is very expensive, but the 2030 analysis gives him reason for optimism.

He first set operating income at 30% because of Cybercab margins, then reduced it to about 25% from vehicle sales, calling that less optimistic. At 5 million vehicles and that margin, the model gives a future value of about $854. The idea rests on Cybercab having potentially 50% fewer parts and mass manufacturing.

He also changed the operating margin to 20% and left everything else the same, which still gives a price target of about $740 against a $375 purchase price per share, a 17% rate of return. He says Tesla's Wall Street forecasts are much less optimistic than these inputs.

The presenter says Tesla's Wall Street forecast implies a price-to-earnings ratio of about 244 on current inputs, and with expected growth of about 47% per year, a PEG ratio of about 4.7. He concludes the stock is probably about 2.2 times too high, and that it would make more sense to buy closer to $170 per share.

He says that is unlikely to happen soon because SpaceX is about to acquire Tesla, likely at a premium. He argues that this gives Tesla stock something like a put option built in, though he stresses that it is not guaranteed.