Tesla's once tech-like margins have been compressed down to typical car company margins. The earnings call and shareholder letter were the typical mix of goal setting and goal moving. This quarter they announced once again production is constrained by supply chain problems. I find it astonishing they still can't get this straightened out.

It's a good problem to have, to have demand soaring, as FSD improves and with high gasoline prices, forcing them to add shifts at their car plants to meet the demand.

Optimus is delayed yet again, from summer 2027 to Q4 2027. Optimus may be a big product one day but I am extremely skeptical that day will be any time soon. I've never seen a video of Optimus doing anything I need a robot for, for example. One problem is that LLMs were trained on enormous data sets, essentially all the writings in human history, which were available almost for free. Similar data sets for sensory inputs like touch do not exist for free, and in most cases do not exist at all and must be created which requires both time and money. I would be delighted to be wrong, but I don't see a generalized AI robot doing using useful tasks any time soon.

Optimus is the future, but as you say, it needs a lot of finishing touches and then training. All that R&D puts a damper on earnings margins, for now, but if Optimus ends up being the major revenue generator for the company, it will have been a good investment.

I believe robotaxi will be a big business, but I don't think it will be a particularly good business, simply because it is competing against the Uber/Lyft model, where the driver/owner provides services like parking, cleaning, and maintenance essentially for free. In any event, it is long way out and Tesla is clearly lagging Waymo technically and importantly is very far behind in the regulatory process. Plenty of competition on the horizon too.

Uber costs about twice as much per kilometer as an unmanned robotaxi, even when you include cleaning and maintenance. How will Uber compete? It currently makes about twice the earnings of Tesla, and most of that volume should go to the low-cost competitor, at better margins.

It is hard to see how robotaxis are lagging Waymo technically - both are very safe, using different technologies, but it is Tesla that can do it at something like half the cost, so if it works, and it looks like it is working, this will have been the correct technical orientation. Tesla can ramp its volume as high as it wants to, although they are probably waiting for v.15 of FSD to really go vertical. Have a look at robotaxitracker.com to see how quickly the rollout is going already.

Teslas using FSD are now at about 10x safer than human drivers, and v.15 may tip that to another order of magnitude, where regulators are not going to be able to sensibly oppose deployment, and although some will drag their feet, it is looking like generalized approval is inevitable.

Cybercab was touted as being a very low cost vehicle that could be produced in large volumes, flooding the robotaxi market. This is shaping up to be blunder. Tesla can only deploy a small number of new vehicles each week so the assembly line is mostly idle. It will take them years before it is running at full capacity. They could have rolled out their robotaxi business using the Model Y at a fraction of the cost.

Tesla is being cautious with its rollout of cybercabs, but it is more than doubling the number in circulation every month. That assembly line won't be idle for long, at the current explosive rate of cybercab introduction. They could have done it with the Model Y, but at much higher costs, since the cybercab is considerably cheaper to produce and to maintain. And if they want to satisfy the recent surge in demand, it makes more sense for them to sell Model Ys to consumers and do the taxiing with cheaper Cybercabs for which they have plenty of capacity.