Thread by @TaylorOgan

Questions/comments for @wintonARK and team re: Tesla 2025 valuation model. First of all, I appreciate the time it took to put this research together, and for responding to questions on Twitter.

@CathieDWood @ARKInvest @TashaARK @skorusARK

2/ In your bear case, you have 17,640 human-driven ride-hail Teslas coming onto a non-existent Tesla ride-hail network THIS YEAR, contributing $529M in rev. Again, this is your BEAR case. How? The network hasn't even been announced.

3/ Also, is there a reason you assume 50% EBITDA margin on Tesla's non-existent human-driven ride-hail network right off the bat (beginning this year), when Uber & Lyft haven't even been profitable?

4/ Beyond the human-driven network, my main qualm is that your model fails to stagger vehicles coming onto the robotaxi network throughout a year. This is hugely important, and I've never seen ARK model this correctly.

5/ In 2025, your bull case has 10.7M Tesla robotaxis joining the network. Not all 10.7M vehicles come online on January 1, 2025. You need to stagger those new vehicles.

6/ You could evenly weigh monthly mileage (110,000mi/12months), but a more accurate way would be to base it off of the average of each month's contribution to full-year deliveries. Seasonality.

In our models, we even break it out by # of days per month, but it does get tedious.

7/ For example, this is how monthly Tesla robotaxis added to the network in 2025 should look. If you add all those up, it gets to your 10.7M. However, only the 208k robotaxis in January '25 (plus the existing fleet) will generate revenue for the full year.

8/ Essentially, this is what your model (bull case) is assuming for net ride-hail revenue in 2025.

9/ Now adjust for seasonality. A new Tesla robotaxi brought online in November 2025 is only going to contribute a month and some change in FY25 revenue.

Using your numbers, adjusted for seasonality, you get $207B in net ride-hail revenue vs. your model's $327B.

10/ I'm not nitpicking here. 2025 is an important year for your bull case; you have Tesla somehow selling nearly TWICE AS MANY new cars than all other years COMBINED. That's a different conversation. When you account for seasonality, it has a HUGE effect on your 2025 PT.

11/ In 2025 alone, if I adjust for the above seasonality of new robotaxis brought online that year, ceteris paribus, your bull case PT falls from $4,128 to $2,884. Again, that's only when I model 2025 seasonality. It would be much lower if I modeled all years in between.

12/ Another question: on the ASP tables, what's the reason for changing base assumptions from last year's model? It has quite a big effect on this year's model without explanation (or maybe I'm missing something?).

13/ Also, check row 84 (max cash deployable given scaling constraint). Columns G-H refer to the wrong cells (even though H is dependent upon 2026 cars producible, which is probably why no one caught it). Small effect on PT (-0.33% for bull).

Just some initial thoughts.