The near 8x increase in compute capacity since March 2023 is the eyebrow raiser for me. My belief is that the vast majority of speculative value in Tesla has almost nothing to do with their hardware and everything to do with the potential of FSD.

If you believe the hypothesis that Americans will pay a huge premium (directly or indirectly by subsidizing the poor build quality) to decrease the mental load by 85% on their twice-daily 35 minute commute, then the company is in good shape.

On top of this, their data moat is enormous and their pipeline is mature. Other car companies seeking this level of autonomous fidelity will need to either race to start harvesting as much data as possible, or make a bet that this quantity will be unnecessary with future models.

Then again, if you don't buy the value-add promise of FSD hypothesis then this company is faltering hard. Cybertruck is flopping in sales (with life-threatening build issues as a bonus), the company is facing deteriorating public perception, the tightening economy makes the 'premium'/'apple' presentation less appealing, serious competitors recently entered the EV market, on and on.