Igor's call, in full

Elon Musk wants to build the world's largest chip factory from scratch, with zero semiconductor experience, and take on ASML. The need is real and the strategy is right: Tesla's AI ambitions need chip volumes no outside supplier will commit to on Musk's timeline, so vertical integration is the correct move. The timeline is the gamble. Based on my 84 percent accuracy rate, I forecast Terafab produces its first meaningful output in 2028 at the earliest, and full scale is a 2030 story. Add two years to whatever Musk promises, the 4680 battery lesson applies. The specs on the table: $25 billion, 100,000 wafer starts per month scaling toward 1 million (about 70 percent of TSMC's output), 100 to 200 billion custom AI chips a year, at 2 nanometers. And I will debunk the viral nonsense: the claim that Tesla builds 1,000 EUV machines a week is fiction. ASML makes fewer than 100 a year, at up to $380 million each, and holds a complete monopoly. ASML is a customer, not a competitor. The real pressure lands on the memory makers, Micron, SK Hynix, and Samsung, if Tesla builds its own HBM at scale.

Adjacent calls

South korea owns the most profitable layer of the AI chip stack, with sk hynix and samsung controlling more than 90% of global hbm by 2027-28.Early callX money is elon musk's everything-app financial layer, the last piece of a 27-year plan. By 2030 it processes over $1 trillion a year and banks start losing deposits.Early callLive sports grows into a $10 trillion attention economy by 2050.On track

Ask Igor your own question. 5 free.