@evrgn11112231@OnodaCapital Yeah, it’s a framework, not some precise analysis — that’s why the framing is generally ~$0.5T or ~$1.0T by some date
But also, it’s an ability to pay analysis for the CSP backlogs, which are 40-50% OAI/ANT (80-90% of AI backlogs)
@evrgn11112231@OnodaCapital Assume labs run breakeven-ish. Lab ARR = csp rev. Take that down to NOPAT to calc ROIC. Pick ROIC threshold to justify spend. Get to “right” capex $ vs lab NNARR. Capex = silicon demand.
How rich people think about money... I find it fascinating how differently those who made the money vs those who didn't approach basically every financial decision, change, etc.
Those who didn't make it themselves are pretty insufferable
@OnodaCapital My fine is my judgment of expectations — I think Oct is looking inline vs informed people’s expects
If OAI can only raise $10-20B, yes, it’s a fake mark, but it’s a laughably low amount of money — feels like the next step is turning training into inference and flooding supply
@OnodaCapital I guess my point is, ANT seems fine (not particularly good) so far + next 2 months prob suck, and that (coupled with OAI maybe declining and out of external funding sources) is the real issue
@P_Remarks This was my AI ARR miss outlook a little while ago - think these views mostly hold still, although Iran getting worse and AI solving alchemy throwing a wrench in some of them
https://t.co/NXuWefuMLj
@rev_cap nGDP revised like 1-6% lower but stays positive… but non-AI capex prob goes up a bit
Bonds up / rates down
USD down
Gold up
SPY/QQQ down
RSP/IJH/IJR maybe a push with beta down vs fundamentals okay / flows good
Ex-US stocks maybe good, especially with currency — latam?
@P_Remarks Non-AI stuff isn’t expensive, and AI capex crowdout goes the other way, consumer is fine for now, idk I just don’t find the world all that dire