for most diseases, the best medicines are soon to come. why?
better disease targets x 10 new/better modalities x better discovery x better clinical precision = golden decade for new drug creation, addressing 2 big opportunities:
1. treat the ~75% of 13,000+ disease segments w no approved medicines
2. replace last generation medicines w more effective and safer meds
https://t.co/6Mnw8aiu4D
noting the summary points of his arguments (below) are at odds with every single company in the space's earnings call disclosures, comments, and forward projections. Zero evidence for these claims, lots of negative evidence for his claims
-AI infrastructure risks collapse as frontier model ARR growth slowed sharply since July after price cuts failed to offset volumes.
-Compute supply next year vastly exceeds inference demand; without broader app breakout, capex peaks around 2028 then rolls over.
@neilsethinew good analysis ๐
in one sentence: failure is now priced in; any level of success at even well below run-rate is upside = assymetric risk reward. + management can buy if market won't
yes, price x quantity = revenue. bear narrstive treats quantity as unimportant, despite huge growth.
LTAs lock in floors and ceilings for price
all available data shows quantity is growing 20-40% per year, supply constrained, with far more demand than can be supplied. try buying a mac studio with 512gb ram right now.
the bear case is also more than priced in now, creating opportunities for take privates before markets regain their sanity
it also doesn't factor in:
1) thresholds of utility. as intelligence gets better it's a step function of use cases that get unlocked growing demand. maybe 1 in 100 people use AI intensively in enterprise.
2) the value of intelligence itself to go into other businesses, not just sell the intelligence
3) the cost reduction of labor possible to grow margins
@TheValueist where this cycle also differs is an entire new world-changing industry powered by what they make, with requirements that make it non-commodity. so basically, it differs entirely
@elonmusk@JessieChimni@fchollet@garrytan *Elon wanting the sector to crash further to destroy his neocloud competitors and acquire chip strategics at 1/5 their fair value for stock*
@RouleurCapital@robin_j_brooks it's also supply shock driven inflation from tariffs and war and loss of trade, so trying to solve it by calming demand with higher rates is like bloodletting to cure disease
one driver of this is probably that failure in an early trial attributable to a new target's biology being less validated by precedent, becomes visible much sooner than failure due to a lack of revenue or deals when your drug ultimately proves not to be best in class.
new target difficulty is appropriately worried about because there isn't as clear a clinical playbook, even if the disease opportunity is more greenfield.
a key question is how many new targets truly "deserve" to be made on the strength of their preclinical and genetic evidence, but are not. @Atelfo
@TheWiseAdapter and at 20-30% net Income yields, why wouldn't they borrow $500B and buy even more if the market doesn't believe their cash counts...
why wouldn't a P/E firm back an LBO
why wouldn't an ibanker pitch nvidia or elon to buy it to fund all their futures accretively..
Dean, i hear you completely. it's an anomaly in the matrix. probably truly just need to wait for the company to be able to buy back in a month. this capitulation is brutal and nonsensical
floors come from:
1) buy backs
2) company borrowing debt to buy back even more
3) accretive acquisition offers from any company that has a higher p/e
4) management LBO attempts with private equity
it's getting to that level as it just collapses while all fundamentals continue to rise.
@VulcanMK5 "itS CYcliCAL". i repeat thing i hear others say. i dont listen mgmt call. i dont know spreadsheet. $200B profits per year? fake money can't be used buy things. iTS cyCLiCAL
will this idea that this is the same product as 2018 commodity memory for video game laptops ever die? different highly technical product now; totally new demand era for totally new uses; new contract structure. this is a company telling you it has 3+ years of pre sold inventory into the biggest technology market in the world.