3/3 It is a pity that king of internet from last 2 decades is just playing catch up in AI world that they brought into the world
They are just a catch up cloud player now and almost struggling to maintain their search revenues.
Is this the legacy $GOOGL carrying into AI world?
1/3 $GOOG Disappointing performance
- $META is doing better in using AI for ad growth
- $MSFT is doing better in growing Azure (and paying down their debt, buying back their stock)
- Anthropic and OpenAI are doing better in AI models while they invented transformer
2/3 $GOOG needs change of leadership to
- Show how TPU is competitive advantage and better margin Cloud growth
- Use Android and Chrome distribution better to grow AI services
- Get competitive in personal agents and aggressive in capturing AI wave
@brian_armstrong Definitely this tweet will haunt @jason some day (probably sooner rather than later based on what I see being long time in Bitcoin ecosystem)
@BlumbergJa71286@burrytracker Most commitments are over a period of 5-10 years. And there are no specific annual commitments. And assumption is that those will ramp up as capacity comes online. Or in other words, usually growing companies are expected to spend more in later year and less in immediate terms
Anthropic IPO is crucial for 2 reasons for stock market
- Cash on books for the objectively best AI model company is good for health of AI spending
- Market listing and GAAP accounting based reporting of the most relevant AI signal is critical to avoid false narratives floating about AI every other day
#anthropicAI may not do best in market immediately on listing as we see with most IPOs , but it is important that we see it play out in public markets rather than private market.
After all, AI trade not only defines a sector but it represents whole world economy now with major chunk of debt of world linking to AI in some form or other.
While airlines comparison feels apt when looking at economics, airlines can not create demand like AI can do.
There are only so many people who would fly even if it was cheaper or better experience.
So with all due respect to Ken Griffin, I think narratives are being conflated.
Super interesting comparison of GPU finance and the AI buildout to airlines. I’ve seen two other people make this comparison now:
1. Ken Griffin
2. Sridhar Ramaswamy
Both are very, very smart so worth learning more.
@JohnTinsmanAOT Asian conglomerates like to pay themselves a fat check in dividend rather than buying stock back only.
The buyback is better narrative is mainly a US thing with shareholder returns kept as first priority and was made even more popular by likes of Berkshire pushing for same.
So we went pretty fast from
“Chinese CXMT is coming for US/Korean suppliers”
to
“Chinese CSPs want Samsung and SK Hynix to do 10 year contracts for memory to ensure they get enough”
$SKHY $MU
@HomoSapienLCY@chontang Yes, every tech company should focus on minting existing tech for profitability and never invest in developing future tech. That would be perfect recipe for making money... Hilarious!!!
@chontang Because you started from Deepseek tokens which is again tiny. Much of paid usage is driven by Opus / Sol.
And again 75% if compute is being used today for training and not for inference.
$SKHY says the prices keep going up while supply is increasing.
Customers are willing to sign 5 year contract for a commodity play which is supposed to be disrupted by Chinese memory next year.
Bear narrative is falling apart on semi trade!!!
$MU $NVDA $AVGO $AMD
Did you understand what he meant? He is referencing value of AI tasks... It is like saying you can use you expensive macbook to play online lottery or build trillion dollar empire.
and also he is telling you how the cost to run a repeated query is negligible for them while user finds same value.
@jimcramer Would you recommend Jensen to buy treasury bonds if yields cross 6%? What kind of argument is this?
Junk bonds trade higher when yields are high. This is normal. CRWV issued coupon of 9%+ and it was never the senior most loan on their assets anyway.
Add to that the amount of money that is already raised or approved by financing and not spent yet.
Are shorts just counting on political opposition to AI now because we do not like data centers in neighborhoods? Is that hope for bears at these valuations in AI trade?
The token demand is going up
The price of AI infrastructure is going up
And Anthropic proved that they can make money on selling inference
What's not to like, really?
Bears are running the narrative recently for AI trade. But what does actual data tell you about demand?
GPU rent trackers show that rents are going up as we speak.
And the AI spenders are flush with cash.
So why so much pessimism? #AI $NVDA $AVGO $MU