Sharing what the Cloud Providers don't want you to know. Tips on negotiating your cloud deal & optimizing your environment. Also post the best cloud memes 🤝
Claude's limits are actually dogshit
It's so clear they're trying to get profitable ASAP
We already know their AWS costs are >100% of revenue today
But at least try and disguise these weak limits
Why Microsoft & OpenAI's re-negotiation is more important than people realize
>Microsoft & OpenAI re-negotiate partnership
>Microsoft now has first right of refusal on providing OAI new capacity
>Microsoft no longer required to subsidize unlimited GPUs compute
>Stargate fills in to meet OAI demand and needs
>SoftBank steps in to fund up to $40B
>OpenAI projects losses may expand up to $20B in 2027
Microsoft just dodged the bullet of being permanently tied to OpenAI's capacity growth
They now will continue to benefit from overall partnership while outsourcing the CAPEX
It's not because Satya is now bearish on the AI Arms Race, it no longer made financial sense because this will not be a "winner take all market"
With Open Source and model competition and diversity, it's the end customers choice.
I am calling it right now, SoftBank will get caught holding the bag on trying to support OpenAI's endless compute growth
Exclusive: Microsoft leaders worried that meeting OpenAI’s rapidly escalating compute demands could lead to overbuilding servers that might not generate a financial return. Learn more: https://t.co/H12eqTJctY
@The_AI_Investor If you believe AMD can capture market share on inferencing workloads, then they’re incredibly undervalued relative to NVIDIA
(I believe they will gain share massively)
All I can say is from what I see in the field, companies are foaming at the mouth from the thought of being able to replace headcount with AI
Doesn’t matter the industry
The is representative of a small sample of public companies
Morgan Stanley warns that Adobe, a 42-year-old software giant, faces severe downside risk if it fails to monetize AI quickly or if competitors leap ahead.
Adopt AI or lose everything.
The scale of the data center built today has reached unprecedented levels. Since 2010, when Azure was first commercially available, until the launch of ChatGPT in 2022, $MSFT has spent $124.2B in Total CapEx as a company (not just data center).
Today, we are talking about a single AI research company spending $MSFT's entire 12-year total CapEx in one year.
There was some quote Satya said once that Azure added more capacity in the last quarter than the first 5 years of Azure combined
The Hyperscalers are at incredible scale