contextwindows's profile picture. Public market investor. Keeping you in the loop on the latest developments in AI and its impact on the public equity market.

Context Window

@contextwindows

Public market investor. Keeping you in the loop on the latest developments in AI and its impact on the public equity market.

Pinned

Welcome to Context Window - your guide to the AI market's ups and downs. We will provide commentary and analysis on the latest developments in AI and its impact on the public equity markets. Let us be your guide through AI winter. Follow along.

contextwindows's tweet image. Welcome to Context Window - your guide to the AI market's ups and downs. 

We will provide commentary and analysis on the latest developments in AI and its impact on the public equity markets. 

Let us be your guide through AI winter. Follow along.

Context Window reposted

200k is just the amuse-bouche


Did you miss me? $NVDA


$NVDA Jensen Huang is done selling for now.

contextwindows's tweet image. $NVDA Jensen Huang is done selling for now.

Another 100k GPU cluster coming online soon $META $NVDA

Meta $META is expected to soon finish its massive GPU cluster featuring 100,000 of Nvidia's $NVDA H100 GPUs, with operations commencing in October or November to train Meta's Llama 4 model.



Blackwell update $NVDA $SMH $SOXX

Nvidia’s $NVDA GB200 is reportedly set for mass production and ramp in December, according to a report from the Commercial Times.



Great points @ResearchQf $NVDA $SMH $SOXX

"I would be very surprised if it is not able to do economically valuable analytical (R&D) work inside large companies." Also reached out to industry practitioners and began getting similar feedback. This alone is big (there are more). Then extrapolate HW & models 1-3 years out.

ResearchQf's tweet image. "I would be very surprised if it is not able to do economically valuable analytical (R&D) work inside large companies."

Also reached out to industry practitioners and began getting similar feedback.

This alone is big (there are more). Then extrapolate HW & models 1-3 years out.


Bingo

My dumb big-picture take: I'd think it'd be new hardware for training, old hardware for inference. And as long as things keep improving so fast, they'll build new infrastructure for each new big model. Which would skew inference towards using the same infrastructure it trains on.



$TSLA we are less than 1 month from Robotaxi event, which will be focused on autonomous driving and likely Tesla's other AI efforts (Optimus). Setup looks vaguely similar to breakout in 2023

contextwindows's tweet image. $TSLA we are less than 1 month from Robotaxi event, which will be focused on autonomous driving and likely Tesla's other AI efforts (Optimus). Setup looks vaguely similar to breakout in 2023

Context Window reposted

🎯 AI is the path to drive cloud growth in the 2020s. SaaS is mature. What's the most important metric for hyperscalers? Cloud growth. 🚀


Ellison. One of a kind. $ORCL $NVDA

To source the 131,072 GPU Al "supercluster," Larry Ellison, appealed directly to Jensen Huang, during a dinner joined by Elon Musk at Nobu. "I would describe the dinner as me and Elon begging Jensen for GPUs. Please take our money. We need you to take more of our money. Please!”



United States Trends

Loading...

Something went wrong.


Something went wrong.