🚨 Wall Street Reacts to DeepSeek. TL;DR:
🔹 Jefferies:
DeepSeek’s lower energy demands for AI training/inference dampen some of the optimism that followed Stargate and Meta’s major commitments last week.
🔹 Bernstein:
1. DeepSeek didn’t create “OpenAI for $5M.”
2. Models are excellent but not a miracle.
3. Weekend panic = overblown.
🔹 Citi:
Groundbreaking? Maybe...but there are questions regarding the actual type of GPUs used to train and fine tune the underlying LLM.
🔹 J.P.Morgan:
The degree to which DeepSeek leverages High-Flyer’s ~50k hopper GPUs (on par with the cluster thought to power GPT-5 training) isn’t fully known, but the dramatic reduction in inference costs seems likely.
🔹 Goldman Sachs:
1. Start-ups could have a better chance of competing against the big players.
2. Less focus on training. More inference.
3. Likelihood of global growth for Chinese companies given cost effectiveness.
@LuckyXBT__ 🍀👆🏼
New vid : Like & Comment "1". 5 will win a mini allocation to my meme coin likely worth 5 figures (Must like/follow/share)
The market is nuking....And it's giving you the best buy points ETH/RWA crypto coins in literal years. Do this NOW
https://t.co/zUcBMlL8ps