Topics
Latest
AI
Amazon
Image Credits:Jonathan Raa/NurPhoto / Getty Images
Apps
Biotech & Health
mood
Cloud Computing
Commerce Department
Crypto
Enterprise
EVs
Fintech
fund raise
Gadgets
Gaming
Government & Policy
Hardware
layoff
Media & Entertainment
Meta
Microsoft
Privacy
Robotics
Security
Social
distance
Startups
TikTok
Department of Transportation
Venture
More from TechCrunch
issue
Startup Battlefield
StrictlyVC
Podcasts
Videos
Partner Content
TechCrunch Brand Studio
Crunchboard
Contact Us
DeepSeek , a Chinese AI startup , became the talk of the proverbial AI town when it release its R1 model on Friday . R1 ’s functionality and accuracy compared to its U.S. counterparts , despite using few resources and less compute power , seems like a win for the overall AI industry . But it is n’t necessarily good news for everyone .
semiconducting material giant Nvidia finds itself in the headwinds of DeepSeek ’s late achievement . The chip giant regard its broth plump 16.9 % from the close of Friday ’s public markets to the close of public market on Monday , according toYahoo Finance datum . Nvidia lost about $ 600 billion off of its grocery store cap . Nvidia ’s store close at $ 142.62 a share on Friday good afternoon . On Monday , it close up at $ 118.58 .
There is speculation that the reason why DeepSeek ’s good example release would bear on Nvidia ’s farm animal is that R1 provides a clear representative that AI models do n’t of necessity need expensive , gamey - end chips or hardware to build an impressive example , which is n’t just great news program for a chipmaker like Nvidia .
“ DeepSeek is an excellent AI progress and a complete example of Test Time Scaling , ” an Nvidia interpreter secern TechCrunch over email . “ DeepSeek ’s work illustrates how new role model can be created using that proficiency , leverage widely - available model and compute that is fully export control compliant . Inference requires significant number of Nvidia GPUs and high-pitched - performance networking . We now have three grading laws : pre - training and post - preparation , which continue , and newfangled trial - time grading . ”
The timing of all of this is interesting because this comes one calendar week after former president Joe Bidensigned an executive director orderthat made further restriction on the export of U.S.-produced modern AI chip to certain countries , with near - cover restrictions on sending chips to country like China , where DeepSeek is found .
At the clip , Nvidia said that the executive order was “ unprecedented and misguided ” and that it would “ derail ” origination and economic growth worldwide .
President Donald Trump has sincereversed Biden ’s executive orderand has contract a dissimilar executive order to make theStargate Project , an substructure program that will commit up to $ 500 billion into AI data point centers .
Join us at TechCrunch Sessions: AI
Exhibit at TechCrunch Sessions: AI
The passing of DeepSeek shows that if the U.S. wants authority over the global AI securities industry , it may need to give attending to more than just chips and AI computer hardware .
