Nvidia, AI and Nemotron 3 models
Digest more
Nvidia said on Monday it acquired AI software firm SchedMD, as the chip designer doubles down on open-source technology and steps up investments in the artificial intelligence ecosystem to fend off rising competition.
Elon Musk warns of an AI hardware “all-out war” as Nvidia’s Blackwell rollout accelerates, pushing rivals to race on speed, cost, and scale.
The new SSDs jointly developed by SK hynix and Nvidia are aimed at high-performance AI workloads enabled by Nvidia’s new Rubin CPX GPUs.
Nvidia is leaning on the hybrid Mamba-Transformer mixture-of-experts architecture its been tapping for models for its new Nemotron 3 models.
While one partner said Nvidia’s acquisition of Slurm maker SchedMD, announced on Monday, makes sense for its AI factory push, another raised concerns about Slurm’s future based on Nvidia’s history with a previous software acquisition.
Nemotron-3 Nano (available now): A highly efficient and accurate model. Though it’s a 30 billion-parameter model, only 3 billion parameters are active at any time, allowing it to fit onto smaller form-factor GPUs, such as the L40S.
It's true that Nvidia is beginning to face additional pressure as companies like Alphabet and Amazon begin to release competitive chips. However, the pie is getting large enough that many companies can eat. Nvidia's CEO Jensen Huang said he expects spending on AI infrastructure to be between $3 trillion and $4 trillion by 2030.