AI News, Research and Tool Reviews
-

Intel Gaudi 3 outpaces Nvidia H100 in specific AI workloads
Intel Gaudi 3 delivers 1.7x faster training times for large language models like Llama2-13B and GPT-3 175B compared to Nvidia H100. The architecture features 128 GB of HBM…
-

Cerebras Wafer Scale Engine 3 and national lab procurement
Sandia National Laboratories deployed Cerebras CS-3 nodes to support AI workloads for national security. The WSE-3 architecture delivers 125 petaflops of AI…
-

Physical Intelligence’s pi0 vision-language-action model explained
Physical Intelligence’s pi0 model uses a 3.3 billion parameter transformer to combine visual perception with continuous motor control. Trained on 10,000 hours…
-

Groq LPU performance and efficiency realities
Groq LPU delivers 800 to 1,000 tokens per second for Llama 3 70B, offering significantly higher speeds than Nvidia GPUs. While providing…
-

The people shaping Anduril’s Lattice AI defense platform
Palmer Luckey and co-founders lead Anduril, a defense firm valued at $61 billion that develops the Lattice AI command system. The company…
-

Cohere revenue growth and Command R+ enterprise adoption
Cohere reached a $5.5 billion valuation in 2025 after revenue climbed to $100 million, driven by a strategy targeting highly regulated sectors.…
-

Scaling AI inference with SambaNova DataScale SN40L
SambaNova’s SN40L RDU architecture manages models with up to 5 trillion parameters using a three-tier memory system. This hardware delivers 10x better…
-

The workforce and defense contracts driving Scale AI
Scale AI manages a workforce facing a class action lawsuit involving up to 20,000 contractors while securing a $500 million agreement with…




