Search results for Nvidia A100
AI & Enterprise
Orbital targets 2028 first launch for space AI data centre vision
Orbital, a startup aiming to build artificial intelligence data centres in space, has raised $5 million in seed funding. The company plans to run AI inference in orbit to avoid power and siting constraints on Earth and reduce regulatory and permitting burdens. Founder Euwyn Poon says the business depends on lower launch costs, with SpaceX\'s Starship seen as key. Orbital targets a first data-processing spacecraft launch in 2028.
AI & Enterprise
Kakao releases updated Kanana-2 open-source models
Kakao said on Jan. 20 it updated its next-generation language model Kanana-2 and released four additional models as open source. Kanana-2 was first released as open source on Hugging Face in December. Kakao said the new models improve high-efficiency, low-cost performance and significantly enhance tool-calling capabilities for agentic AI. It said the models are optimised to run on Nvidia A100-class general-purpose GPUs and use a mixture-of-experts architecture.
AI & Enterprise
LG AI Research says K-EXAONE proprietary AI model tops benchmark
LG AI Research on Jan. 11 unveiled its proprietary AI foundation model K-EXAONE, saying it consolidates five years of foundation model technology and reduces memory and compute needs by 70 percent versus EXAONE 4.0. LG said K-EXAONE scored 72 points on average across 13 benchmarks in the first round of a government project evaluation and posted 97.83 points on KGC-SAFETY. It will offer the K-EXAONE API free until Jan. 28.