G42 partners with Liquid AI for enterprise edge computing
New G42-Liquid AI partnership targets private enterprise AI models
#UAE #edge - Abu Dhabi-based AI powerhouse G42 has entered into a multifaceted commercial partnership with Cambridge Massachusetts-based foundation model startup Liquid AI, a spin-out from the Massachusetts Institute of Technology. The collaboration aims to jointly develop and deploy private generative AI solutions internationally across sectors including investment, banking, consumer electronics, telecommunications, biotech and energy. G42 subsidiaries Core42 and Inception will participate in various aspects from AI infrastructure development to co-designing multimodal foundation models, targeting markets from the Middle East and North Africa to the Global South.
SO WHAT? - This partnership could prove to be a significant strategic alliance enabling private AI solutions from the core to the edge for G42 customers. Liquid AI is focused on designing faster, leaner AI models addressing critical limitations in large language models and moving beyond Transformer architecture. The partnership also sits well alongside G42’s subsidiary Core42’s focus on delivering sovereign AI solutions to governments and large enterprises in the UAE and across the world.
Here are some key details about this announcement:
G42 has formed a multifaceted commercial partnership with foundation model startup Liquid AI to facilitate the creation, training and commercialisation of generative AI solutions powered by Liquid Foundation Models.
The partnership targets private generative AI deployment across multiple sectors including investment and banking, consumer electronics, telecommunications, biotech and energy from MENA to Global South markets
G42 subsidiaries Core42 and Inception will participate in AI infrastructure development and co-designing multimodal foundation models, leveraging the holding group's established technological capabilities
Liquid AI is developing alternatives to Transformer architecture that underpins most current large language models, focusing on solutions that support edge deployment on for smartphones and other edge devices
In April, Liquid AI announced Hyena Edge, a convolution-based multi-hybrid model engineered to outperform Transformer baselines on computational efficiency and language model quality for edge computing.
Convolutional Neural Network (CNN) is a type of deep neural network that uses convolutional layers, and is frequently used for image and video processing.
The new convolution-based model promises faster, more memory-efficient AI performance at the edge, addressing critical limitations of current smartphone and device applications
ZOOM OUT - The new partnership with Liquid AI provides further evidence that G42 is adopting a more diversified approach to AI models. The move follows a strategic alliance announced last month with Paris-headquartered AI company Mistral AI. G42 stated that the collaboration with Mistral AI spans the entire AI value chain from model training to infrastructure development across Europe, the Middle East and Global South markets. The group has long been a partner of OpenAI and has developed its own Arabic-centric large language model called Jais.
[Written and edited with the assistance of AI]