Artificial intelligence models require an enormous amount of computing power to operate. And now, a company in Spain is raising hundreds of millions of dollars to solve that through quantum physics.
What’s happening:
- Spain based quantum software company Multiverse Computing has raised $215M USD in a Series B financing round
- The new funding round was led by Bullhound Capital and included notable participation from HP Tech Ventures, Toshiba and others
The big idea:
- Multiverse Computing has developed quantum inspired technology known as CompactifAI that leverages tensor networks to compress large artificial intelligence models
- By applying tensor networks to large language models, Multiverse Computing is able to identify and remove any redundant or unnecessary correlations in the model and only keep what is essential for performance
- Multiverse Computing’s technology is able to reduce the size of large language models by up to 95%, without causing a reduction in the speed or accuracy of the model
Why it matters:
- Multiverse Computing’s CompactifAI software could represent one of the largest breakthroughs in compressing artificial intelligence models and enabling them to run on smaller devices such as laptops or mobile phones
Going deeper:
- Many large language models are unable to run locally on small devices as they require too much memory and computational power, which ultimately leads to most popular artificial intelligence models needing to be run in the cloud instead of directly on devices
- Multiverse Computing has previously landed multiple venture capital investors and notable backers, including JME Ventures, Columbus Ventures Partners and even the European Innovation Council Fund
