April 3, 2024
luminous computing ai 105m series bill

Luminous Computing Raises $105M in Series A Funding

Luminous Computing Inc., a startup creating a novel class of supercomputers for running AI models, said today that it has completed a $105 million investment round.

There were more than six investors who took part in the Series A financing. Bill Gates, Gigafund, 8090 Partners, Neo, Third Kind Venture Capital, Alumni Ventures Group, Strawberry Creek Ventures, Horsley Bridge, Modern Venture Partners, and others were among the contributors.

Based in Mountain View, California, the supercomputer being built by Luminous will run AI models. Based on silicon photonics, the supercomputer can considerably accelerate some processing workloads.

The information processed by a conventional processor is stored as electricity. To do computations, the processor manipulates this electricity, such as by transferring it between transistors. A distinct method is used by silicon photonics technology, which encodes data as light instead of electricity.

In some cases, light may move far more quickly than electricity. The idea behind silicon photonics is to make use of the fact that light moves more swiftly to carry out computations than is feasible with a conventional processor in which data is encoded as electricity. That is the goal that Luminous is attempting to achieve.

Few technical information concerning Luminous technique has so far been disclosed. The business did, however, disclose its development objectives in conjunction with today’s investment release.

Luminous wants to create a silicon photonics device that performs 3,000 times better than a third-generation Tensor Processing Unit, or TPU, circuit board. To execute AI models, Google LLC designed TPUs, specialized semiconductors. The search engine giant utilizes the chips to power internal apps and makes them available through its public cloud.

The more hardware is needed to train a neural network, the more complicated it is. Due to the increasing complexity of AI models, this presents difficulties: According to OpenAI, between 2012 and 2018, the computing power needed for the biggest AI training runs increased every three and a half months. According to Luminous, their technology will make it simpler to deliver the expanding computing capacity required for the advancement of AI. The business will quadruple its technical staff thanks to its most recent $105 million fundraising round. To expand its chip technology, Luminous will work with photonics designers and other specialists. The company will also create companion software to make the technology easier for future users.