Carlsbad – Gigaio wants to be a one-stop shop for artificial intelligence inferences in infrastructure.
The company offers a scalable processing infrastructure that can combine power of up to 64 GPUs on a single server via interconnected technology. That level of performance is particularly useful for AI companies that are training the inference capabilities of large-scale language models.
“We're looking forward to seeing you in the future,” said Alan Benjamin, CEO of Gigaio. “We use less power. Often we use 30% less power to achieve the same result, making our systems more efficient.”
Gigaio's revenues increased more than five times in 2024, according to Benjamin.
That increased demand requires additional working capital to build the company's sales and marketing team. Gigaio announced in mid-July it had secured its first $21 million in two tranches for the Series B funding round.
According to Benjamin, the company aims to close the second tranch in the round, estimated to be about half of the first tranches by the end of September.
“There's a market window here. We want to make sure we have capital letters,” he said.
“Inference is getting bigger. We have a great solution for reasoning. That's when we capitalize.”
The funds are the first Gigaio raised in nearly four years and place the company in position when it comes to revenue before a positive interest, tax, depreciation and amortization (EBITDA) test in 2026.
Computation solutions of all sizes
Since launching in 2012 by La Jolla's Evonexus Startup Incubator, Gigaio's product line is centered around a supernode that combines calculation, GPU, storage and networking resources for one system that allows users to avoid assembling the entire computing system and the entire processing system.
“Our customers, including most of the National Laboratories and some of the largest AI homes, want to bring the system together,” Benjamin said. “So we're (and) from Supernode.”
The company has also created a small version of the supernode system called Gryf. The customer company approached Gigaio and requested a device with similar features, but small enough to be used as carry-on luggage on a plane. This system provides EDGE users with the ability to take advantage of the computing power that is usually only seen through data centers.
“The customers loved it right off the go, the eyes were like, 'I need this, and now I need it,'” Benjamin said. “And we remember coming out of the meeting. We said to our group, “There was good news and bad news. The good news is that we love it. The bad news is that we don't know how to do this.”
With that, Gigaio has partnered with Massachusetts-based Sourcecode, a digital infrastructure and hardware co-design company. Gigaio writes GRYF software, uses multiple patented components in the system, and Sourcode generates the system.
According to Benjamin, expanding GRYF adoption will be a major focus of Gigaio's marketing efforts following the funding round.
“From a working capital perspective, it frees us,” he said of his partnership with Sourcode. “They make (gryf), we sell it, and we both share the reward for it.”
Gigaio mainly sells to major labs such as cloud service providers, national security-related entities such as the US Department of Defense, and allied governments, as well as Lawrence Livermore National Laboratory and corporate customers.
The company's goal is to provide data center-level computing capabilities in multiple formats to both cloud and enterprise customers, Benjamin said.
“We have customers who want the concept of bringing together the whole system,” he said.
“In some cases, this means going out and buying all your GPUs, making it the easiest one-stop shop possible for customers.
Gigaio
Established: 2012
CEO: Alambenjamin
Headquarters: Carlsbad
Business: AI Inference Infrastructure
Employees: 40
Funding: Approximately $40 million
Website: gigaio.com
contact: [email protected]
Notable: Gigaio is an Evonexus Startup Incubator alumni and an earlier part of the Connect Capital Match program.
Eli is an award-winning reporter who primarily covers Tech and Life Sciences Industries. He previously worked as a San Diego City Hall reporter for the local Wire City News Service. He also covers public health, transportation, state and local politics in the San Francisco Bay Area on local news issues, and is a nonprofit for the local Wire Bay City News Service, overseeing the development of outlet public health and Covid-19 news and resources webpages and daily content management. He is also a contributing writer covering minor league baseball for the baseball prospectus on the analysis and commentary website. Eli is a graduate of San Francisco State University and is a native of Northern California.
