AI Chip Startup Mythic Reaps $40M from SoftBank, Lockheed Venture Arms

Mythic, which last year unveiled its novel computer chip designed for artificial intelligence, announced today that it has raised $40 million in a Series B fundraising round led by the venture capital arm of Japanese telecom and tech giant SoftBank.

Mythic, based in both Redwood City, CA, and Austin, TX, is one of the upstart companies benefiting from a surge of investing in the chip-making field, where the growing computational demands due to the use of AI are creating an opening for young entrepreneurs with out-of-the-box ideas.

The boxes Mythic wants to break customers out of are their big data centers and the rooms full of servers run by Web-based data processing and storage companies. The startup has developed a new chip design it says is small enough, and cheap enough, to pack AI-strength computing power into smartphones, drones, robots, virtual reality headsets, wearables, and self-navigating cars. Rather than sending the data these machines collect out over the Internet to mega-processors elsewhere, the devices themselves could crunch data and make some conclusions or decisions. This falls into a category dubbed “edge computing,” which shifts some of the workload of AI analysis to devices at the periphery of a network.

Mythic co-founder and CEO Michael Henry (pictured) says the startup’s technology is competitive with that of entrenched chipmakers such as Nvidia (NASDAQ: [[ticker:NVDA]]) and Qualcomm (NASDAQ: [[ticker:QCOM]]), which are among the companies racing to provide units with computing power that can stand up to the needs of artificial intelligence applications like the radar or computer vision needed to navigate an autonomous vehicle.

“If you have Nvidia’s $5,000 box in the trunk, we can replace it with a $100 box with far lower power consumption,” Henry says.

Since Mythic emerged from stealth almost exactly a year ago, it has been working to prove its big claims to investors and large companies it hopes to attract as customers, its CEO says.

The fundraising round announced today not only provides an influx of cash, but also a benefit as important—validation that its evidence has won over some major players, Henry says.

In an announcement on the fundraising, SoftBank Ventures managing director JP Lee called Mythic “a clear leader in the rapidly growing AI edge chip market.”

SoftBank has recently become a cornucopia of investments in tech companies of all kinds, distributing funds from its $100 billion SoftBank Vision Fund and from its corporate resources. In a deal closed in January, SoftBank became the largest single shareholder in ride-hailing company Uber, the same month when the Vision Fund led an $865 million fundraising round for Menlo Park, CA-based Katerra, a tech-enhanced construction firm.

Henry says the Vision Fund was involved in the discussions with Mythic that led to an investment by SoftBank Ventures, which handles early-stage investments. The Vision Fund is dedicated to making major global investments in “the next stage of the Information Revolution.”

Back in 2016, SoftBank puzzled tech observers by spending nearly $32 billion to acquire U.K.-based ARM, the design company behind the core architecture of processors made by almost all chipmakers. In February, ARM announced a new machine learning platform called Project Trillium, which aims to bring advanced computing power, and functions such as object detection, to mobile devices. ARM plans to preview a suite of IP from the edge computing project in April, and make it generally available in mid-2018.

ARM president Rene Haas will be joining Mythic’s board as a result of the SoftBank Ventures investment. But Mythic’s CEO says his company is working separately from Project Trillium. Mythic occupies a different technology niche for SoftBank, Henry says.

“They’re putting together the pieces for what, I think, is going to be the future world,” Henry says.

Mythic and other edge computing innovators are trying to create small, scaleable chip designs with low power demands that extend battery life and don’t require energy expenditures to keep them cool, like big server arrays do. One of Mythic’s tactics is keeping costs low by basing its chip design on a 40-nanometer standard that was state of the art when the company was founded in 2012. Mythic packs the chip with flash memory, and shifts computing tasks to memory cells, eliminating a separate step formerly assigned to processors. Paired with other innovations, the computational speed of the chip is competitive with more miniaturized AI chips now made to the current state-of-the-art standard of seven to 10 nanometers, Henry says.

SoftBank Ventures was joined in Mythic’s Series B round by another new strategic investor, Lockheed Martin Ventures, whose general manager Chris Moran called Mythic’s technology “game-changing.” It “will allow us to provide our customers with the unmatched ability to obtain timely, actionable information from distributed, power constrained systems,” Moran further stated in the fundraising announcement.

Also participating in the round were earlier investors DFJ, Lux Capital, Data Collective, and AME Cloud Ventures. Mythic also disclosed an individual new investor, Andy Bechtolsheim, the billionaire co-founder of both cloud networking company Arista Networks and pioneer software and computer company Sun Microsystems.

The new money brings Mythic’s fundraising total to $55.2 million. Henry says the company, which now has 45 employees, will beef up its engineering staff to help prospective customers adopt its chip for applications in a broad range of sectors, including autonomous vehicle navigation, robotics, smart home devices, connected industrial machines, and virtual reality. Engineers design software toolkits for the various applications, he says.

Those efforts will serve a crucial goal for 2018—attracting “lighthouse customers” in important industries that could further validate Mythic’s technology, Henry says.

“Companies are in the process of selecting their AI,” Henry says. “In 2019, people will be making decisions.”

Photo courtesy of Mythic

Author: Bernadette Tansey

Bernadette Tansey is a former editor of Xconomy San Francisco. She has covered information technology, biotechnology, business, law, environment, and government as a Bay area journalist. She has written about edtech, mobile apps, social media startups, and life sciences companies for Xconomy, and tracked the adoption of Web tools by small businesses for CNBC. She was a biotechnology reporter for the business section of the San Francisco Chronicle, where she also wrote about software developers and early commercial companies in nanotechnology and synthetic biology.