Kernel has raised $53 million to accelerate development of a new generation of brain tech, helping companies leverage insights from the latest advances in neuroscience.
General Catalyst led the round, which marks the first outside funding the company has accepted. Kernel had previously been self-funded by founder and CEO Bryan Johnson. Khosla Ventures, Eldridge, Manta Ray Ventures, and Tiny Blue Dot also participated in the round.
Johnson previously founded Braintree, which PayPal acquired for $800 million in 2013, and had backed Kernel with $54 million of his own money. More recently, Johnson said he had been on a quest to figure out how businesses could tap into brain signals to refine their own products and services.
“We can quantify and measure pretty much everything in the known universe, from distant stars to our calories and our steps and our genome,” Johnson said. “Our mind is one of the only things that we cannot consistently measure and quantify. And humans do remarkable things when we can measure something.”
The past decade has seen various cycles of hype around the field of neurotech. Several companies created headsets that could read certain types of brainwaves. These were equivalent to commercial versions of an electroencephalogram or EEG machine, but they never reached mainstream success.
Johnson said that’s because an EEG reading of the brain is like hovering over New York City at 30,000 feet and trying to listen to a conversation. So with Kernel, the company has gone a different route.
Kernel has built two new devices. Flow reads the magnetic fields generated by the brain’s neural activity, while Flux tracks cortical hemodynamics (blood flow in the brain) to represent neural activity. Together, the devices provide the power of a Functional magnetic resonance imaging, or FMRI, at a fraction of the cost.
“The question for us: Was there technology out there that we could build that would basically do for measuring and quantifying the brain what mainframes did when they became PCs and when the billion-dollar genome became a $1,000 genome?” Johnson said. “What if we could get on a cost curve like that and have broad access to high-quality neural data?”
Flow and Flux are currently being tested and used by partners as the company further refines the technology. Rather than selling them directly at this time, Kernel is making them available via a subscription platform it calls “neuroscience-as-a-service” (NaaS).
The main goal is figuring out the best use cases for this technology. Johnson believes access to rich neural data has applications in fields ranging from brain health to education, AI and machine learning, natural language processing, and consumer behavior.
“I think the most important thing on our side is we need to be mindful that this is an emerging category,” Johnson said. “And it’s going to follow a similar discovery curve [as] other technologies have done.”
Johnson believes the technology could help improve AI and ML, particularly generating the underlying data for image recognition.
“Image recognition is done by a person on Mechanical Turk who would look at the image of a cat with a couple of million pixels and then write the words for the annotation,” Johnson said. But with Kernel’s tools, “You could have the person look at the cat for 500 milliseconds for the annotation, plus you would have 500 milliseconds of neural activity for a higher dimensional rotation. And that’s been shown to improve algorithms.”
Kernel will focus on the NaaS model introduced in May over the next year or so. After honing in on the best applications, Johnson said he’ll begin to think about direct sales. But after getting the technology into the hands of customers, Johnson is clearly thrilled about its potential impact.
“If we can quantify and measure the thing that makes up our brains and our mind, it really is an introduction to a new era in the world,” Johnson said. “And not just for all the obvious things that we know, but for a whole host of brand-new things.”