This founder cracked firefighting — now he’s creating an AI gold mine

Sunny Sethi, laminitis of HEN Technologies, doesn’t dependable similar idiosyncratic who’s disrupted an manufacture that has remained mostly unchanged since the 1960s. His institution builds occurrence nozzles – specifically, nozzles that it says summation suppression rates by up to 300% portion conserving 67% of water. But Sethi is matter-of-fact astir this achievement, much focused connected what’s adjacent than what’s already been done. And what’s adjacent sounds a batch bigger than occurrence nozzles.

His way to firefighting doesn’t travel a tidy narrative. After nabbing his PhD astatine the University of Akron, researching surfaces and adhesion, helium founded ADAP Nanotech, an outfit that developed a c nanotube-based portfolio and won Air Force Research Lab grants. Next, astatine SunPower, helium developed caller materials and processes for shingled photovoltaic modules. When helium landed adjacent astatine a institution called TE Connectivity, helium worked connected devices with caller adhesive formulations to alteration faster manufacturing successful the automotive industry.

Then came a situation from his wife. The 2 had moved from Ohio to the East Bay extracurricular San Francisco successful 2013. A fewer years aboriginal came the Thomas Fire — the lone megafire they’d ever see, they thought. Then came the Camp Fire, past the Napa-Sonoma fires. Then, successful 2019, came the breaking point. Sethi was traveling during evacuation warnings portion his woman was location unsocial with their past three-year-old daughter, nary household nearby, facing a imaginable evacuation order. “She was truly huffy astatine me,” Sethi recalls. “She’s like, ‘Dude, you request to hole this, different you’re not a existent scientist.’”

A inheritance spanning nanotechnology, solar, semiconductors, and automotive had made his reasoning “bias escaped and flexible,” arsenic helium puts it. He’d seen truthful galore industries, truthful galore antithetic problems. Why not effort to hole the problem?

In June 2020, helium founded HEN Technologies successful adjacent Hayward. With National Science Foundation funding, helium conducted computational fluid dynamics research, analyzing however h2o suppresses occurrence and however upwind affects it. The result: a nozzle that controls droplet size precisely, manages velocity successful caller ways, and resists wind.

In HEN’s examination video, which Sethi shows maine implicit a Zoom call, the quality is stark. It’s the aforesaid travel rate, helium says, but HEN’s signifier and velocity power support the watercourse coherent portion accepted nozzles disperse.

But the nozzle is conscionable the opening – what Sethi calls “the musculus connected the ground.” HEN has since expanded into monitors, valves, overhead sprinklers, and unit devices, and is launching a flow-control instrumentality (“Stream IQ”) and discharge power systems this year. According to Sethi, each instrumentality contains custom-designed circuit boards with sensors and computing power—23 antithetic designs that crook dumb hardware into smart, connected equipment, immoderate powered by Nvidia Orion Nano processors. Altogether, says Sethi, HEN has filed 20 patent applications with fractional a twelve granted truthful far.

The existent innovation is the strategy these devices create. HEN’s level uses sensors astatine the pump to enactment arsenic a virtual sensor successful the nozzle, tracking precisely erstwhile it’s on, however overmuch h2o flows, and what unit is required. The strategy captures precisely however overmuch h2o was utilized for a fixed fire, however it was used, which hydrant was tapped, and what the upwind conditions were.

Why it matters: Fire departments tin tally retired of h2o otherwise, due to the fact that there’s nary connection betwixt h2o suppliers and firefighters. It happened successful the Palisades Fire. It happened successful the Oakland Fire decades earlier. When 2 engines are connected to 1 hydrant, unit variations tin mean that 1 motor abruptly gets thing arsenic a occurrence continues to grow. In agrarian America, h2o tenders, which are tankers shuttling h2o from distant sources, look their ain logistical nightmares. If they tin integrate h2o usage calculations with their ain inferior monitoring systems to optimize assets allocation, that’s a elephantine win.

So HEN built a unreality level with exertion layers, which Sethi likens to what Adobe did with unreality infrastructure. Think Individual à la carte systems for occurrence captains, battalion chiefs, and incidental commanders. HEN’s strategy has upwind data; it has GPS successful each devices. It tin pass those connected the beforehand lines that the upwind is astir to displacement and they’d amended determination their engines, oregon that a peculiar occurrence motortruck is moving retired of water.

The Department of Homeland Security has been asking for precisely this benignant of strategy done its NERIS program, which is an inaugural to bring predictive analytics to exigency operations. “But you can’t person [predictive analytics] unless you person bully prime data,” Sethi notes. “You can’t person bully prime information unless you person the close hardware.”

HEN isn’t monetizing that information yet. It’s implementing information nodes, putting devices successful arsenic galore systems arsenic possible, gathering the information pipeline, creating the information lake. Next year, says Sethi, it volition commencement commercializing the exertion furniture with its built-in intelligence.

If gathering a predictive analytics level for exigency effect sounds daunting, Sethi says really selling it is tougher, and he’s proudest of HEN’s traction connected that front.

“The hardest portion of gathering this institution is that this marketplace is pugnacious due to the fact that it’s a B2C play erstwhile you deliberation of convincing the customers to buy, but the procurement rhythm is B2B,” helium explains. “So you person to truly marque a merchandise that resonates with radical – with the extremity idiosyncratic –  but you inactive person to spell done authorities purchasing cycles, and we person cracked some of those.”

The numbers carnivore this out. HEN launched its archetypal products into the marketplace successful the 2nd 4th of 2023, lining up 10 occurrence departments and generating $200,000 successful revenue. Then connection started to spread. Revenue deed $1.6 cardinal successful 2024, past $5.2 cardinal past year. This year, Hen, which presently has 1,500 occurrence section customers, is projecting $20 cardinal successful revenue. 

HEN has competition, of course. IDEX Corp, a nationalist company, sells hoses, nozzles, and monitors. Software companies similar Central Square service occurrence departments. A Miami company, First Due, which sells bundle to nationalist information agencies, announced a monolithic $355 cardinal round past August.  But nary institution is “doing precisely what we are trying to do,” insists Sethi.

Still, Sethi says that the constraint isn’t request – it’s scaling accelerated enough. HEN serves the Marine Corps, US Army bases, Naval atomic labs, NASA, Abu Dhabi Civil Defense, and ships to 22 countries. It works done 120 distributors and precocious qualified for GSA aft a year-long vetting process (that’s a national seal of support that makes it easier for subject and authorities agencies to buy).

Fire departments bargain astir 20,000 caller engines each twelvemonth to regenerate aging instrumentality successful a nationalist fleet of 200,000, truthful erstwhile HEN is qualified, it becomes recurring gross (is the idea), and due to the fact that the hardware generates data, gross continues betwixt acquisition cycles.

HEN’s dual extremity has required gathering a precise circumstantial team. Its bundle pb was formerly a elder manager who helped physique Adobe’s unreality infrastructure. Other members of HEN’s 50-person squad see a erstwhile NASA technologist and veterans from Tesla, Apple, and Microsoft. “If you inquire maine method questions, I would not beryllium capable to reply everything,” Sethi admits with a laugh, “but I person specified bully teams that [it] has been a blessing.”

Indeed, it’s the bundle that hints astatine wherever this gets interesting, due to the fact that portion HEN is selling nozzles, it’s amassing thing much valuable: data. Highly specific, real-world information astir however h2o behaves nether pressure, however travel rates interact with materials, however occurrence responds to suppression techniques, however physics works successful progressive occurrence environments.

It’s precisely what companies gathering alleged satellite models need. These AI systems that conception simulated representations of carnal environments to foretell aboriginal states necessitate real-world, multimodal information from carnal systems nether utmost conditions. You can’t thatch AI astir physics done simulations alone. You request what HEN collects with each deployment.

Sethi won’t elaborate, but helium knows what he’s sitting on. Companies grooming robotics and predictive physics engines would wage handsomely for this benignant of real-world physics data.

Investors spot it, too. Last month, HEN closed a $20 cardinal Series A round, positive $2 cardinal successful task indebtedness from Silicon Valley Bank. O’Neil Strategic Capital led the financing, with NSFO, Tanas Capital, and z21 Ventures participating. The circular brought the company’s full backing to much than $30 million.

Sethi, meanwhile, is already looking ahead. He says the institution volition instrumentality to fundraising successful the 2nd 4th of this year.