SHI Fall Summit 2025 | Brent Willis, Prohawk AI

Clips
More from SHI Fall Summit 2025

Brent Willis

Co-Founder & COO

ProHawk AI

play_circle_outline ProHawk AI offers real-time pixel-by-pixel computer vision technology for safety and visibility.

play_circle_outline Transforming Industries: How Technology Enhances Safety and Efficiency in Manufacturing, Healthcare, and Transportation Through Anomaly Detection

play_circle_outline Enhancing Real-Time Video Analytics: ProHawk’s Strategic Partnerships with HPE and NVIDIA for Comprehensive AI Solutions

Info
Transcript

Brent Willis, Prohawk AI

Brent Willis

Co-Founder & COO ProHawk AI

At the SHI Fall Summit, Brent Willis, president, co-founder and chief operating officer of ProHawk AI, joins theCUBE’s Paul Gillin to unpack how “physical AI” is restoring degraded sensor data in real time to improve safety and decision-making. Willis explains ProHawk’s pixel-by-pixel computer vision approach that converts incoming video into 33 million pixels per second at the extreme edge – optimizing each pixel without model training so operators immediately see actionable information. He details why ProHawk “restores” rather than “enhances” imagery to mee... Read more

explore Keep Exploring

What is the role of the physical AI interface in processing incoming data from various sensors? add

What is a real-life application of your technology? add

What role does the NVIDIA chip play in the capabilities of the solution being discussed? add

bolt Powered by CUBE AI

Brent Willis, Prohawk AI

search

Paul Gillin

>> This is theCUBE. I'm Paul Gillin. We're here in Somerset, New Jersey for the SHI Fall Summit 2025. We're talking all about AI here and all about good uses of AI for the public benefit. And one of those examples is what my next guest is doing with his company, ProHawk. Let's welcome President, Co-founder, and COO of ProHawk AI.

Brent Willis

>> Thanks, Paul. Happy to be here.

Paul Gillin

>> So what does ProHawk do?

Brent Willis

>> We are the only real-time pixel-by-pixel computer vision technology or computer vision management system in the world. And what that translates to in real language is we take the risks and obscurity of night and we convert them to the visibility and safety of day to help save lives and to make critical decisions at the critical point in time.

Paul Gillin

>> And you do this in real time with images from all kinds of different sources and all kinds of different quality. Is that correct?

Brent Willis

>> Yeah, we're not really the best in the world at it. We're just the only ones in the world that had it, that can take that degraded video or degraded data and then make something actionable and usable out of it. And think about AI, right? Most people can't spell it, and they don't know that there's all sorts of different types of AI. There's agentic AI, generative AI, quantum AI. We're physical AI and for HPE at the initial interface. So we take that incoming data from any sensor, radar, LiDAR, MRI, CT scan, infrared or optical, and right after that comes into any feed or device, from that sensor, we restore that. So the user, real time without being in the loop at all, sees actionable information. And then there's a conduit and a pass down to any analytics or any agentic AI or any of the things that customers are going to use in AI in the future. So this is the starting point really in AI where customers can make a difference for productivity today. Not in six months, not in six years, not in six days. In six hours if they implement it. And it's ready to go, we're making a difference around the world already.

Paul Gillin

>> This is a neural network technology, is that correct?

Brent Willis

>> Well, yes and no because what we're doing is we take any incoming video feed and we convert that instant feed into 33 million pixels, optimize each individual pixel, the two green, one red, one blue, the concentric pixels all around that, and then we just do that 33 million times every second. So we don't need any learning or networks, or anything. We do this at the edge of the extreme edge locally and take that data and then optimize each one of those pixels so you can get that unmatched clarity and actionable data such that all gen AI applications are going to work better from the future. But from a learning standpoint, what we do now is we take that incoming data, we approve it, we do the analytics, and then we can circle back to teach the sensor and to teach the computer to work better next time to make everything more efficient.

Paul Gillin

>> So make this real. What's a real-life use case where we might see your technology at work?

Brent Willis

>> Well, the problem for us is we work horizontally for HPE as part of the Unleash AI initiative across all industries, from energy to transportation, to healthcare, to federal, to safety and security, you name it. But one potential solution is, take manufacturing. At the end of every manufacturing line, they have an optical sensor. And that optical sensor typically sees the finished good, or let's say in the beverage industry, which I spent a lot of time in, fill heights, as a metric of safety. Now what we can do instead of just that optical sensor, we can see through all of the liquid to see if there's any particulate matter in there or any kind of anomaly, or, if it's a glass bottle, any kind of glass in there that doesn't meet standards for consumer safety, and meeting the right consistency and tolerances. And that kind of technology never existed before, but now it does with ProHawk, as we can see through all of those details pixel by pixel, in this case for consumer safety,

Paul Gillin

>> You can literally turn a picture taken at night into a daytime photo.

Brent Willis

>> Yeah, it's kind of Hollywood stuff, but yes. And it's just math, right?

Paul Gillin

>> Take fog out of an image, a foggy image.

Brent Willis

>> And think of maritime use cases> so you get sun glare in the morning, that's hard to take out. You get fog or rain in the afternoon, and then night half the time. This is the only technology that overcomes all of those environmental obstacles. So you can have better efficiency and safety in harbor management or those kinds of things, or ships on the water. Or think autonomous driving. Because you get such visual clarity, it allows us to see 20 times farther on average and 80 times faster. And again, pixel by pixel, so I can take out every single raindrop, we take out every single snowflake, every single piece of particulate matter. So the autonomous driving algorithms can see objects or what they need to do 20 times farther, 80 times faster. So actually autonomous driving will work in degraded environmental conditions, and typically they don't. So the use cases just go on and on. And that's the challenge and the benefit is why it's so important for HPE and NVIDIA as one of these Unleash AI partners is because we work so well with the Kamiwazas of the world on the agentic AI side and some of the generative AI companies, all as part of that HPE Unleashed ecosystem underpinned by HPE infrastructure and NVIDIA infrastructure to make it all work in a turnkey solution.

Paul Gillin

>> Tell us about your own journey to this company. Can tell by your sweater that you are a West Point graduate.

Brent Willis

>> Don't hold it against me, Paul.

Paul Gillin

>> Military career, worked in consumer goods all over the world. How did you decide that this was what you wanted to do?

Brent Willis

>> It's hard. I love serving the country and making a difference for the country and the military. And this can make a real difference also, this technology, because our mission is to save lives by making the world a better and safer place. And I used to be president for the Coca-Cola company for about half of Latin America, and I ran a small little beer company in Belgium. Or I was the number two at a small little beer company in Belgium. And being a West Point guy, we were kind of into global domination. I am too. And we figured out the economics in the beer industry, and we became the world's largest beer company in three years. But it did take three years. But one of my negatives that I was just embarrassed about is we put in so much plastics and plastic bottles in the environment, and the size of the great Pacific Garbage Patch now is the size of 25 Floridas. And it's getting worse because 92% of all plastics-

Paul Gillin

>> Wow....

Brent Willis

>> end up in landfills or the ocean. It is a huge, huge problem, and it's not getting better. So I just felt so... I was proud at the time to be in leadership positions in these big companies, and today I'm just embarrassed and I want to do better. And I just know we can make a huge difference for the planet with this technology. And not a difference in the long-term, a difference today across manufacturing, across maritime, across federal, across safety and security. It's just a fantastic technology that will make an impact on society that people just need to know more about. That it is available and available now and very inexpensive, and turnkey ready to go underpinned by all the resources of TD SYNNEX, HPE, and NVIDIA.

Paul Gillin

>> I want to pick up on something you said earlier about improving the quality of training models, how your technology contributes to generative AI training, agentic AI training. How so?

Brent Willis

>> Well, the negative for all of those technologies is if you get garbage in, you get garbage out. And so the achilles heel for all of AI is quality of data. That's the first thing we fix. We restore that incoming data, so all of the other AI models and AI components are going to work and work better and work more efficiently. But what we also do now is we have what we call the AI flywheel, so we restore that incoming video. And the reason why we say restore versus enhance is because we're actually restoring actuals. Enhance implies changing the evidence.

Paul Gillin

>> Changing. You're not changing anything.

Brent Willis

>> If you change evidence, it is not admissible in a court of law. So we're actually restoring actuals to meet the Daubert standard for all of the law enforcement and forensic applications. But what we do now is we cycle back again every three milliseconds and teach the models and teach the sensors and learn the model such that we can do all of the identification, and the faster and more efficient improvement of the data and the videos going forward before we pass it down to the agentic AI systems.

Paul Gillin

>> You mentioned you're the only one who does what you do. You also have some important partnerships. You're part of the HPE AI program, allied with NVIDIA. What are some of the key alliances you have?

Brent Willis

>> Well, the truth is we couldn't do what we do on our own. We could never get this real-time three millisecond benefit without the NVIDIA chip. And any NVIDIA chip, so it doesn't have to be the latest. But any NVIDIA chip, we couldn't do it without them. And even though we're a big part of the whole video management solution, customers want and need more, right? This is a way to get started with AI, but they need and want more. They want the agentic AI. And so we're such an easy fit and handoff to any video analytics companies. Any other partner, whether it's BigID or Kamiwaza on the agentics AI side, and it's kind of an ABC. And what HPE has done is integrated it all, picked the best in world out there, and ProHawk just got lucky to be in there because we were the only one. Not best in the world, just the only one at what we do. And HPE has taken all the guesswork out of it for customers to pick who are those best in the world, to pull it together to say, "Here is your turnkey solution," and nobody else has done it. And they're fast in doing it and really effective, and low cost for the customers as well.

Paul Gillin

>> And why are you at an SHI conference?

Brent Willis

>> SHI is the interface to the customers. Even though HPE and other OEMs, and the major distributors like TD SYNNEX, sell and distribute. The key interface with the customers is groups like SHI, and they're really, really good because they've got their relationships and they've got those interfaces directly with those customers, and they can translate all of this AI stuff into something that somebody can spell. I know it's only two letters, but-

Paul Gillin

>> It's not day one.

Brent Willis

>> Yeah, a lot of the customers still are just learning about it. I don't think SHI, HPE, any of us really know much about AI today versus what's going to happen. We just know more than anybody else and know where we can make a difference today. And SHI really knows where to make a difference across industries with customers today. So that's why I'm here with SHI. They're so good and one of the best in the world, and one of the key priorities for both HPE and NVIDIA, key customer partner ecosystem priority, is to drive the implementation of AI to make a difference.

Paul Gillin

>> And that's what you're doing with ProHawk AI. Brent Willis, doing good with AI and making a difference. Thank you so much for joining us here on theCUBE.

Brent Willis

>> Thank you, Paul.

Paul Gillin

>> I'm Paul Gillin. This is theCUBE, the source for technology news and analysis. We'll be back.