Atombeam is a company at the cutting edge of data reduction for IoT and data storage applications. Pulse 2.0 interviewed Atombeam co-founder and CEO Charles Yeomans to learn more about the company.
Charles Yeomans’ Background
What is Charles Yeomans’ background? Yeomans said:
“I’m a former U.S. Navy intelligence officer who went on to hold leadership roles in investment banking, insurance brokerage and financial services companies before diving headfirst into managing technology companies. Before founding Atombeam, I led a biotechnology company, Trigemina, a Stanford Medical School spinoff for which I secured funding, managed and later sold successfully. I have an AB from Kenyon and an MBA from Stanford, and currently reside in Orinda, California.”
Formation Of Atombeam
How did the idea for Atombeam come together? Yeomans shared:
“The idea for what would become Atombeam actually came together through connections made during a Boy Scout troop. My co-founder is a long-time technology product specialist with extensive experience in data storage and management at some of the largest technology companies in Silicon Valley. His son and my son were in the same Boy Scout troop, and that’s how we initially met. I was the business guy and he was the technology guy, a pretty classic combination.”
“We soon realized that we had some revolutionary technology on our hands – technology that reduces the size of data in a way that hadn’t been done before. The idea of using machine learning to build codebooks leverages the speed of computers to create a technology that is a truly new, ultra-efficient approach to the growing data problem. The basic idea is to have small codewords represent much larger patterns machine learning finds in a sample data set. Not only does this enable an average of 4x more bandwidth for IoT, but it also can encrypt the data in the same light computing step, and the data can be searched and accessed in its reduced form. A simple idea, but with revolutionary implications.”
“Before long, we were talking to data scientists and patent attorneys, and had founded Atombeam. After we raised our initial funding, we quickly added Josh Cooper as Atombeam’s Chief Scientist. Josh is a brilliant mathematician and data scientist who was also a professor at the University of South Carolina. We quickly started filing patents and interviewing customers to focus our product market fit, and began building early software prototypes.”
“Several major companies are now working and partnering with us to utilize our innovative, game-changing Data-as-Codewords technology.”
“As CEO, I’m responsible for guiding the company through the development of our products, helping to ensure that our technological advances translate to solving real-world industry problems. I’m also responsible for building a world-class team that can both deliver on the promise of our technology today, and shepherd its application to new industries – and to solve future challenges.”
“I’m also the primary driver for the company’s funding, something I’m especially proud of, as we’ve chosen to crowdfund the company instead of seeking the involvement of institutional investors, at least up to now. I strongly believe that crowdfunding gives ordinary investors the opportunity to back innovative companies directly and participate in what could be a significant upside – and, it gives companies like ourselves the opportunity to focus on building a great company, not managing a board of directors made up of your biggest investors, all which have their own opinions and ideas for what we should be doing. In a real sense, it is like being a public company without as great of a regulatory burden or focus on quarterly results; we can keep our eyes on the long term.”
“All in all, we prefer to spend our time and effort on building the best product possible to solve the world’s growing data problem. The opportunity before us is very big.”
Favorite Memory
What has been your favorite memory working for the company so far? Yeomans reflected:
“That is easy. Bringing in the team we have built this year and seeing it build a real product that big companies now are piloting in the field. And those companies are excited about it. There is literally no other way to solve their data congestion problem in a way that works for them, and both our end-user customers and our channel partners, including Viasat and Ericsson, see a win-win for everyone. Imagine buying a gateway that has 4x more bandwidth than the specifications, and can add encryption, but still makes the data 75% smaller; that is the power of Atombeam’s Neurpac. And it is really cool to see it happen for real customers in the real world.”
Core Products
What are Atombeam’s core products and features? Yeomans explained:
“Here is the key to understanding why we think our approach is revolutionary: all three products work together to make data far more informationally dense, like compression does, but unlike compressed data, the Atombeam-densified data is still searchable, readable and usable. This can drive massive efficiencies for computing systems. Imagine what it can mean if you could squeeze the same usable information into a quarter of the data – computing and network speed could quadruple. And you can do so with just a software update, no new hardware. Atombeam can impact every connected device, and the network that connects it. That is why we are winning prizes for technology innovation.”
“Atombeam’s technologies use what we call “Data-as-Codewords” technology, which brings a new, more efficient (and more secure) approach to data. It is a fundamental rethinking of how data is structured at the most basic level. Since the advent of digital computing, computers generate data essentially the same way that we translate English into French, except in computing you are translating data into binary – but computers do not need all of that human contextualization and repetition. That is the code we cracked with Data-as-Codewords. It crams the same information into a quarter of the data, and consequently makes computers and networks 4x more efficient.”
“Neurpac is the first product, and it works for situations for which you need exact, 100% replication of the original data on the receiving end, like telemetry and tracking data. The difference is you need, on average, 75% less data to send the same information – codewords, versus the original data. You might say, compression can do that too, but not for the little messages machines send, IoT data, plus that kind of data is tough to encrypt, so 98% is just sent in the clear. Plus if you compress data it is unusable, unless you decompress it first. Add it all up and you see why Neurpac is such a different technology than compression. That is why so many companies (and the U.S. Department of Defense) are piloting the technology now.”
Challenges Faced
What challenges have Yeomans and the team face in building the company? Yeomans acknowledged:
“The crazy thing is, when we first started talking about Neurpac, almost no one believed us. Or they thought it was some kind of compression variant. One guy said, oh, that’s just the audio compression I was using in the 1970s(!). But then they test it on their own data, and their mouths drop open. So the cool thing is, you just need to take it out for a spin and see for yourself. As we have become better known, this problem seems to be going away, though, and I think we are better at explaining how it works. Engineers and data scientists get it right away, plus we have big partners who are promoting it, so now we are starting out from a different place.”
Evolution Of The Company’s Technology
How has the company’s technology evolved since launching? Yeomans noted:
“The best part is how our technology continues to evolve. Neurcom is our second product based on Data-as-Codewords, and is still in development under a contract with the U.S. Air Force. It is also about making data informationally dense, in this case images and video. Neurcom uses a neural net AI that integrates into state of the art video compression algorithms to enhance their performance. Like Neurpac, Neurcom is about packing more information into data. Compared to conventional compression, Neurcom can double effective network bandwidth for image and audio files, while maintaining equal quality.”
“Neurcom’s capabilities are particularly valuable for applications in constrained environments, such as satellite communications and defense systems, where bandwidth is at a premium. Neurcom addresses a critical need in sectors dealing with high-volume image and audio data transmission, including aerospace, telecommunications, and media industries.”
“Our third product is a new approach to generative AI that we call the Lightweight Codeword Model, or LCM. The LCM exploits the informationally dense data generated by Neurpac and Neurcom to make generative AI faster and lighter, so it can train faster, use less power and be able to operate on the edge, all major goals of generative AI systems. Not only that, but because the LCM uses Neurpac and Neurcom data in the model, it is natively multimodal, using data from any source – including raw sensor data, genomic information, and time-series data – to perform its calculations, thus eliminating the need to pre-process and post-process data into the “language” that LLMs require.”
“By emulating human learning processes at an unprecedented scale and speed, LCMs could adapt and improve their predictions in real-time. This breakthrough could yield AI systems that are not only more versatile and robust but also capable of navigating the intricacies of complex, real-world scenarios across a multitude of industries.”
“Added together, Atombeam’s three technologies represent a completely new approach to the structure of data – a far more efficient and secure way that, from end-to-end, can drive massive efficiencies in computing and networking. Imagine a data lake with petabytes of stored data that was not only reduced in size in a similar proportion to compression, but is searchable and randomly accessible. Not only that, it is secure while it is stored – a hacker can steal data, but it will be unusable without the codebook used to generate it.”
Significant Milestones
What have been some of the company’s most significant milestones? Yeomans cited:
“We’ve secured multiple active contracts with the United States Department of Defense, and have forged partnerships with technology giants such as Hewlett Packard Enterprise, Intel, Nvidia, Viasat, Ericsson, Microsoft and Amazon. These collaborations underscore the wide-ranging applicability and potential of our technology across various sectors. Ericsson, in particular, is interested in leveraging Atombeam’s technology to enhance the effective bandwidth of their Cradlepoint gateways. If you are Ericsson, and your gateways can send 4x more data than your competitors, that is a pretty strong competitive advantage.”
“We have also built up an excellent patent portfolio, with 84 issued and allowed patents and an additional 115 pending. We have truly new, highly valuable intellectual property and have put real resources behind protecting it.”
Customer Success Stories
When asking Yeomans about customer success stories, he highlighted:
“In addition to our work with the Department of Defense in various applications aimed at improving the transmission of data in real-time, we’ve started working with Viasat, a major satellite operator, and one of their oil company customers who needed to collect and send more data from its remote operations than was possible over Viasat’s (or other satellite provider) networks.”
“The oil company currently needs to send about 20% more data than was possible over satellites, and in the near future they want to send almost four times more data than they are sending now. They needed this data flow to optimize production and for preventative maintenance, and until Atombeam’s technology was brought to bear, they had no answers. We reduced their data load by 77%, making it possible for them to send all of the data they need now – and to also cut their data transmission costs. Atombeam’s technology even enables them to future proof their data requirements. They are excited that there is a solution that works flawlessly in the field – and are moving to deployment.”
Funding
When asking Yeomans about the company’s funding details, he revealed:
“We just completed a Reg A+ funding round with StartEngine on January 30, 2025, and have successfully raised $20 million, one of the fastest crowdfunding raises ever. This builds upon the $15 million previously secured through a combination of family offices, angels and crowdfunding. We are garnering interest from a number of institutional funds and corporate investors as well.”
Total Addressable Market
What total addressable market (TAM) size is the company pursuing? Yeomans assessed:
“For just Neurpac, we estimate that the technology is applicable to virtually any connected IoT device, which IoT Analytics estimates to be 18.8 billion in 2024. Our pricing is based on heavy data users for now, but over time we plan to go down market and enable even very limited devices. We also plan to make Neurpac available for data centers to enable far greater access to stored data for their customers.”
“Neurcom initially will focus on specialized applications like images transmitted from satellites, and may in the longer term be able to address a broader market opportunity, such as video and optical images. The LCM could be a very big market, but it is difficult to determine the dimensions of the market at this fairly early stage.”
“Overall, Atombeam estimates its TAM as at least $60 billion, and likely higher.”
Differentiation From The Competition
What differentiates the company from its competition? Yeomans affirmed:
“There are no other companies addressing the problem of data size and transmission in the way that Atombeam is. And we’re not just solving current industry problems; we’re redefining how data in the broadest sense is approached altogether. By challenging the fundamental assumptions and approaches about data structure and processing, we believe we can establish a new standard for efficiency, security, and versatility in data handling.”
“The standout feature of our innovative Neurpac technology is its capability to make data searchable even in its minimized form. Unlike traditional compression methods that sacrifice accessibility for size reduction, Atombeam’s technology uniquely preserves data searchability in encoded, minimized form. By enabling quick and efficient searches of massive data lakes, Atombeam has the potential to help companies unlock valuable insights from previously underutilized data stores owned by companies and governments. It is efficient, and green; data can be encrypted and compacted in a cheap edge device, such as a sensor, then transported to a data center and stored. Sending less data saves power and bandwidth, and encryption is ultralight and built in, avoiding the need for special hardware and software. There is also no need to decrypt or compress the data at the data center. This is the future of data.”
“Looking forward, if successful, this sort of paradigm shift has the potential to drive a whole new era of technological advancements.”
Future Company Goals
What are some of the company’s future goals? Yeomans concluded:
“Initially, our focus is in serving the needs of our DoD and commercial customers with our first released product, Neurpac. We anticipate ramping our revenue later this year and early next year, and will also be devoting significant resources to our other two products, Neurcom and the LCM. We both hope to turn both into commercial products next year.”
“We also plan to build the company’s infrastructure. We think we have a tremendous opportunity before us, but also a tremendous challenge. We need to continue to recruit ‘A’ players in technology, marketing, sales and administration to create a world-class company. That is the only way we can realize the potential of these groundbreaking technologies.”
Additional Thoughts
Are there any other topics you would like to discuss? Yeomans concluded:
“I would like to put a spotlight on the core of what Atombeam is about. This is not about a better compression, this is a complete overhaul of data from the ground up. Instead of all of these steps you need to undertake now – compressing, encrypting, decrypting, trying to figure out where the data you need is stored, decompressing, and on and on – with Atombeam’s technologies, that is all yesterday’s news. We eliminate the extra hardware, computing, power use, and human time waste as well. We can encrypt and compact data at the edge with a ten cent processor, send it more efficiently and securely, store it more efficiently and securely, and never perform any additional processing of the data until you need it.”
“Unlike conventional stored data, if it is stolen by a hacker, Neurpac encoded, compacted data will be virtually impossible to access without the codebook that was used to encode it. And you can find it when you need it easily, since the data is searchable when Neurpac encodes it, a radical difference from compressed data.”
“Finally, the icing on the cake is the LCM, which takes maximum advantage of the informationally dense structures of Neurpac and Neurcom to make generative AI at the edge possible, shortening training times and providing native multimodality.”
“This is a truly new paradigm for data, the centerpiece of why we have networked computing devices, making them massively more efficient and secure.”
