Founder Q&A: 5 Questions With Atombeam’s Charles Yeomans

It’s safe to say Atombeam has resonated with the crowd here at StartEngine.

A growing community of more than 10,000 shareholders has invested over $35 million to date into the data compaction firm — including three straight sellouts.¹

And Atomeam’s current campaign has already raised $2.8 million from everyday investors like you.

So what is it about Atombeam that’s impressed the crowd so much?

We decided to ask Founder and CEO Charles Yeomans. With over 30 years of experience in both executive management and investment banking, Charles has been the CEO or COO of three companies, and helped launch two of the nation’s largest insurance brokerages.

Charles and our team discussed the company’s vision of real-time intelligence for machines, and how the crowd has helped fuel that mission.

See below to read more, or explore Atombeam’s latest campaign.

This Reg A+ offering is made available through StartEngine Primary, LLC, member FINRA/SIPC. This investment is speculative, illiquid, and involves a high degree of risk, including the possible loss of your entire investment. For more information about this offering, please view Atombeam’s Offering Circular and Related Risks.

If Neurpac and PCM (Persistent Cognitive Machine) become widely adopted, how could they change the economics of AI, data infrastructure, and connected devices?

They attack different problems, but both are economic problems at their core.

Neurpac changes the cost of moving data.²

Every connected device — every vehicle, every IoT sensor, every satellite, every server in a data center — transmits data over a link that someone pays for. Cellular, satellite, carrier-provisioned fiber — it’s all priced per byte or per unit of bandwidth. Neurpac reduces that data by an approximate average of 75%, losslessly and in real time.

For a single connected vehicle, that’s the difference between a $150/year cellular bill and a $37.50/year cellular bill. For a fleet of 10 million vehicles, it’s the difference between $1.5 billion and $375 million. For a mid-sized utility with 2 million smart meters on a cellular network, we estimate 
Neurpac could potentially unlock a 75% reduction in data volume that could save $9-36 million annually, or allow the utility to quadruple data frequency at the same cost.

Multiply by the number of facilities, fleets, or satellite terminals in the world, and you begin to see the scale. The hyperscalers are spending $690 billion on data center infrastructure in 2026, and the single biggest bottleneck is moving data through the network fast enough. We believe Neurpac can make every existing link significantly more productive without building anything new.

PCM changes the cost of knowing things.³

Current AI systems — large language models — store knowledge as frozen weights in neural networks. They cannot learn new information without expensive retraining. They cannot tell you whether what they’re saying is grounded in evidence or fabricated. And their memory requirements grow linearly with the amount of information they ingest.

PCM is designed to store knowledge as geometric structure on a mathematical manifold, where memory scales logarithmically — meaning it should be able to absorb exponentially more experience for the same compute and storage.

It's intended to learn continuously without retraining.

And our latest architecture, PCM 2.0, includes a three-layer hallucination resistance system designed to mathematically detect when reasoning has gone wrong, distinguish between “I need more evidence” and “the evidence contradicts itself,” and refuse to consolidate conclusions that aren’t supported. That’s not a filter on top of a model. It’s built into the geometry of how the system thinks.

If both are adopted at scale, we think the combined effect could be significant: the data that moves through networks is 75% smaller, and the AI systems that process it are potentially orders of magnitude more memory-efficient, continuously learning, and architecturally resistant to hallucination. We believe those are foundational cost curves in the infrastructure economy.

Explore Atombeam’s Current Campaign »

This Reg A+ offering is made available through StartEngine Primary, LLC, member FINRA/SIPC. This investment is speculative, illiquid, and involves a high degree of risk, including the possible loss of your entire investment. For more information about this offering, please view Atombeam’s Offering Circular and Related Risks.

***

You’ve secured $2.4+ million in U.S. DoD contracts, a major commercial deal with Trilliant, and partnerships with NVIDIA, L3Harris, Ericsson, and others. What does it take to earn the trust of organizations at that level?

It takes the technology actually working. That’s the price of entry, and there is no substitute for it.

The Space Force didn’t give us a contract because of a pitch deck. They tested Neurpac on their satellite data, measured 77% average compaction, and decided the results justified a $776,000 extension, increasing the contract to a $1.98 million total Space Development Agency agreement.

Trilliant didn’t sign a reseller agreement because we asked them to. They evaluated Neurpac against their 40-million-device smart grid network and concluded it was instrumental to enabling what they call “Grid Intelligence” — their term, not ours.⁴

L3Harris, Ericsson, Intel, NVIDIA, HPE — these companies have teams whose job is to evaluate technology. They are not easily impressed and they do not partner casually.⁴

Atombeam’s partnership relationships vary between companies. See footnote 4 for more details.

But there’s a second part that matters almost as much: intellectual property. We hold 165 issued patents and 249 pending applications across Neurpac, PCM, and other technologies. When a defense prime or a Fortune 500 company evaluates a technology partner, they need to know the technology is protected — that a competitor can’t simply replicate it.

We believe our patent portfolio gives them that assurance. It’s the reason we can have conversations about IP core licensing and silicon integration without worrying about giving away the store.

Explore Atombeam’s Current Campaign »

This Reg A+ offering is made available through StartEngine Primary, LLC, member FINRA/SIPC. This investment is speculative, illiquid, and involves a high degree of risk, including the possible loss of your entire investment. For more information about this offering, please view Atombeam’s Offering Circular and Related Risks.

***

After raising $35+ million from 10,000+ investors, including three straight sellouts, why do you think Atombeam has resonated so strongly with the crowd?

I think investors recognize a real technology company when they see one.

We don’t have a pitch that requires you to suspend disbelief. We have production deployments processing millions of transactions daily. We have government contracts with serial numbers on them. We have a patent portfolio that we believe would take a competitor a decade and tens of millions of dollars to replicate. And we have two technology platforms — Neurpac and PCM — that we believe address problems large enough to justify significant company valuations independently.

I also think our investors appreciate that we communicate honestly. We don’t overstate what we’ve accomplished. When PCM is at TRL 4–5 and running on prototype hardware, we say that — we don’t pretend it’s shipping. When we are excited about a big development, we share it, but we also share disappointments.³

Investors are smart enough to evaluate a company on facts, and we respect them enough to provide those facts without embellishment.

Explore Atombeam’s Current Campaign »

This Reg A+ offering is made available through StartEngine Primary, LLC, member FINRA/SIPC. This investment is speculative, illiquid, and involves a high degree of risk, including the possible loss of your entire investment. For more information about this offering, please view Atombeam’s Offering Circular and Related Risks.

***

What has crowdfunding enabled you to accomplish that might not have been possible otherwise?

Crowdfunding gave us the ability to build a real company without giving up control to investors who might have pushed us to optimize for short-term metrics at the expense of long-term value. That matters enormously when you’re developing fundamental technology.

Neurpac was not an obvious bet in 2017 when Atombeam was founded. Data compaction as a category didn’t exist. The idea that you could build data-type-specific codebooks that would outperform general-purpose compression by 5–10x on structured data — most VCs would have passed on that.²

Crowdfunding let us build the technology, prove it worked, get it deployed commercially, win government contracts, and build a patent portfolio — all while maintaining the ability to pursue PCM as a second platform without having to justify it to a board that wanted us to focus exclusively on near-term revenue.

PCM is a good example. A VC-backed company at our stage would almost certainly have been told to shelve a fundamental AI research platform and focus entirely on selling Neurpac. That would have been a rational short-term decision and, we believe, a strategic mistake.³

PCM is now at TRL 4–5 with a prototype running on NVIDIA hardware, a Navy demonstration planned for April, a DARPA proposal submitted and significant progress underway. If it works as the mathematics suggest it should, we think it could ultimately be even more valuable than Neurpac.

Our investors gave us the freedom to pursue both, and I believe that decision will be validated.

Explore Atombeam’s Current Campaign »

This Reg A+ offering is made available through StartEngine Primary, LLC, member FINRA/SIPC. This investment is speculative, illiquid, and involves a high degree of risk, including the possible loss of your entire investment. For more information about this offering, please view Atombeam’s Offering Circular and Related Risks.

***

What do you believe is the strongest reason for investors to join your latest campaign?

Timing. Everything seems to be falling into place at Atombeam.

• Neurpac is commercially shipping, deployed in production, and generating revenue. We just signed Trilliant as a reseller into the smart grid industry. We have a joint solution with Digital Barriers that unlocks the video surveillance and oil and gas markets we couldn’t serve alone. We’re actively selling into data centers at a moment when the hyperscalers are spending $690 billion and can’t build fast enough. There are numerous companies evaluating Neurpac right now, including a big computer maker and one of the three major cell phone carriers.²
• PCM is at the point where we’re demonstrating it to the Navy, briefing DARPA, and presenting to Gartner analysts. The PCM 2.0 architecture includes what we believe may be the first mathematically grounded hallucination resistance system in AI — not a filter, not a guardrail, but geometric structure built into how the system reasons. If the April Navy demonstration goes well, we believe it could materially accelerate the path to a PCM-specific government program of record.³
• The patent portfolio — 414 issued and pending across both platforms — is a durable asset regardless of which specific deployment path moves fastest. The team is 55 people, 30+ of them engineers. The technology is real, the customers are real, the contracts are real, and we believe the market is actively looking for what we built. That’s not a bet on potential. We think it’s a bet on execution at a moment when the execution is visibly working.

Explore Atombeam’s Current Campaign »

This Reg A+ offering is made available through StartEngine Primary, LLC, member FINRA/SIPC. This investment is speculative, illiquid, and involves a high degree of risk, including the possible loss of your entire investment. For more information about this offering, please view Atombeam’s Offering Circular and Related Risks.

1. The amount raised may include insider investments, which may go toward meeting the minimum offering amount. Atombeam has previously crowdfunded $32,648,634, including Reg A+ and Reg CF offerings conducted on other platforms and on different terms from the current offering. Please see the offering document for additional information.

2. Neurpac’s power efficiency projections are based on prototype testing and theoretical modeling. Actual results may vary significantly from these estimates. Statements about potential revenue or commercial success are forward-looking and involve risks and uncertainties. Actual results may differ materially from any projections.

3. The PCM technology is still in development, and there are substantial technical risks that could prevent us from achieving these efficiency gains at scale. Competitive advantages in technology are often temporary, and competitors may develop alternative approaches that match or exceed our efficiency claims. Market adoption of power-efficient AI is not guaranteed, and regulatory, technical, or economic factors could impact the viability of our approach. This technology has not been validated in large-scale commercial deployments, and significant engineering challenges remain before commercial release. Investors should consider this a high-risk, early-stage technology investment with uncertain outcomes.

4. The partnership relationship varies between companies and can include the following: inclusion on a preferred vendor list, invitations to participate in certain forums, listed on the other company's website, and introduction and networking opportunities.

MAR-13974