Hi Everyone,

IoT Analytics put out a survey on June 15 of what maintenance vendors showed at the Dortmund and Hannover trade fairs this spring. It is framed around the idea that the trillion-dollar industrial downtime problem is turning into a knowledge problem. It is worth reading, because a lot of what the analysts and vendors describe as the hard problems in the space are problems Neurpac is designed to solve.* They are not writing about us, which is part of why it is useful. It is an outside read on where the demand is.

The knowledge angle is really PCM. The rest of the article is Neurpac.

The argument runs like this. Prediction is largely a solved problem now, but the people who actually know how to fix the machines are retiring and taking undocumented know-how with them. AWS told the conference it expects 65% of maintenance teams to lose critical knowledge. Cognite said junior technicians take three to three-and-a-half times longer to diagnose and repair than the people they are replacing. That problem, holding onto what the expert knew and giving the next technician something trustworthy to work from, is what PCM is for, and I will come back to it.** What I want to flag first is that most of the article is not about that. It is about the sensors, the links, and the deployment setup that all have to work before any AI ever sees the data, and I believe that is Neurpac’s territory.

Wireless sensing keeps running into the battery-versus-data tradeoff.

IoT Analytics calls wireless vibration monitoring one of the fastest-growing parts of the smart maintenance market. It has passed $1 billion in size, with 30%+ growth projected for 2026. And it names the central engineering problem plainly: you are constantly trading measurement frequency against battery life. Every vendor in the segment is wrestling with the same thing. SKF’s Enlight Collect sensor only gets four to five years of battery by sleeping most of the time and waking once a week for a single full measurement. El-Watch sensors, which Status Pro promotes, get ten years by taking readings every couple of minutes and sending them over low-power 868 MHz radio. What is actually constraining these devices is not the sensing. It is the energy it takes to move the bytes off the device over a slow radio.

That is the part Neurpac changes. In a battery-powered sensor the radio is the main drain, and how long the radio stays on depends on how many bytes it has to send. Neurpac cuts structured telemetry by 70 to 90 percent, losslessly, in microseconds, running on the edge gateway where the data is collected before it goes out. If you are sending a fraction of the bytes, you can spend that however you want: longer battery life, more frequent readings, or more channels off the same device, with no change to the sensor and no new silicon. The article treats battery-versus-frequency as a fixed wall everyone designs around. With Neurpac in the path it is not fixed.

The article’s own examples, read through Neurpac:

More data per device just moves the bottleneck to the link, which is good for us.

Endress+Hauser told the conference the biggest near-term shift is APL connectivity. A device that used to report one value over the forty-year-old electrical standard can now send flow, temperature, viscosity, density, and concentration at the same time. More devices going wireless, plus more data coming off each one, both push more bytes across links that are short on bandwidth or battery. So the market is producing more of exactly the kind of structured machine data Neurpac handles best, and putting it on the constrained links where saving bytes is worth the most. The bigger the segment gets, the more there is for us to compact.

The “cloud hesitancy” section is really about data sovereignty, which plays to us.

IoT Analytics reports that resistance to the cloud, driven by security and data-sovereignty concerns and strongest in Europe, is reshaping how these systems get deployed. Vendors are building cellular gateways that skip the customer’s IT entirely (Schaeffler uses SIM-based gateways for this) and private-routing setups that drop raw data straight into the customer’s own servers (Status Pro’s MQTT split). The article’s read is that cloud hesitancy is not a hard blocker but a cost-and-complexity tax, because it forces fragmented, custom data pathways instead of clean centralized ones.

This is where Neurpac fits. It runs as a transparent compaction tunnel on standard edge gateways, including Cradlepoint, which Ericsson has already tested and validated through SD-WAN. It compacts and encrypts in the same pass, without the attacker penetration problem that made the rest of the industry stop compressing encrypted traffic years ago. For a customer who will not put raw operational data on a public cloud, that makes the private-routing setup they are already building cheaper to run, since there are fewer bytes going over expensive private or cellular links, and harder to attack, since the data is encrypted on the wire and the codebook is meaningless without our software. One vendor in the piece calms nervous customers by telling them the data is only temperature and vibration readings. With Neurpac in the path they can say it is encrypted and a tenth the size too.

We are already shipping into the same kind of deployment.

We are not guessing about whether this data profile works. Atombeam signed a reseller deal with Trilliant to put Neurpac in smart grid that covers something like 40 million devices, and on piloted oil and gas monitoring over satellites at 76 percent compaction. Smart grid and remote SCADA are basically the same as industrial maintenance telemetry: big fleets of devices, small structured packets, links that are metered or scarce, and a customer who pays for the connectivity. So the market the article describes sits right next to one we already sell into.

Where PCM comes back in.

The article is also blunt that the real barrier to maintenance AI is not the models, it is data quality and trust. Hexagon is deliberately keeping its AI away from proprietary customer data until it has hallucination risk under control, on the logic that a confident wrong recommendation in a plant is worse than no recommendation at all. That is more or less why PCM exists. we believe between the knowledge angle and the trust angle, the article is describing demand for both products: PCM for the knowledge and trust layer, Neurpac for the data movement and sovereignty layer underneath. I am not going to turn that into a packaged-product pitch right now. I just think it is worth noticing that an outside survey of a trillion-dollar problem managed to describe both of the things we make.

The proper hedges.

A few caveats. We do not have a named maintenance customer out of this. It is a market read backed by real deployments next door, not a deal. Neurpac also does not solve the knowledge problem the article opens with. We believe it solves the plumbing that makes the knowledge systems cheaper to run and easier to keep sovereign, which matters, but it is a different thing, and I do not want us blurring the two with investors. The article itself calls cloud hesitancy a cost-and-complexity tax rather than a hard blocker, so our sovereignty pitch is “we make your setup cheaper and safer,” not “you cannot deploy without us.” And the battery benefit happens at the gateway, where the telemetry is gathered and sent out, not down at the sensor itself. That still covers most of these deployments, but we want to be precise on that.

I am sending this around because an independent analyst looked hard at a trillion-dollar problem and kept naming the same constraints we built Neurpac to deal with, whether that is battery life against sampling rate, more data coming off each device, or data that customers will not let leave the building. The wireless monitoring piece of this is already past a billion dollars and growing about 30 percent a year. Our plan is to be selling into it.

Charles

*Neurpac’s power efficiency projections are based on prototype testing and theoretical modeling. Actual results may vary significantly from these estimates.

**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.

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 the Offering Circular and Related Risks.

Today’s the day. At 9am PT / 12pm ET, the Atombeam team is going live.

This is your chance to meet:

• Charles Yeomans, CEO and Chairman
• Cody Martinson, President of the Neurpac Division
• Alexandria Tucker, AI Scientist for PCM
• Nadine Alam, Chief of Staff
• Julien Dersy, Chief Product Officer
• Pam Dodge, VP of Marketing

We’ll have a team-focused discussion about Atombeam’s technology, growth strategy, product development pipeline, and the opportunity at the heart of this funding round.

Join us at the link below and get your questions ready!

[Register Here]

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 the Offering Circular and Related Risks.

If you haven’t registered yet, now’s the time. Tomorrow’s webinar is your chance to hear directly from the leaders behind Atombeam.

CEO and Chairman Charles Yeoman will lead a discussion between key members of the team who deal in every aspect of the business, from product development to operations to marketing.

You can get direct insight from them about how things are going and where we’re heading next, and ask any questions you might have. With this many minds in one place, someone will have the answer.

Register today and join us tomorrow, Monday June 29, at 9am PT / 12pm ET!

[Register Here]

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 the Offering Circular and Related Risks.