This year, the amount of data in the world is expected to reach ninety zettabytes. For reference, one zettabyte is equal to a trillion gigabytes.
But right now, the storage capacity of all the data centers globally is about two zettabytes. This overwhelming shortage is placing mounting pressure on networks, slowing and overloading them. So, what’s the answer?
AtomBeam has developed a potential solution. It’s a data-compaction software company that shrinks, secures, and speeds data transmission.
Using machine learning and Artificial Intelligence (AI), it reduces the size of data files by seventy-five percent, and adds a much-needed layer of security. This technology has the potential to free up space for all of this data and enable it to be shared quickly and securely.
AtomBeam is led by an experienced team. Its CEO has twenty-five years of financial experience, and led two prior companies. Its Chief Technology Officer managed software teams at HP, Siemens, and Red Hat. Its Chief Scientist is a mathematics professor at the University of South Carolina, and holds degrees from MIT and the University of California, San Diego. And its Chief Revenue Officer managed sales teams at SAP and Oracle.
The company’s software, called Neurpac, merges data reduction, security, and search ability into a single algorithm. With an average of seventy-five percent reduction in data size, Neurpac maintains data integrity by encoding standard data into much smaller codewords. This can make networks up to four times more efficient and secure, paving the way for technologies like AI.
To a computer, a tiny codeword has all the same information as the much larger pattern it represents, generally allowing the computer to operate far more efficiently by processing far less data. Notably, a computer doesn’t need all of the human-driven context and repetition that bulks up data, but instead can work on the core information without wasting its resources.
AtomBeam’s second product, Neurcom, will address data types that fall under “lossy” categories: images, video, and audio. This will be a complementary product to Neurpac, which is suited for data like telemetry and tracking messages. Neurcom is under development as part of a $1.2 million contract with the U.S. Air Force and is expected to be available for commercial use in 2026.
In the meantime, AtomBeam has achieved significant progress. It generated $640,000 in revenue from the U.S. Air Force and U.S. Space Force in 2023. And in mid-2023, the company was awarded a Phase II contract for $1.2 million by the Air Force Research Laboratory to develop Neurcom for use in satellites and aircraft. AtomBeam holds 121 patents.