Albert Dadon’s AEREDIUM secured a patent for a multi-model AI system to audit digital asset and stablecoin reserves continuously in real time.
Key Takeaways
AEREDIUM CEO Albert Dadon won U.S. Patent 12,694,407 B1 on July 29 for a 4-model AI audit system.The broad USPTO grant blocks rivals from using off-the-shelf AI to audit 100% of stablecoin reserves.The company is preparing global patents and engaging U.S. regulators following its 7th patent grant.According to a media statement, AEREDIUM’s patented system replaces periodic reports with continuous monitoring using four independent AI models. Under the architecture, no single model can determine an outcome independently. A verdict requires consensus from at least three of the four models, subject to the compliance engine’s override power.
The system issues status levels ranging from PASS to HALT. When a HALT status is issued, the system automatically instructs connected smart contracts to reject pending token transactions, preventing an impaired reserve from continuing to trade. Audit findings are recorded to a primary blockchain within minutes and cross-anchored to a second independent blockchain, creating an immutable public ledger.
“Trust in this industry has always been retrospective — you found out whether the reserves were there after it mattered,” Dadon said. “This patent makes trust a live property of the asset itself. The verification travels with the token.”
“Machine-learning components in the abstract belong to everyone, but applied to this use case — the autonomous auditing of digital asset reserves — they are protected,” Dadon said. “A competing issuer or auditing firm deploying these models against stablecoin reserves is operating inside the granted claims, whatever shelf the components came from.”
He added that AEREDIUM is actively preparing international patent filings.
Addressing potential operational hurdles such as latency from custodial bank APIs, Dadon explained that the architecture is designed to accommodate slower off-chain data feeds. The cycle utilizes the freshest available data from each source, treating stale or diverging feeds as risk signals rather than system failures.
Dadon also clarified that the AI ensemble is not intended to replace legally mandated accounting audits or diminish issuer accountability.
“Legal responsibility for a stablecoin’s reserves sits with the issuer, and whatever audit a regulator mandates, the issuer must run,” Dadon told Bitcoin.com News, describing continuous verification and traditional audits as complementary processes.
However, Dadon indicated that regulatory expectations may shift as real-time verification tools become established, noting that AEREDIUM is currently engaging with U.S. regulatory bodies.
“We expect regulators to move from accepting continuous verification to demanding it,” Dadon said. “In the United States, reserve requirements for payment stablecoins are set in statute, and continuous verification against a statutory standard is precisely what this architecture was built for.”
Hero/Feature image credit: SBS Hebrew/sbs.com
