Algorand Shows Up in an Air Force Patent Filing: What the Research Actually Proves

What the Patent Filing Actually Says

The application is titled "Systems and methods for blockchain-based secure multiparty computation," publication number US20250363486A1, filed May 24, 2024 and published November 27, 2025. Google Patents lists the assignees as Florida International University (FIU) and the United States Department of the Air Force, with six named inventors: Oscar G. Bautista Chia, Richard Hernandez Garcia, Kemal Akkaya, Selcuk Uluagac, Soamar Homsi, and Mohammad Hossein Manshaei. The filing includes the standard government-rights clause noting federal funding was involved in the invention.

One correction worth making up front: several outlets covering this story this week described the patent as having been "officially granted" in June. The primary record on Google Patents lists the application's legal status as pending, not granted, as of this writing. That's not a small distinction. A pending application means the US Patent and Trademark Office has published the filing for examination; it has not yet issued as an enforceable patent. Readers should treat "granted" claims about this specific filing as unverified until a grant number shows up in the primary record, and treat what follows as a description of a patent application, not an issued patent.

The Actual Research: Secure Multi-Party Computation Over Algorand

Secure multi-party computation, SMPC, lets multiple parties jointly compute a function over their private inputs without revealing those inputs to each other, useful for anything from joint statistical analysis to coordinated decision-making where the underlying data has to stay compartmentalized. The problem SMPC has always had is communication overhead: parties need a reliable, verifiable way to exchange intermediate values across multiple rounds of computation, and that channel has to resist tampering and forged messages, especially in a zero-trust environment where you can't assume every participant's network path is safe.

The FIU-led research, sponsored by the Air Force Research Laboratory, tested whether an Algorand blockchain could serve as that communication layer. The design attaches SMPC messages to the note field of Algorand transactions, then lets Algorand's gossip protocol handle distribution, broadcasting those transactions to every node in the network rather than requiring each computation party to maintain a direct peer-to-peer connection with every other party. Transaction signatures do double duty as message authentication, letting a recipient verify a message came from the participant it claims to be from and wasn't altered in transit. Because the ledger retains a record of every message, the design also has a built-in recovery path: a party that drops out mid-computation can, in principle, reconstruct earlier rounds from chain history instead of forcing the whole computation to restart.

Why a Private Network, Not the Public Chain

The test deployment ran on a private Algorand network in the cloud, not Algorand's public mainnet, with up to seven computation nodes and four relay nodes in the tested configurations. The researchers were explicit about why: a private network gave them stability, easier customization, open-source deployability, faster transaction propagation, and immediate access to transactions as they arrived at each node, without needing to acquire ALGO to pay public network fees.

That detail matters more than it might look like at first glance. A private network reuses Algorand's open-source software and architecture, but it generates none of the fee activity, token demand, or observable on-chain usage that a public mainnet application would. The research supports a technical case for Algorand's codebase, specifically its gossip protocol and transaction-signing model, being suitable for a specialized communications problem. It is a materially weaker claim than an economic case for ALGO itself, and the paper doesn't attempt to make the stronger one. It's also worth noting the researchers didn't present Algorand as irreplaceable: the paper acknowledges that another blockchain with comparable broadcast speed, customizability, and transaction accessibility could, in principle, serve the same role. Algorand met the requirements. It wasn't framed as the only chain that could have.

What the Performance Numbers Show

The researchers benchmarked their Algorand-based broadcast channel against conventional TCP peer-to-peer communication, using total computation completion time as the primary metric. Reported results showed average delay improvements of 24% at a 4-kilobyte note size and 32% at a 16-kilobyte note size, with the advantage growing as matrix sizes in the underlying computation got larger. That's a meaningful signal for a specific, communication-heavy workload: coordinating message exchange through blockchain transactions and a gossip protocol outperformed pairwise TCP connections in the tested range.

It's also a bounded result. The tests covered a cloud environment with a small, fixed set of nodes. Nothing in the published account demonstrates how the approach holds up at significantly larger scale, under different workload types, or inside the kind of operational security constraints an actual defense deployment would carry. A benchmark showing lower delay under controlled conditions is evidence the architecture works as designed. It isn't evidence of production readiness.

What This Doesn't Prove, and a Second Claim Worth Separating Out

Coverage of this story has also connected it to a separate claim involving Vertex Corporation, reportedly building an Algorand-based marketplace for aircraft parts producible through 3D printing, and said to have received roughly $2 million in Air Force contracts for a "secure digital fabric." No contract document or primary source URL has surfaced alongside that claim, and it should not be treated as the same project as the FIU patent filing. They may both be real signals of institutional interest in Algorand's architecture. They are not one story, and merging them overstates what either individually supports.

Zooming out, the honest summary is: an Air Force Research Laboratory sponsored project, run through FIU, found a private Algorand configuration useful enough for a secure multi-party computation experiment to file a patent application and publish performance results. That's a legitimate technical reference for Algorand's underlying software, the kind of validation that shows up in academic and defense research contexts independent of price action or public adoption metrics. What it doesn't show is a live military system, a public-chain deployment, ALGO procurement, or a completed patent grant. Those are different categories of evidence, and this filing only supports the first one.