Three Provisional Patents Filed

Bimini Technologies has filed three provisional patent applications with the United States Patent and Trademark Office (USPTO), covering the core technology that powers its market intelligence platform. The filings were submitted on May 22 and May 23, 2026.

The three filings protect distinct but interconnected systems: how data enters the platform and becomes intelligence (Intelligence Refinery), how anomalies are detected within it (Ghost Pattern Detection), and how the output reaches the user in a compliance-grade format (Chain of Logic). Each system is independently novel. Together, they protect an architecture where no single component can be removed without breaking the pipeline.

---

Patent 1 — Detection

Ghost Pattern Detection

System and Method for Detecting Pre-Disclosure Behavioral Anomalies in Market Microstructure Across Correlated Instruments Using Reconstructed Market State Data

This patent covers Bimini's methodology for detecting statistically significant behavioral anomalies in market microstructure — patterns in order flow, volume, and price behavior that have historically preceded major market events. The system reconstructs raw exchange data into a proprietary multi-dimensional analytical format, applies statistical anomaly detection to identify behavioral signatures consistent with informed positioning, and validates signals across structurally independent instruments using Bayesian evidence accumulation.

The filing also covers Zero-Detection Day analysis — the detection of the absence of expected anomalies as an independent predictive signal. Under normal conditions, detections occur daily across monitored securities. A day with no detections is statistically anomalous and has historically preceded high-uncertainty events within 24 to 72 hours.

The detection methodology is grounded in approximately two decades of peer-reviewed research from Georgia Tech's Scheller College of Business, published in the Review of Financial Studies and the Journal of Financial Economics, which demonstrated that pre-disclosure behavioral anomalies in market microstructure are real, detectable, and carry measurable significance. The convergence between this research and Bimini's independent live-market findings is documented in the Bimini white paper. Four worked examples of the detection in action are presented in Four Signals Before the News Broke.

---

Patent 2 — Synthesis

Intelligence Refinery

System and Method for Transforming Commodity Public Data into Proprietary Market Intelligence Through Multi-Domain Synthesis, Institutional Response Model Matching, and Accumulation into a Compounding Longitudinal Dataset

This patent covers the system that ingests publicly available data from eight heterogeneous domains — government economic releases, Federal Reserve communications, market microstructure, regulatory filings, corporate earnings, academic research, open source intelligence, and reported & non-reported news — and transforms each data point through a proprietary analytical framework into intelligence that does not exist in any public data source.

The filing covers three novel elements. First, multi-domain synthesis: normalizing data from structurally different sources into a common analytical schema and correlating across domains to produce forward-looking market assessments classified as Tailwind, Headwind, or Crosswind. Second, institutional response model matching: comparing current data configurations against published models of how institutions such as the Federal Reserve have historically responded to similar conditions, enabling anticipation of policy actions before announcement. Third, compounding longitudinal dataset architecture: the Market Substrate, where each new assessment is evaluated against every prior assessment and where analytical value scales with the interactions between entries rather than the entry count.

The architectural shift this represents — from systems of record to systems of intelligence — is explored in From Data Terminal to Intelligence Layer.

---

Patent 3 — Delivery

Chain of Logic

System and Method for Generating Tiered, Compliance-Grade Intelligence Reports with Linked Reasoning Chains, Evidence-Based Admission Rules, and Historical Base Rate Classification

This patent covers the system that automatically generates structured reasoning trails accompanying every market intelligence assessment. The Chain of Logic is the compliance-grade document that Registered Investment Advisors (RIAs) need to meet their fiduciary obligations under Section 206 of the Investment Advisers Act of 1940.

The filing covers tiered rendering architecture — the same underlying reasoning chain rendered differently based on regulatory context. Display-tier output shows full directional language on screen with no export capability. Export-tier output produces compliance-grade documentation with classification framing, historical base rates, sample-size qualifications, mandatory failure-case disclosure, source citations, and non-removable legal disclaimers. The filing also covers the admission rule engine that gates which detected signals may enter a reasoning chain, and a removal test architecture that verifies each step in the chain is independently necessary for the final assessment.

---

Why three filings

The three patents cover the full intelligence pipeline. The Intelligence Refinery determines how data enters the system and becomes intelligence. Ghost Pattern Detection determines how anomalies are identified within it. Chain of Logic determines how the output reaches the user in a compliance-grade format.

Each system is independently novel. Together, they protect an architecture where no single component can be removed without breaking the pipeline. A competitor who builds one cannot replicate what the three produce together.

---

Patent pending

Each application establishes “patent pending” status and a priority date with the USPTO — a flag planted that this architecture was invented here, by this team.

All three systems use only publicly available data as input. The patents protect the methods of transformation, detection, and structured reasoning — not the underlying data sources. Bimini does not access, solicit, or process material non-public information (MNPI).

---