Mr. Mehdi Azodi reports

BELGRAVIA HARTFORD REPORTS GRAVITIO.AI TECHNICAL AND USER GROWTH AS RECORDED AI-AGENT PREDICTIONS SURPASS 10,000

Belgravia Hartford Capital Inc. has provided a further technical and growth update on Gravitio.ai, the company's wholly owned artificial-intelligence-powered prediction intelligence platform.

Since the company's June 19, 2026, technical update, Gravitio has continued to advance the infrastructure supporting its AI-agent prediction system, with recent development focused on the GRV-2.1 inference architecture, prediction market intelligence, historical market data, automated outcome verification and feedback from verified prediction performance.

Platform growth and engagement

Since the company's June 19, 2026, technical update, Gravitio has grown from more than 4,000 registered users to 5,500 registered users. Based on internal platform activity records, 3,960 users have recorded active platform usage, representing approximately 72 per cent of the registered user base.

Platform activity has also expanded materially. Recorded AI-agent predictions have increased from 5,874 reported on June 19 to 10,409, an increase of approximately 77 per cent, while cloned AI agents have grown from more than 8,000 to 9,077. These metrics provide additional evidence of platform usage and the continued expansion of Gravitio's underlying prediction performance data layer.

Importantly, this growth has been achieved without a formal marketing campaign. User acquisition to date has been driven primarily by early platform discovery, direct on-boarding, referrals, app store availability and initial market exposure. With the platform's technical foundation continuing to mature, the company is preparing to begin a more structured marketing and user acquisition program intended to expand awareness of Gravitio and support the next stage of platform growth.

GRV-2.1 AI inference architecture

GRV-2.1 advances Gravitio's centralized AI inference architecture by separating data acquisition, AI reasoning, prediction storage and outcome verification into distinct components.

Market, news, sports and prediction market services collect and normalize information before relevant structured data are supplied to the AI reasoning layer. This architecture is designed to improve consistency, data control, caching, API rate management, fault handling and scalability across Gravitio's expanding network of AI agents.

The system generates structured prediction outputs that may include prediction direction, confidence level, time horizon and supporting reasoning. Gravitio's master clone architecture is also designed to reduce unnecessary duplicate AI workloads by allowing qualifying cloned agents to reference centralized system-level intelligence while maintaining corresponding performance histories.

In business terms, this modular architecture is designed to reduce duplicate AI processing and unnecessary external data requests, improve operating efficiency, and support a larger user and agent base without requiring a proportional increase in AI-compute workloads.

Prediction market and multisource intelligence

Gravitio has expanded its prediction market infrastructure through integration with Polymarket market data, including available probability, pricing, spread and order book information.

These signals can be evaluated alongside other data sources, including market prices, technical indicators, sentiment, news and historical information. Gravitio's crypto-focused agents similarly incorporate broader market context, including price and volume data, bitcoin dominance, technical indicators, historical market behaviour, sentiment measures, and structured news intelligence.

The objective is to provide the AI inference layer with multiple independent sources of information rather than generating predictions based on a single market indicator or data source.

Automated verification and performance feedback

Building on the proprietary prediction performance data layer described in the company's June technical update, Gravitio records qualifying AI-agent predictions and subsequently compares eligible predictions against observed outcomes.

Verified results can be incorporated into agent performance records, including recent prediction performance, category- or symbol-level results, and directional tendencies.

Historical performance information can then be returned to the AI inference process as contextual data for subsequent predictions, creating a continuing technical cycle: data collection to AI inference to structured prediction to outcome verification to performance measurement to feedback.

This architecture is intended to support Gravitio's development as an adaptive prediction intelligence system in which verified outcomes and accumulated historical information can contribute to future AI reasoning.

Management commentary

Bobby Moshar, chief executive officer of Gravitio, commented: "Since our June update, our priority has been to strengthen the intelligence infrastructure behind Gravitio while preparing the platform for greater scale. Recorded AI-agent predictions have now increased to 10,409 while cloned agents have reached 9,077. GRV-2.1 separates data collection, AI reasoning, prediction storage and outcome verification so that each component can be measured and improved independently.

"At the same time, reaching 5,500 registered users, with 3,960 users recording active platform usage, before launching a formal marketing campaign provides Gravitio with an important early user and engagement base as we enter the next stage of growth. The combination of broader data intelligence, automated prediction verification, historical performance feedback and a scalable AI-agent architecture is creating the technical foundation we believe is required before accelerating user acquisition.

"Our objective remains to build Gravitio as an adaptive prediction intelligence platform where AI agents, market data, human prediction activity and verified real-world outcomes continuously contribute to an expanding intelligence and performance layer."

Mehdi Azodi, chief executive officer of Belgravia, commented: "Our long-term objective has always been to build an operating company with the potential to be spun out and distributed to Belgravia shareholders as a dividend, subject to applicable corporate, securities, tax and regulatory requirements. The current microcap market is highly selective and increasingly rewards projects that demonstrate quality, technical maturity and execution. We believe our continued efforts to advance Gravitio are aligned with creating long-term value for Belgravia shareholders."

No decision has been made to pursue a spinout or shareholder distribution. Any such transaction would be subject to, among other matters, approval by the company's board of directors, applicable corporate and securities laws, tax considerations, exchange requirements, and other customary conditions.

The company will continue developing Gravitio's supported data sources, AI-agent capabilities, prediction categories and performance measurement infrastructure while advancing planned marketing, user acquisition and commercialization initiatives.

About Belgravia Hartford Capital Inc.

Belgravia is an investment issuer focused on technology, finance, artificial intelligence, digital assets and related investment opportunities. Listed for trading on the Canadian Securities Exchange and OTCQB, the company is focused on the technology and finance sectors of the bitcoin ecosystem. The company's 2018 investment policy specifies cryptocurrencies, artificial intelligence, media and digital streaming opportunities.

Belgravia invests in a portfolio of private and public companies located in jurisdictions governed by the rule of law. Belgravia and its investments are considered very high-risk holdings, and may expose shareholders to significant volatility and losses.

We seek Safe Harbor.