HSBC Appointed as Tokenization Agent for Hong Kong Project Ensemble Pilot Program

HashKey collaborates with Bosera to launch a tokenized currency market ETF, supporting Hong Kong's "Project Ensemble" initiative.

Hong Kong Securities and Futures Commission: A financial institution participating in the Ensemble project sandbox has completed its pioneering simulation test.

HKMA Project Ensemble Partners with Brazil and Thailand for Tokenization

HKMA Project Ensemble partners with Brazil, Thailand for tokenization

Huaxia Fund (Hong Kong) announced its participation in the Hong Kong Monetary Authority's Ensemble project sandbox for fixed income and investment fund use cases.

The Hong Kong Monetary Authority's Ensemble project sandbox first phase has a total of 25 institutions participating, including HSBC, Bank of China Hong Kong, Hang Seng, and Standard Chartered.

Hong Kong Launches CBDC Sandbox for Project Ensemble to Explore Tokenization in Banking

Hong Kong's Project Ensemble Advances in Asset Tokenization Testing

Ant Group and Ant International join the Hong Kong Monetary Authority's Ensemble project sandbox.

HSBC: Has completed three concept validation use cases in the Ensemble sandbox and become one of the first financial institutions to explore use cases.

Bank of China Hong Kong: One of the first banks to participate in the Hong Kong Monetary Authority's Ensemble Project Sandbox

Hong Kong Monetary Authority Reports Progress on Ensemble Project with Ant Chain's Assistance

HashKey Group will experiment with tokenized financial solutions in the Hong Kong Monetary Authority's Ensemble project sandbox.

Hong Kong's HKMA Advances Asset Tokenization with Project Ensemble Sandbox

The Hong Kong Monetary Authority announced the official launch of the Ensemble project sandbox.

Hong Kong Monetary Authority Launches Ensemble Project Sandbox to Promote Tokenized Asset Applications

Hong Kong Monetary Authority: Ensemble Sandbox's four tokenized asset use cases include fixed income and investment funds, liquidity management, green and sustainable finance, etc.

Hong Kong Monetary Authority Initiates Ensemble Project Sandbox for Asset Tokenization

Hong Kong Monetary Authority Initiates Ensemble Project Sandbox for Tokenization

Hong Kong Monetary Authority Launches Ensemble Project Sandbox to Promote Tokenization Applications

Hong Kong's Project Ensemble Advances in Asset Tokenization Testing

HKMA establishes ENSEMBLE project working group to support tokenization market development

HKMA establishes ENSEMBLE project working group to support tokenization market development

Hong Kong Monetary Authority establishes ENSEMBLE task force to support tokenized market development

Ant Financial joins HKMA’s ENSEMBLE project to explore tokenized assets

Hong Kong Launches Project Ensemble to Support Tokenization with WCBDC

Hong Kong's Central Bank Launches Project Ensemble for Tokenization Market Support

Ant Group's First International Sustainability Report: Increasing Investment in Crypto Technology Applications, Blockchain Platforms Handle Over One-Third of Global Transactions

HSBC intensifies its blockchain initiatives! Focus areas: digital asset custody, HSBC Orion, and tokenization Orion has issued over $1.5 billion in bonds across HK HSBC appointed tokenization agent for China AMC in HK’s Project Ensemble #HSBC #Blockchain #Tokenization #CryptoNews #ProjectEnsemble https://t.co/ZJDLpxhynt

Wu Shuo learned from the Hong Kong Economic Journal that HSBC's Chief Digital, Data and Innovation Officer, Zhu Kuang Lee, stated that HSBC is focusing on three major areas of blockchain technology: digital asset custody, the private blockchain platform HSBC Orion, and tokenization. In the past year and a half, the HSBC Orion platform has completed more than 10 bond issuances in Hong Kong and other regions, with a total size of over $1.5 billion. In addition, HSBC was recently appointed by China Asset Management as the tokenization agent for Hong Kong's Project Ensemble pilot program, assisting in the minting and burning of money market fund shares. https://t.co/kxsB4Z3sKb

Planet Midday News: GMGN X Account Ban Unclear, Whales Increase Holdings in ETH and HYPE, Bitcoin ETFs Continue Net Inflows

Here is the translated content in English: Announcing the next class of Union Fellows: @esprezzoapp @EnsembleCodes @conduct_xyz @nativerse_xyz @themafinancial @Hyperweb_ @SphinxProtocol https://t.co/tgj2ZGftoK

Forecasting real estate prices in volatile or underserved regions is a high-variance problem. Sparse data, shifting local conditions, and nonlinear shocks cause traditional models to misfire or overfit. Static models—even ensembles—struggle in low-liquidity environments because they lack adaptivity. When ground truth distributions shift, yesterday’s model performance becomes a poor signal of tomorrow’s. Allora addresses this using a decentralized forecasting protocol. Within this structure, each worker can: • Submit an inference for a target variable (e.g., future price) • Forecast the loss (expected error) of other workers’ inferences under current conditions These second-order predictions—loss forecasts—are essential. They allow Allora to generate forecast-implied inferences: • Weighted combinations of primary inferences • Informed by context-aware expectations of accuracy from peer forecasters This leads to a network-level mechanism known as Inference Synthesis. Unlike static ensemble techniques, Allora reweights each contribution based on: • Historical performance • Real-time, contextual loss forecasts • Predicted regret: the expected shortfall of a worker’s inference versus the network’s past performance The weights are determined using a smooth potential function applied to normalized regrets. This yields dynamic inference weights that favor models likely to perform well under current conditions. The system doesn’t assume stationarity. It learns where predictive accuracy is likely to concentrate—using real-time assessments from competing workers—and shifts aggregation toward those workers in a given context. This adaptivity is particularly powerful in fragmented real estate markets, such as: • Rural and peri-urban areas • Zones undergoing regulatory or infrastructural transformation • Geographies where transaction data is delayed or unreliable Crucially, Allora’s structure allows contributors with niche models or hyperlocal insights to participate meaningfully. Its reward structure ties compensation to: • Unique contribution to network accuracy • Not stake, reputation, or frequency of participation This means the system doesn't reward volume, but marginal utility. No single model dominates in Allora. Instead: • The network synthesizes a context-aware intelligence layer • It reflects the most effective predictive elements at any moment in time • It improves continually as new patterns, contributors, and conditions arise Historically, real estate markets have suffered from information asymmetries. Access to high-quality forecasts has been gated by data centralization, proprietary tooling, or geographic privilege. Allora changes that dynamic: • Intelligence is decentralized • Performance governs visibility • Adaptivity is built-in at the protocol level By design, Allora enables actionable forecasting in places that static models cannot serve—urban fringe zones, remote townships, or rapidly transitioning micro-markets. In these environments, volatility is not just noise—it is often the signal. The future of real estate inference is not about discovering a single “perfect” model. It’s about designing systems that: • Continuously learn which models to trust, when • Adapt structurally to dynamic data environments • Incentivize useful contributions from anyone, anywhere Allora does exactly this—turning volatility from a liability into a structural advantage.

LLM Use Case – Development Status 🧠 NuNet’s Large Language Model (LLM) initiative continues to progress steadily, with current efforts focused on exploring fine-tuning techniques using decentralized infrastructure. ⚙️ Key Milestones: Tested Models: 📌 DistilGPT2 (lightweight transformer, optimized for speed and efficiency) 📌 Mistral 7B (modern, high-performance architecture – testing initiated but not fully completed due to infrastructure transitions) Deployment Architecture: 🧪 Demonstrated the feasibility of fine-tuning via ensemble deployment 🧩 Testing is in its initial phase on a single-node setup, with multi-node support in the pipeline Infrastructure & Progress: ⏳ Temporary delays were encountered due to the decommissioning of the Fluxus machines 🔧 Progress resumed with a newly configured environment and successful deployment on a new node 🧪 Active testing is underway using the Bitext Travel QA dataset [View dataset] 📂 Additional datasets are being considered for future experiments What’s Next: The team remains in an experimental and exploratory phase, currently focused on: 🔍 Evaluating fine-tuning performance within decentralized environments 🛠️ Laying the groundwork for multi-node orchestration 📊 Expanding dataset diversity and model testing as infrastructure matures This use case not only advances our technical roadmap but also plays a vital role in validating NuNet’s broader vision of scalable, decentralized AI—one node at a time.

🌐 As tokenisation transforms global finance, investor and market education remains critical. At a recent seminar hosted by @HKUST Crypto-Fintech Lab, Harmon Venture, and GBA RWA Incubator, our HashKey Tokenisation team, led by Andy P. Lee and Lianxuan Li, shared how tokenisation is unlocking new opportunities in real-world assets (RWA) and redefining liquidity. ✨ Key Highlights: ✅ Liquidity Unleashed Blockchain is unbundling liquidity, much like fintech disrupted traditional banking. With stablecoin regulations emerging, tokenised products are poised to deliver global investor value. ✅ Regulatory Evolution Hong Kong's Project Ensemble is setting the framework for tokenisation’s future, balancing compliance with innovation. ✅ Real-World Impact Our collaboration with Bosera Asset Management (International) Co. led to the creation of the world’s first Tokenised Money Market ETFs, featuring: 🔹 Low Subscription Fees → Low-Cost Access 🔹 T+0 Settlement → Instant Finality 🔹 Full Compliance → Regulatory Approval and Security 🚀 The discussions were truly inspiring, and we’re excited to continue partnering with industry leaders to drive this transformation forward.

elizaOS Builder Demos: EnsembleCodes and HashgraphOnline https://t.co/1lrT85FZwg

🚧 What We’re Building This Week at NuNet 🔹 Current Priorities 🔧 Job orchestration design using Contract Database 📄 Finish test matrix and feature files for the functional tests 🧪 Dynamic ensemble: adding/removing nodes 🔹 Under Review 🧩 Define a Test Policy For more updates 👉 https://t.co/5Q2Bl90zNj

✨ NuNet Appliance v1.4 Now Available! We officially released Appliance 1.4 last week! This release wraps up everything from our April milestones, delivering all planned features and several important bug fixes. Here’s a quick recap of what’s inside: 📦 April Milestones: 🔐 Local Web App & SSH Access: Manage your appliance through a streamlined web interface or via SSH for direct control. 🔌 Plugin Integrations: Extend appliance functionality with plugins like Victoria Metrics for telemetry, and enhanced monitoring via Node Exporter and cAdvisor. 🔁 Simplified Update System: Easily upgrade your DMS, operating system, and menu through a seamless update process. 👥 Organization Roles: Select your role within NuNet Organizations — as a Compute Provider or an Orchestrator. 🛡️ Boot & Security Enhancements: Improved network booting with robust image verification through hash checks for secure deployments. ✅ Key improvements to appliance stability and performance 🐞 Major bug fixes based on community feedback 🧠 Feature refinements to enhance the overall user experience — for example, advanced ensemble manager 📥 We invite you to download the latest release, test it, and share your feedback. Every insight helps us refine the network for everyone. 🔗 Link to download the latest appliance: https://t.co/SQSYds9ocX ⚠️ Note: Do not click on any links that are not from our official handle. You may fall prey to scams and security compromises. 💬 Run into issues or have suggestions? Drop them in the 💬│testers-chat channel on discord https://t.co/dxuR3QJe6v

📖 Chart Analysis and Financial Models Using CNN (Professor Jang Hee-soo, Department of Finance, Soongsil University) Powered by @BanklessKorea CNN-based price prediction models demonstrate high accuracy and Sharpe ratios by learning chart images, and are superior in interpretability and flexibility compared to traditional analysis. They are particularly effective in short-term predictions, and performance is further improved when applied to various markets in an ensemble manner.👉 Detailed article: https://t.co/eCZZnwpZYB#CDFI #월스트리트디파이 #DeFi

PIP LabsがStory Protocolのために$80MのシリーズB資金調達を行い、評価額は$2.25 Billionに達しました。