Uniswap Labs Partners with Fireblocks for Institutional DeFi Access, U.S. Regulated Settlement Network, Anchorage Digital Offers Liquid ETH Staking...
Hi Everyone,
Welcome to QX Snapshots - a weekly recap of the key news on emerging technologies. In this newsletter, you will get a "digest" of the latest info on Enterprise Blockchain, AI, Quantum Technology, and more.
Hope it brings you value :)
[Blockchain] Uniswap Labs Partners with Fireblocks for Institutional DeFi Access, U.S. Regulated Settlement Network, Anchorage Digital Offers Liquid ETH Staking, and Fortlake Tokenizes Sigma Fund with Chainlink and Sonic. Uniswap Labs and Fireblocks have announced a strategic collaboration that enables Fireblocks' 2,000 institutional clients to access the Uniswap Protocol directly through their platform. This integration allows financial institutions like asset managers and hedge funds. The Uniswap Trading API, developed by the team behind the protocol that has processed $2.5T in volume, simplifies DeFi trading by handling technical complexities like gas estimates and routing options, while offering improved pricing compared to leading aggregators. Meanwhile, major U.S. financial institutions have completed a proof-of-concept study exploring the Regulated Settlement Network (RSN), a shared ledger technology platform for settling multi-asset and cross-network transactions. The project, managed by SIFMA, demonstrated the feasibility of simultaneous 24/7 settlement of tokenized assets including bank deposits, U.S. Treasuries, and other regulated assets. Participants included Citi, J.P. Morgan, Mastercard, and others, with technology provided by Digital Asset and Swift. The study found no legal barriers to implementation and showed potential for reducing inefficiencies in current siloed settlement systems. Also, Anchorage Digital has become the first U.S. federally chartered bank to support liquid Ether staking through Liquid Collective's LsETH token. The bank aims to serve institutional clients, including VC firms and wealth managers, amid growing demand for institutional staking solutions. This development comes as ETF sponsors anticipate potential approval for staking in U.S. ETH ETFs. Liquid Collective, focusing on regulatory compliance and security, has $430 million in total value locked. And, Backed, Sonic (formerly Fantom), and Chainlink are partnering with Fortlake Asset Management to tokenize their Sigma Opportunities Fund. The fund will be tokenized on Sonic's blockchain using Backed's Tokenization-as-a-Service platform, with Chainlink providing critical infrastructure for NAV data, collateralization verification, and cross-chain operations. The Sigma Fund, managed by Fortlake, targets 7-10% returns above the fed funds rate.
[AI] AWS Unveils Trainium-Powered Ultracluster for Anthropic, Google DeepMind Launches Genie 2 for Interactive 3D Worlds, and OpenAI Introduces ChatGPT Pro with Advanced o1 Reasoning Model. Amazon Web Services has announced an "Ultracluster" supercomputer using its homegrown Trainium chips, to be used by AI startup Anthropic in 2025. The company also unveiled its new Ultraserver, featuring 64 interconnected chips. These developments, along with Apple joining as a chip customer, demonstrate AWS's strategy to compete with Nvidia in the AI chip market, currently valued at $117.5 billion. The initiatives are driven by AWS's Austin-based Annapurna Labs, which focuses on creating cost-effective alternatives to Nvidia's GPUs. The company is already working on Trainium3, promising four times more power than Trainium2. Meanwhile, Google DeepMind has unveiled Genie 2, a foundation world model that can generate interactive 3D environments from single prompt images. Building on Genie 1's 2D capabilities, this new model creates playable worlds that respond to keyboard and mouse inputs, suitable for both human players and AI agents. Trained on large-scale video data, Genie 2 demonstrates emergent capabilities in physics, object interactions, and character animation. The model works with Imagen 3 for text-to-image prompts, enabling the creation of diverse training environments for embodied AI agents, with simulations lasting up to one minute. Also, OpenAI has launched ChatGPT Pro, a new $200 monthly subscription tier featuring their advanced o1 reasoning model (codenamed "Strawberry"). The upgrade includes o1 pro mode, which offers enhanced problem-solving capabilities with a 34% reduced error rate, particularly excelling in data science, programming, and case law analysis. The service includes unlimited access to o1-mini, GPT-4o, and Advanced Voice features. OpenAI is also awarding Pro grants to medical researchers. The premium tier targets professional users needing research-grade intelligence.
Recommended by LinkedIn
Quantum Technology] China Unveils 504-Qubit Tianyan-504 Quantum Computer, IonQ Launches First European Quantum System in Switzerland, and AWS Partners with NVIDIA for Advanced Quantum Simulations on Braket. China has unveiled the Tianyan-504, its most advanced quantum computer featuring a 504-qubit chip called "Xiaohong." Developed by China Telecom Quantum Group, Chinese Academy of Sciences, and QuantumCTek, the system reportedly matches IBM's performance in key metrics like qubit lifetime and readout fidelity. It will be integrated into China Telecom's Tianyan quantum cloud platform, launched in 2023. This development follows China's previous quantum achievements, including the Jiuzhang and Zuchongzhi systems. The platform has attracted over 12 million visits from users in more than 50 countries. Meanwhile, IonQ has launched its first European quantum computer, Forte Enterprise, at the uptownBasel campus in Switzerland, in partnership with QuantumBasel. The system has achieved a record algorithmic qubit count of AQ #36, capable of processing over 68 billion possibilities simultaneously. This marks IonQ's first data-center-ready quantum computer outside the United States and Switzerland's first commercial quantum system. Operating from data centers in Washington D.C., Seattle, and now Basel, IonQ's Forte Enterprise will provide European enterprises, governments, and research institutes with local access to advanced quantum computing capabilities. Also, AWS and NVIDIA have partnered to integrate NVIDIA's CUDA-Q quantum development platform into Amazon Braket, enhancing hybrid quantum-classical computing capabilities. The integration enables up to 350x faster quantum circuit simulations using NVIDIA GPUs compared to CPUs. Researchers can seamlessly transition from simulations to quantum hardware from providers like IonQ, Rigetti, and IQM. The partnership focuses on developing infrastructure for future quantum computing challenges, including error correction and AI-enabled simulations. The pay-as-you-go model allows researchers to access powerful GPU resources and quantum devices without managing complex infrastructure.
[General technology] DroidBot Targets Banks and Digital Assets Platforms, JFrog Uncovers Security Flaws in Top ML Frameworks, and Europol Dismantles Manson Market and Related Cybercrime Networks. A new Android remote access trojan called DroidBot has targeted 77 banking institutions, cryptocurrency exchanges, and national organizations. Discovered in October 2024 but active since June, the malware operates as a service for $3,000 monthly, with 17 affiliate groups identified. DroidBot combines VNC and overlay attack techniques with spyware capabilities, using dual-channel communication (MQTT and HTTPS) for enhanced flexibility. The malware has primarily affected European countries and is distributed through fake security apps and banking applications. Meanwhile, researchers at JFrog have identified multiple security vulnerabilities in popular open-source machine learning frameworks including MLflow, H2O, PyTorch, and MLeap. The flaws could enable code execution and affect ML clients, potentially allowing attackers to perform lateral movement within organizations and access sensitive information. The vulnerabilities include cross-site scripting in MLflow, unsafe deserialization in H20, path traversal issues in PyTorch and MLeap. JFrog warns that even ML models loaded from "safe" formats like Safetensors can achieve arbitrary code execution if not properly vetted. Also, Europol has shut down Manson Market, a clearnet marketplace launched in 2022 for trading stolen financial data obtained through phishing schemes. German-led operations resulted in seizing over 50 servers, 200 terabytes of data, and €63,000 in cash and crypto assets. Two suspects were arrested in Germany and Austria. The marketplace facilitated fraud by selling stolen data sorted by region and account balance, causing losses exceeding €250,000 to 57 victims. The operation involved six European countries and coincided with takedowns of Crimenetwork marketplace and MATRIX messaging app.
If you enjoyed today’s QX Snapshots, consider subscribing! 𝘑𝘰𝘪𝘯 𝘰𝘶𝘳 +500 𝘵𝘳𝘶𝘴𝘵𝘦𝘥 𝘳𝘦𝘢𝘥𝘦𝘳𝘴 𝘢𝘵 𝘜.𝘚 𝘉𝘢𝘯𝘬, 𝘌𝘶𝘳𝘢𝘴𝘪𝘢𝘯 𝘋𝘦𝘷𝘦𝘭𝘰𝘱𝘮𝘦𝘯𝘵 𝘉𝘢𝘯𝘬, 𝘌𝘠, 𝘔𝘰𝘳𝘨𝘢𝘯 𝘚𝘵𝘢𝘯𝘭𝘦𝘺, 𝘉𝘕𝘠 𝘔𝘦𝘭𝘭𝘰𝘯, 𝘔𝘪𝘤𝘳𝘰𝘴𝘰𝘧𝘵, 𝘝𝘦𝘳𝘪𝘻𝘰𝘯, 𝘍𝘢𝘤𝘦𝘣𝘰𝘰𝘬, 𝘚𝘦𝘳𝘷𝘪𝘤𝘦𝘕𝘰𝘸, 𝘊𝘰𝘯𝘴𝘦𝘯𝘚𝘺𝘴, 𝘗𝘰𝘭𝘺𝘨𝘰𝘯 𝘓𝘢𝘣𝘴, 𝘋𝘶𝘬𝘦 𝘜𝘯𝘪𝘷𝘦𝘳𝘴𝘪𝘵𝘺, 𝘚𝘬𝘺, 𝘐𝘯𝘧𝘰𝘳𝘮𝘢, 𝘍𝘳𝘦𝘴𝘩𝘧𝘪𝘦𝘭𝘥𝘴 𝘉𝘳𝘶𝘤𝘬𝘩𝘢𝘶𝘴 𝘋𝘦𝘳𝘪𝘯𝘨𝘦𝘳, 𝘚𝘢𝘯𝘵𝘢𝘯𝘥𝘦𝘳, 𝘢𝘯𝘥 𝘰𝘵𝘩𝘦𝘳𝘴.