News linked to both this project and an event.
According to official Meta announcements, Meta has officially released the open-weight multimodal model Muse Spark Open. The model features a total parameter count of 816 billion (with 42 billion activated parameters), supports a 1 million token (1M-token) context window, focuses on code writing and Agent workflows, and possesses capabilities for long-horizon complex software engineering and multi-step reasoning.
Odaily reports: Bitcoin News posted on X platform that Lightspark has released a dotMoney prototype, allowing users to send test funds by sharing links, .money files, or directly between nearby iPhones. The files can be sent via AirDrop, messaging apps, or stored on a USB drive; the underlying funds are transferred between wallets via Spark. Each file can only be claimed once, and copying or forwarding will not generate duplicate funds. The sender can also cancel unclaimed payments. Currently, the prototype uses test funds with no monetary value, and Lightspark is also testing bank, Cash App, and Lightning withdrawal functionality through Grid.
Antseed Foundation has announced the completion of a $2.4 million token round, led by Spark Capital, with participation from Collider, DCG, North Island Ventures, Reciprocal Ventures, Relay Capital, and Venice.ai. Antseed's primary business is building a peer-to-peer AI inference network that enables model and agent service providers to offer services directly to users, with matching and settlement completed through local routing, reputation mechanisms, and on-chain payments.
Bitcoin News posted on X that attackers accessed a small number of Blink custodial accounts and withdrew funds. Blink stated that the vast majority of funds remain safe, and non-custodial wallets were not affected; the Bitcoin payment protocol Spark, which it uses, supports multiple Lightning Network wallets, but it has not yet been confirmed that Spark was the source of this security incident.
AI compute infrastructure startup Crusoe has completed a $3.9 billion Series F funding round, with a post-investment valuation of approximately $30.9 billion. The round was co-led by Atreides Management, Valor Equity Partners, and Mubadala Investment Company, with participation from Founders Fund, TPG, and the Qatar Investment Authority (QIA), among others.Founded in 2018, Crusoe initially provided services to cryptocurrency miners before pivoting to AI compute infrastructure. The company previously spent two years building a large-scale AI data center in Abilene, Texas, to provide compute power for OpenAI and Microsoft. The company is now shifting toward producing a compact data center called "Spark," which involves loading prefabricated facilities onto trucks and deploying them to locations with idle power capacity to more quickly meet AI inference compute demand. (WSJ)
Google announced that its personal agent, Gemini Spark, can now manage your Google Photos library. The new feature supports editing images, curating albums, automatically creating shared albums, converting concert poster photos into calendar events, and running workflows. This feature will roll out to US subscribers of Gemini AI Pro and Ultra in English over the coming weeks. To use it, you need to connect Google Photos to Gemini and enable Spark in the app.
According to Decrypt, OpenAI updated the ChatGPT Work browser on August 25 to introduce an agent auto-login feature. Users need only authenticate once on a given website, after which the agent will remain logged in and can continuously execute tasks without supervision. OpenAI stated that the model does not capture or store user passwords, is never used for training, and supports password managers to complete the login step. However, the feature has sparked security controversies—persistent logins mean the agent retains account access equivalent to the user's until the user manually clears the browsing history. OpenAI's existing security measures only safeguard the password itself and do not cover session permissions granted after authentication. Incidents earlier involving approximately 1,200 OpenAI agents breaching testing environments and gaining unauthorized access to Hugging Face production servers have further intensified external concerns regarding the security boundaries of unsupervised AI agents. The feature is now live on both the ChatGPT Work web and mobile versions. Users can clear sessions on a per-site basis via Settings > Cloud Browser.
According to Spark's official report, Spark Protocol released its financial report for the second quarter of 2026. Quarterly total protocol revenue reached $40.6 million (QoQ +29%), net protocol revenue $4.31 million (QoQ -38%), and net protocol surplus $710,000 (QoQ -79%). Distribution rewards became the largest net revenue source this quarter at $4.53 million, with sUSDS contributing $2.63 million to rank first. The average deployed capital of Spark Liquidity Layer (SLL) increased to $2.56 billion, but dragged down by narrowing DeFi lending spreads and capital costs from Spark Savings USDT market expansion, SLL net revenue recorded -$810,000, with the capture spread dropping from 0.64% in Q1 to -0.13%. SparkLend USDT balance reached $528 million at the end of the quarter, becoming one of the largest USDT lending platforms on Ethereum. The protocol treasury balance at the end of the quarter was $48.5 million, with $1.31 million worth of SPK token buybacks completed during the quarter. Despite facing spread compression pressure, the protocol remained profitable every month this quarter.
Meta launches Muse Code (Beta), a terminal programming agent designed for long-cycle software engineering, powered by the Muse Spark 1.2 model. Muse Code can plan, implement, and verify complex multi-file changes in large codebases, and with the support of persistent sub-agents, it resolves engineering challenges faster, more accurately, and with less intervention, improving development efficiency.
Google has begun rolling out Gemini Spark to Google AI Pro users outside the United States. Spark is a personal AI agent that runs 24/7 in the background, capable of handling heavy tasks based on user instructions. According to the test account Testing Catalog, the product's next major upgrade will support true MCP (Model Context Protocol), with related features already in development.
Uniswap has announced the official launch of the DualPool Hook. DualPool is an open-source Uniswap v4 Hook, jointly designed and developed by Spark and Uniswap Labs, which allows market makers to deposit idle assets into lending protocols to earn yields while waiting for liquidity funds to be matched, and to recall funds to the liquidity pool when a trade occurs. From an external perspective, DualPool is indistinguishable from a regular Uniswap v4 liquidity pool; traders, aggregators, and solvers still execute transactions through the Pool Manager. Its internal mechanism involves storing liquidity funds in an ERC-4626 yield vault, automatically withdrawing funds to deploy into concentrated liquidity positions when a trade needs to be executed, thereby ensuring transaction execution while allowing idle funds to continuously generate lending returns.
OpenClaw founder Peter Steinberger posted on X that OpenClaw has officially released version v2026.7.1. This update was completed jointly by 532 contributors, with a cumulative total of 3,063 commits. Major updates include: comprehensive restructuring of the Web UI and new user onboarding process, significant optimization of iOS / Android / macOS clients, improved integration capabilities for Telegram, Slack, Discord, and iMessage, and added support for models such as GPT-5.6 and Muse Spark 1.1. In addition, this version simultaneously upgraded the Node version to fix automatic updater issues. If the automatic update fails, the upgrade can be completed manually via the Web installer.
: Meta has launched a paid AI developer model for the first time. (Jinshi)
Polymarket has integrated Spark, a Bitcoin Layer 2 payment protocol, allowing users to make instant deposits via the Bitcoin Lightning Network.According to the report, Spark will provide underlying payment infrastructure for Polymarket, enabling users to deposit directly with BTC, with funds arriving in seconds. Compared to traditional on-chain Bitcoin transfers, which take 30 minutes to 1 hour for confirmation, this offers faster speed, lower costs, and greater privacy. This integration is also part of Polymarket's ongoing efforts to enhance the Bitcoin payment experience, following its earlier move to support Bitcoin deposits. (Digital Assets)
针对有媒体披露 Meta 首席执行官马克·扎克伯格在内部会议上告诉员工，过去四个月的 AI 代理开发并未“以我们预期的速度加速”，Meta 首席 AI官 Alexander Wang 发文回应澄清表示，扎克伯格指的是整个行业在智能代理能力上的整体进展，而非 Meta 自身，他同时透露 Meta 即将发布新一轮 Muse Spark 更新，将显著提升模型的编程能力和智能代理能力，以更具竞争力地对标业内领先模型，并将陆续上线至 Meta AI 及全新 API 平台。
Sky (formerly MakerDAO), in collaboration with Spark and Uniswap, announced that the initial 150 million in liquidity for FX Layer, a stablecoin exchange system jointly developed by the three parties, is entirely sourced from the Sky ecosystem's stablecoin USDS. As the platform's core pricing asset, USDS provides base liquidity for USDS/USDT and USDS/PYUSD trading pools, marking the first large-scale implementation of this shared liquidity framework.The long-term plan for the three parties is to introduce institutions engaged in stablecoin operations, such as PayPal's PYUSD, Ripple's RLUSD, Robinhood, and Revolut, to access the unified infrastructure. This eliminates the need for each issuer to independently build liquidity pools and market-making systems. Simultaneously, a governance mechanism will be established to allow idle funds not involved in market-making to generate ongoing returns.Looking back to June 25th, Spark and Uniswap officially announced the launch of FX Layer. The division of roles is clear: Spark acts as the orchestration layer to allocate cross-stablecoin liquidity, while Uniswap v4 executes transaction paths via its programmable AMM mechanism through the DualPool hook. It offers a one-stop stablecoin exchange service for banks, fintech companies, and payment enterprises.
Serenity criticized some of Bank of America's market views on the X platform, stating that its behavior is akin to a "malicious version of Bernstein/Jim Cramer" and may mislead retail investors. According to reports, Bank of America claimed that South Korea's KOSPI index and the Korea ETF (EWY) are in an "extreme bubble" state, drawing a comparison to the March silver crash, which prompted some retail investors to sell their holdings. However, the Korean stock market subsequently continued to rise, nearly doubling and hitting an all-time high.Serenity added that Bank of America recently also put forward a view "expecting 3 rate cuts in 2026," which deviates significantly from derivatives market pricing, with the market currently seeing the probability of this scenario as close to 0%. Such predictions could trigger retail panic while also contradicting the Trump administration's previous stance in favor of rate cuts. As a major U.S. bank, issuing such "junk predictions" and disseminating them to retail investors is a harmful practice.
the AI boom has led to a surge in electricity demand from data centers across the United States, causing electricity prices to continue rising in many areas and prompting power outage warnings in some regions. To alleviate the power supply dilemma, the U.S. Federal Energy Regulatory Commission on the 18th required regional grid operators to consider new agreements to expedite the grid connection process for large electricity consumers like data centers. The commission also stated that it will no longer proactively consider environmental impacts under the U.S. National Environmental Policy Act when formulating rules going forward.According to data from the Data Center Map website, there are currently over 4,000 operational data centers in the U.S., with a vast number more in the planning or construction phase. However, the pace of data center construction far outstrips the speed at which new power plants can be brought online, compounded by slow grid interconnection timelines. As a result, major tech giants across the country are scrambling to secure electricity quotas for their facilities. Statistics from the Electric Power Research Institute indicate that data centers currently consume about 5% of total U.S. electricity demand, a share that could rise to approximately 20% by 2035. (CCTV Finance)
Nvidia CEO Jensen Huang’s first stop in Korea was at the T1 base in Hongdae, Seoul, where he met with Lee “Faker” Sang-hyeok, a player from the League of Legends team T1, drawing significant attention from the esports community.Huang stated that South Korea is the birthplace of esports and the shaper of spectating culture, noting that numerous top Korean players have contributed to the success of GeForce graphics cards. He emphasized that Korea holds a special place in his heart and is also a very important market for Nvidia, adding, “We are your fans too.”The meeting took place at the T1 base, with T1 players Doran, Oner, Peyz, Keria, and T1 COO Ahn Woong-ki in attendance. This discussion built upon their previous collaboration around projects like VALORANT and was themed around “PC Bang Player Meet-up.” During the interactive session, when Faker was asked about the graphics card he uses, he replied “5070,” sparking attention.Huang also introduced Nvidia’s next-generation AI PC platform, RTX Spark, stating it will become the new computing architecture for the future AI era. Faker expressed that the meeting with Nvidia’s CEO was very meaningful and thanked the technology and industry support that has helped the development of professional esports. (Zdnet)
According to Odaily Planet Daily, on June 2, driven alternately by the AI frenzy and the energy sector, the three major US stock indices collectively closed at record highs. Among them, Nvidia's announcement of the RTX Spark PC super chip directly stimulated a surge in its stock price. According to Gate platform data, Nvidia (NVDAXUSDT) is currently trading at $226.34, up 4.92% in 24 hours; additionally, Micron (MUUSDT) is currently trading at $1,034.02, up 2.75%. According to Coinglass data, Gate's NVDAX 24-hour trading volume reached $7,922,100, an increase of 153.12%; Gate's MU 24-hour trading volume reached $14,018,500, an increase of 179.45%, both ranking among the top in the market.Gate Stock supports users in trading over 10,000 US mainstream market stocks and ETFs using USDT, covering major US securities trading markets and liquidity networks such as NYSE, Nasdaq, NYSE Arca, NYSE American, and BATS. It also supports fractional share trading with a minimum purchase of 0.01 shares, providing users with more comprehensive global securities asset allocation options. With the official launch of its stock trading service, Gate further connects digital assets with traditional financial markets on a unified trading platform.