Bangko Sentral ng Pilipinas (BSP) aims to suspend new registration of Payment System Operators (OPS) for 12 months to review classification and licensing frameworks. Pending applications will be evaluated but not approved. The draft also mandates stricter controls for Merchant Acquirers serving Virtual Asset Service Providers (VASPs), including enhanced due diligence and risk-based measures. If passed, the rule takes effect 15 days after publication; BSP is now soliciting public comments.

Ethereum execution client ethrex has announced the launch of a public testnet on Glamsterdam to test EIPs tied to the Ethereum privacy roadmap. The testnet supports five EIPs: EIP-8141 (frame transactions), EIP-8250 (key randomness), EIP-8272 (recent root), EIP-7805 (FOCIL), and EIP-8369 (VOPS profile). Each implementation meets the latest specification requirements. The frame transaction implementation aligns with the latest execution spec test release. The testnet is permissionless, allowing anyone to sync, establish peer connections, transact, and fund accounts via a faucet.

NVIDIA announced the RTX Spark device powered by the N1X chip, shipping in October, featuring up to 128GB unified memory and Blackwell GPU delivering petaflop-level performance. The company also released the open-source NVIDIA PAIR tool, which pools idle computing power across home devices to accelerate AI agent tasks, cutting completion time by over half.

Binance Futures said it will revise the funding rate settlement schedule for nine USD-M TradFi perpetual contracts starting at 16:15 on Sept. 4, 2026, Beijing time. Under the change, settlement will move from once every eight hours to once every four hours. Binance also said the upper and lower limits for the funding rate on the affected contracts will be adjusted to ±1.00%. The contracts listed in the notice are KODEX200USDT, NAVERUSDT, LGELECTRONICSUSDT, HANMIUSDT, SAMSUNGEMUSDT, CXMTUSDT, ZHONGJIUSDT, CSOPSAMSUNG2LUSDT and CSOPSKHYNIX2LUSDT. The update was disclosed in an official announcement cited by BlockBeats on Sept. 2.

ETHShanghai 2026 will take place in Shanghai on Sept. 22 under the theme “Ethereum Renaissance.” The conference will examine Ethereum’s early history, the values the crypto industry may have lost through speculation and attention-driven markets, the relationship between AI and Web3, and possible directions for the ecosystem’s next phase. Its program will also address Ethereum’s changing identity as it works through coordination between Layer 1 and Layer 2 scaling, governance questions, the link between ecosystem growth and ETH value, and its broader role as infrastructure for an on-chain economy. The event will use the CROPS framework of censorship resistance, capture resistance, open source, privacy and security as a recurring set of values. Two roundtables will focus on open-source AI, decentralized compute, data sovereignty, verifiability, agents, autonomous identity and privacy. ETHShanghai has opened global sponsorship applications with Diamond, Platinum and Gold tiers. The event is organized by ETHPanda, PANews and The Excited Few, with support from GCC, Wanxiang Blockchain Labs, LXDAO, Web3Buidler.Tech and ETH HK Hub.

Shenzhen-based AI hardware startup Violoop has completed an angel round and a pre-A round worth more than RMB 100 million, according to IT Juzi, with backing from Lenovo Capital, CICC Porsche, BlueRun Ventures, Genesis Capital, Qifu Capital and Zero2IPO Holdings. Xiangyang Capital acted as the long-term exclusive financial adviser. Rather than competing head-on with major tech firms for software entry points on the desktop, Violoop is building a handheld-sized external device that connects to a computer through HDMI and USB, reads screen information, and carries out actions through keyboard and mouse signals. The company says the goal is a physical personal AI agent that combines always-on context, long-term memory, local inference, and hardware-based authorization. Founder He Jialin told IT Juzi that the company wants to move beyond prompt-driven interaction and build what it calls “Artificial Intuition,” where the system prepares replies, gathers files, and organizes tasks before a user explicitly asks. Violoop says part of that depends on lifting on-device AI compute from 6 TOPS to 26 TOPS and keeping some key active interactions within one second. The company plans to launch through overseas channels in mid-September, begin a first small-scale domestic sale, and expand market coverage after October. About 12,000 overseas users have already paid deposits, the report said.

MAGNE.AI is no longer being framed only as a Web3 smartphone project. Public materials reviewed from RootData, regulatory filings and the company’s own disclosures show a broader buildout across devices, edge AI, identity rails and machine payments. On the hardware side, MAGNE Phone Gen1 is listed at the PVT P2 stage after Pre-EVT and EVT P1, with first-batch manufacturing targeted for the third quarter of 2026. The company has also published FCC-linked materials, a 3D phone structure viewer and staged production updates. On the enterprise side, MAGNE AI BOX is being positioned for private AI and edge-compute deployments, with a Release Gate framework, one local run validation for a Qwen3.6-35B-A3B Q4_K_M artifact, and a Gate 1 plan to build three prototypes for performance, thermal, offline and recovery testing. Agent Pay, meanwhile, has shown a local contract demo and testnet evidence, but end-to-end HTTP 402 flows are still under development and production payments are not open. MAGNE.AI disclosed a $2.64 million strategic follow-on round on Aug. 5, 2026, bringing cumulative public financing to $12.64 million by its own account. Even so, the article argues the key questions now are delivery, prototype validation, public testnet payment proof and whether community participation can convert into actual device users.

OpenAI says its in-house Jalapeño inference chip, developed with Broadcom, outperformed Nvidia’s GB300 on power efficiency and response time in internal testing, according to remarks by chip lead Richard Ho at Stanford University’s Hot Chips conference. The chip is designed for inference rather than model training and was formally introduced in June after a development cycle that OpenAI says took nine months from project start to tape-out. A 128-chip Jalapeño system delivers 1.7 exaFLOPS of 4-bit MXFP4 compute, or about 13.4 petaFLOPS per chip, and includes 27.5 TB of HBM4 memory. Technical analysis cited by The Register said the system posted 1.5x to 1.9x higher peak throughput and 1.7x to 3.6x lower end-to-end latency in OpenAI’s custom InferenceX benchmark, with even larger gains in ultra-low-latency scenarios. Bloomberg, however, said the results come with at least four caveats: Jalapeño was not compared with Nvidia’s newer Vera Rubin chips; speculative decoding was excluded from the benchmark; the system trails Nvidia GB200 NVL72 and GB300 NVL72 racks in raw compute and memory capacity; and OpenAI has not disclosed actual power draw. The report added that volumes will remain limited in 2026, with mass production only expected in 2027.