Synapse Software has integrated Google's TurboQuant algorithm into its Synap OCR IX, enhancing document AI while cutting costs. TurboQuant reduces memory use in LLMs and VLMs, facilitating faster processing and supporting CPU environments for enterprises.

NVIDIA's N1 development board, co-developed with MediaTek, has appeared on a second-hand market for about $1,370. Featuring 10 high-performance and 10 energy-efficient CPU cores, it aims to boost gaming and content creation on Arm-based Windows PCs. Official details are expected at GTC 2026.

Gate News message. On April 14, according to Gate market data, as of the time of publication, BLESS (Bless) is trading at $0.02091. In the past 24 hours, it has risen 136.95%, reaching a high of $0.03747 and falling to a low of $0.00844. The total trading volume over the past 24 hours was $5.68M. The current market cap is approximately $38.5129 million. Bless is a decentralized edge computing network that provides on-demand CPU and GPU capabilities for AI, machine learning, and advanced data tools near end users. The network allows users to share their devices’ idle compute resources by installing a desktop application or a browser extension, so they can earn corresponding rewards without special hardware or complex setup. Bless offers one-click access to computing resources, supporting AI model training

SK Telecom, Arm, and Rebellions sign a trilateral memorandum of understanding to jointly develop AI data center inference server solutions, combining Arm’s AGI CPU and the RebelCard chip to optimize high-density inference and large-scale AI deployments.

An Ethereum Foundation researcher presented the zkAttester client solution at the EthCC conference, addressing the computational burden of node re-execution, allowing validators to verify zero-knowledge proofs directly, lowering CPU performance requirements, shortening the time for new node synchronization, and helping enable Ethereum light verification and scaling.

Arm launches its self-developed data center CPU AGI CPU, targeting billion dollars in annual revenue by 2031, transitioning from IP licensing to direct chip sales. Facing competitive risks from self-developed chips, Arm hopes that AI-era computing power demand can support its development.

Nvidia unveiled Vera CPU at GTC conference, its first processor designed for Agent AI and reinforcement learning, featuring 88 proprietary cores with multi-task parallel processing support. Its high-bandwidth memory reduces power consumption, and multiple cloud service providers and hardware manufacturers have already planned integration. It is expected to enhance AI programming and streaming data processing performance, with shipments anticipated in the second half of the year.

Nvidia announced at its GTC conference the launch of the Space-1 space computing module, which integrates 2 Rubin GPUs and 1 Vera CPU, delivering AI inference computing power 25 times that of the H100, and supports on-orbit data processing. Founder Jensen Huang pointed out that thermal dissipation challenges in space remain to be solved. Initial partners include multiple space enterprises, but the launch timeline has not been determined.

Gate News Bot Message, February 12 — According to CoinMarketCap data, as of press time, BLESS (Bless) is reported at $0.00 USD, up 39.74% in the past 24 hours, with a high of $0.01 USD and a low back down to $0.00 USD. The current market cap is approximately $12.4 million, an increase of $3.54 million from yesterday. Bless is the world's first shared computer network. Users can share their device's idle computing resources by installing a browser extension or desktop application to earn rewards. The network supports various applications such as AI model training and inference, gaming infrastructure, data processing, and scientific research computing, while also providing behavior data market insights based on user consent. Bless has over 5 million user-maintained nodes, offering on-demand GPU/CPU computing resources to users worldwide, with ultra-low latency and zero downtime.

The Prysm development team recently released an after-action report detailing the mainnet anomaly that occurred after the Fusaka upgrade on December 4, 2025. The issue once threatened the stability of the Ethereum network but was ultimately resolved thanks to the diversity mechanism of client implementations. The report indicates that the problem occurred at epoch 411,392 (December 4 at 21:49 UTC) after the Fusaka upgrade was activated. Prysm consensus clients, when processing certain proof data, triggered大量历史状态的重复计算, leading to rapid exhaustion of CPU and memory resources, causing nodes to experience denial-of-service (DoS)-style performance degradation. This was not a protocol design flaw but an implementation issue within clients under specific boundary conditions.