Tech giant IBM has announced a strategic partnership with blockchain service provider Casper Labs to promote integrated Web3 and AI technology services. The two parties will collaborate to deploy generative AI models and optimize large language models (LLM), while using blockchain anti-tampering and transparency features to address key issues such as intellectual property (IP) and governance, and to create a safer and more transparent environment for AI deployment.

IBM has released a new encryption signature technology called IBM Hyper Protect Offline Signing Orchestrator (OSO) for processing digital assets in cold storage, reducing the risk associated with manual programs, while keeping assets at a certain distance from the internet. OSO helps protect high-value transactions by providing additional security layers, including disconnected network operations, time-based security, and electronic transaction approval by multiple stakeholders. IBM's long-term partner in the cryptocurrency field, Metaco, a subsidiary of Ripple, is using the new OSO technology.

IBM has installed a 127-qubit quantum computing system at the University of Tokyo, marking the arrival of the first "utility-scale" quantum system in the region. The system, called Quantum System One, will be used for research in various fields including bioinformatics, materials science, and finance. Meanwhile, Alibaba has decided to close its quantum computing laboratory and donate its equipment to Zhejiang University, citing cost-cutting measures and uncertainty caused by the US partial export ban on computer chips to China. The quantum computing sector is expected to grow by over $5.5 billion between 2023 and 2030, leading some experts to worry about the state of quantum computing research outside of the US and China.

IBM has developed a new prototype AI chip called NorthPole, which is claimed to be faster and more energy-efficient than any other chip available. The chip has achieved a 25 times higher energy metric and a 22 times lower time metric of latency, which could lead to post-GPU performance at a fraction of the cost in energy requirements. The NorthPole chip integrates the memory component onto the processing chip itself, bypassing the von Neumann bottleneck. The chip's benchmark results indicate that it could perform well in tasks such as autonomous surgery, self-driving cars, and robotics-related endeavors. IBM Research is already working on the next chip using the NorthPole architecture.