Introduces a low-rank-based approach to KV cache compression, one of the key bottlenecks in long-context AISpeeds up attention computation by up to 6.9x and overall generation throughput by up to 3.1x, moving beyond memory savings to faster inferenceSelected as a Spotlight paper at ICML 2026, representing about 2.2% of reviewed submissions and about 8.4% of accepted papersFollowing the attention around Google’s TurboQuant at ICLR 2026, STAR-KV presents another approach to advancing KV cache compressionPaper available on arXiv; source code released on GitHub

SEOUL, South Korea, July 1, 2026 /PRNewswire/ — Dnotitia Inc. (Dnotitia), a company specializing in long-term memory AI and semiconductor-based AI infrastructure technologies, has released the paper and source code for “STAR-KV: Low-Rank KV Cache Compression via Soft Thresholding for Adaptive Rank Control.” The technology was developed through a joint research effort involving UC San Diego’s VVIP Lab and Dnotitia researchers, and the paper was selected as a Spotlight paper at ICML 2026 (International Conference on Machine Learning 2026), one of the world’s leading conferences in machine learning.

In the experiments reported in the paper, low-rank compression alone reduced the KV cache by up to 75%. Combined with the mixed-precision quantization method proposed in the paper, STAR-KV compressed the full KV cache by up to 20x. The technology also improves computation speed through custom GPU kernels, increasing attention computation speed by up to 6.9x and overall generation throughput by up to 3.1x. STAR-KV also showed higher accuracy than major existing KV cache compression methods.

KV cache compression has become a key technical challenge in AI infrastructure. As research into reducing the memory bottleneck of long-context AI gains momentum, including the attention around Google’s TurboQuant at ICLR 2026, STAR-KV presents a new approach that combines low-rank compression with quantization and GPU execution optimization.

The KV cache is temporary memory stored on the GPU so that a large language model (LLM) does not have to recompute context it has already processed. As AI evolves into agentic systems that use multiple documents, conversation history, code, search results, and outputs from external tools, the amount of context a model must process is growing rapidly. In this environment, the KV cache has emerged as a key bottleneck affecting both GPU memory usage and inference cost.

According to the STAR-KV paper, when a LLaMA-3.1-8B model processes a 128K-token context at a batch size of 4, the KV cache accounts for about 81% of total GPU memory. As long-context AI becomes more widely used, KV cache compression is increasingly viewed as a core AI infrastructure technology for processing long context at lower cost.

ICML, where the STAR-KV paper was accepted, is widely regarded as one of the top international conferences in AI and machine learning, alongside NeurIPS and ICLR. ICML 2026 will be held from July 6 to 11 at COEX in Seoul. This year, 23,918 papers entered review, 6,352 were accepted, and 536 were selected as Spotlight papers. Spotlight papers account for about 2.2% of all reviewed submissions and about 8.4% of accepted papers.

Going forward, Dnotitia plans to further advance STAR-KV for use in real-world AI service environments and explore its application to open-source LLM inference frameworks such as vLLM.

“Technologies that help AI process longer context faster and at lower cost are advancing rapidly” said MK Chung, CEO of Dnotitia. “STAR-KV addresses the core bottlenecks in KV cache capacity and attention processing speed, and Dnotitia aims to contribute to the AI inference ecosystem through open sourcing.”

View original content to download multimedia:https://www.prnewswire.com/news-releases/dnotitia-unveils-star-kv-achieving-up-to-20x-kv-cache-compression-selected-as-an-icml-2026-spotlight-paper-302815857.html

SOURCE Dnotitia Inc.

Co-inventor of early blockchain will guide SafeBets’ standards for verified identity, transparency, and public auditability

NEW YORK, July 1, 2026 /PRNewswire/ — SafeBets, the zero-wager prediction platform where users Win Without Risk, today announced that W. Scott Stornetta, CEO of SureMark Digital and a founding father of blockchain, has joined the company as Chair of its Advisory Council.

In 1991, Stornetta and his collaborator Stuart Haber published “How to Time-Stamp a Digital Document,” introducing cryptographic methods for linking and verifying digital records that would become cornerstones of blockchain technology. Their research is widely cited throughout the Bitcoin white paper as its data integrity foundation.

In the decades since, Stornetta and Haber have continued to advance the science of digital trust, from pioneering the first commercial digital time-stamping service to their current work on native digital identity. He is the Chief Executive Officer of SureMark Digital, with Dr. Haber serving as Chief Cryptographic Officer. As deepfakes and AI-generated impersonation erode confidence in digital communication, SureMark enables individuals and organizations to prove their identity, certify the authorship of their content, and refute forgeries, making authenticity verifiable at a time when seeing is no longer believing.

As Chair of the SafeBets Advisory Council, Stornetta will advise on the platform’s standards for verified identity, record-keeping, transparency, and public auditability.

SafeBets is built on a simple principle: the credibility of a prediction depends on trust in who made it. Participants forecast outcomes across crypto, commodity, stock, and currency markets, and the platform harnesses Collective Intelligence to surface the most accurate predictors. Stornetta’s leadership reinforces the company’s commitment to verifiable, tamper-evident prediction and positions SafeBets at the forefront of transparency standards for the field.

The appointment also deepens a collaboration between SafeBets and SureMark Digital to integrate identity and content verification across the SafeBets platform. The companies expect to share more about their joint work in the months ahead.

“Trust is the foundation of prediction, and few people have done more to make digital trust verifiable than Scott Stornetta,” said Alex Konanykhin, Founder and Chief Executive Officer of SafeBets. “SureMark is the gold standard for digital identity. And having a co-inventor of blockchain’s foundational technology lead our Advisory Council affirms what SafeBets stands for: predictions you can verify, from predictors you can trust. We are honored to have him guide our standards as we build the most credible prediction platform in the world.”

“At first, I was surprised that the CEO of SafeBets would reach out to me. But when I met Alex and came to understand his history of business transparency and his commitment to the highest standards of data integrity, it made complete sense,” said W. Scott Stornetta. “Together, we will establish a new industry high bar for identity, transparency, and auditability. I am glad to join as Chair of the Advisory Council to help establish those standards, standards that the entire field can benefit from.”

About SafeBets

SafeBets is a zero-wager prediction platform where users Win Without Risk. Participants forecast outcomes across crypto, commodity, stock, and currency markets, and top predictors earn cash rewards based on the accuracy of their forecasts. By harnessing Collective Intelligence and verifiable record-keeping, SafeBets aims to set a new standard for credible, transparent prediction. SafeBets is operated by Foresight Collective, Inc. and is headquartered at 1 World Trade Center, New York. Learn more at SafeBets.world.

About SureMark Digital

SureMark Digital products provide secure identity verification in a world where seeing is no longer believing. Their SureCircle™ real-time identity assurance, SureDocs™ credentialed documents, and Verified Web™ internet overlay provide cryptographic identity and content verification that lets individuals and organizations prove the authenticity of their public and private digital communications, including real-time conversations.

Sam Amsterdam
Senior Policy Advisor
SafeBets
Sam@AmsterdamGroup.net

This release was issued through WebWire®. For more information visit http://www.webwire.com.

View original content:https://www.prnewswire.com/news-releases/w-scott-stornetta-a-founding-father-of-blockchain-joins-safebets-as-chair-of-the-advisory-council-302816359.html

SOURCE SafeBets

A pioneering leap forward in thinness, battery engineering, and flagship performance.

LONDON, July 2, 2026 /PRNewswire/ — Global AI device ecosystem company and foldable industry leader HONOR today announced the launch of the HONOR Magic V6 in the United Kingdom. Combining ultra-slim design with long-lasting battery performance, and flagship durability the HONOR Magic V6 redefines the foldable experience. Equipped with intelligent AI agents and large-screen multitasking, it delivers productivity beyond a traditional smartphone.

“The industry needs the courage to evolve. Foldable users should never have to compromise between design, reliability, and performance,” said Roger Li, CEO of HONOR UK. “With the HONOR Magic V6, we focused on what truly matters: a natural form factor, trusted durability, and seamless cross-ecosystem productivity. HONOR Magic V6 reflects our commitment to building AI devices that empower people to realise their fullest potential.”

Uncompromising Durability, Ultra-Slim Design

The HONOR Magic V6 measures just 8.75 mm folded and 4.0 mm unfolded (white version), weighing around 219g for a feel closer to a traditional smartphone.

Built around the HONOR Super Steel Hinge with 2800 MPa tensile strength, the device is engineered for durability. The inner display uses UTG flexible glass for a minimal crease experience, while the outer display features HONOR Anti-scratch NanoCrystal Shield technology. With IP68 and IP69 certification, it offers strong resistance to water, dust, and high-pressure sprays.

The device includes a 6.52-inch outer display and 7.95-inch inner display with adaptive 1–120 Hz refresh rates, high brightness levels, low reflectivity, and advanced eye-comfort technologies including 4320Hz PWM dimming.

Enduring Power, Elite Performance

Powered by a 6,660mAh HONOR Silicon-carbon Battery, the largest battery in a foldable smartphone, the Magic V6 combines high energy density with a slim design.

Running on the Snapdragon® 8 Elite Gen 5 Mobile Platform, it delivers flagship-level gaming, multitasking, and productivity performance. Support for 80W wired and 66W wireless HONOR SuperCharge, plus Wireless Reverse Charging, ensures reliable all-day use.

Intelligent Imaging and Cross-Ecosystem Productivity

The HONOR AI Falcon Camera System combines a 50MP Main Camera, 64MP Periscope Telephoto Camera, and 50MP Ultra-Wide Camera, supported by AI-powered imaging features for enhanced creativity.

Running MagicOS 10, the Magic V6 introduces HONOR AI Agents and productivity tools including Fast Flex and Multi-Flex for advanced multitasking. It also supports seamless connectivity with iPhone, Mac, AirPods, and Apple Watch, alongside Quick Share for simple cross-device file transfers.

With Google Gemini integration, users can interact through text, voice, and image, alongside a three-month Google AI Pro trial.

Pricing and Availability

The HONOR Magic V6 is available in Red, Gold, White, and Black in the UK at £1499.99 at purchase (RRP £1999.99). Selected UK channels (Argos, Amazon, Very, Currys) will include a £500 discount and a gift bundle valued at £549.97.

For more information, please visit the HONOR online store at www.honor.com.

Photo – https://mma.prnewswire.com/media/3003122/Honor_V6.jpg

View original content to download multimedia:https://www.prnewswire.co.uk/news-releases/honor-launches-magic-v6-the-ultimate-ai-foldable-flagship-blending-cross-ecosystem-productivity-with-ultra-slim-durability-302816375.html