Gate News message, April 15 — Synapse Software has integrated TurboQuant, a vector quantization algorithm developed by Google Research, into its Synap OCR IX solution to enhance document AI capabilities while reducing enterprise AI adoption costs.
TurboQuant compresses KV cache used by large language models (LLMs) and vision language models (VLMs), reducing memory consumption when processing long documents. This enables faster data processing on the same GPU infrastructure. The technology addresses memory bottlenecks that previously required expensive GPU infrastructure, particularly for handling extended contexts and large-batch processing without performance degradation.
Synapse Software also supports CPU-based environments for enterprises unable to deploy GPU infrastructure. Through precision profiling, the company achieved CPU-only inference processing of approximately 100 documents per minute with quality loss kept below 1%, lowering total cost of ownership (TCO). The flexible infrastructure approach enables AI OCR adoption in security-sensitive sectors such as finance and public administration, where on-premises deployment and infrastructure constraints have previously hindered AI implementation.
Disclaimer: The information on this page may come from third parties and does not represent the views or opinions of Gate. The content displayed on this page is for reference only and does not constitute any financial, investment, or legal advice. Gate does not guarantee the accuracy or completeness of the information and shall not be liable for any losses arising from the use of this information. Virtual asset investments carry high risks and are subject to significant price volatility. You may lose all of your invested principal. Please fully understand the relevant risks and make prudent decisions based on your own financial situation and risk tolerance. For details, please refer to
Disclaimer.
Related Articles
Perplexity AI 助理 Personal Computer 正式上線！會比 Claude Cowork 好用嗎？
Perplexity has launched its new "Personal Computer" feature, integrating with Mac applications to automate tasks across local files, native apps, and browsers. This service is available to Perplexity Max subscribers and includes advanced automation capabilities and remote access via iPhone.
ChainNewsAbmedia6h ago
Uniswap Labs Launches Developer Platform with AI Tools and Expanded API Access Across 18 Blockchains
Uniswap Labs launched a Developer Platform offering tools for DeFi integration, including an AI toolkit, API playground, and liquidity management features. With over 500,000 developers accessing resources, the platform aims to simplify integration and enhance developer capabilities.
GateNews6h ago
Gulf Edge Partners with Kore.ai to Deploy Enterprise AI Solutions in Thailand
Gulf Edge Co. has partnered with Kore.ai to utilize AI for enterprise solutions in Thailand's telecom, banking, insurance, healthcare, and public sectors. The collaboration aims to enhance AI adoption as Thailand's AI market is projected to grow significantly by 2030.
GateNews7h ago
Hesai Launches KOSMO, an AI-Powered 3D Spatial Intelligence Hardware for Physical AI Training
Hesai Technology has launched KOSMO, an AI-powered device for comprehensive 3D spatial capture, aimed at advancing robotics with high-quality data. The company also announced its entry into the robotic power module market.
GateNews7h ago
Agile Soda Launches Agentic OCR Platform with 98% Document Classification Accuracy
Agile Soda launched Agentic OCR, an AI-driven document automation platform that eliminates pre-training and allows for instant deployment. It offers high accuracy in classification and extraction, improving continuously through user corrections, with plans for future enhancements.
GateNews8h ago
American Express to Acquire AI Expense Startup Hyper in Q2 2026
American Express will acquire AI startup Hyper to enhance its expense management tools for commercial clients. The acquisition, expected to close in Q2 2026, follows a partnership that launched a co-branded rewards card in 2024.
GateNews10h ago
