AI & Machine Learning– category –
-
AMD Unveils Instinct MI455X AI Accelerator with 432GB HBM4 Memory, Surpassing Nvidia’s Rubin GPU, to Challenge Dominance in Data Centers with CDNA 5 and Helios Rack-Scale Architecture
Tom's Hardware USA Overview AMD has announced the Instinct MI455X AI accelerator, integrating its CDNA 5 and Helios rack-scale architectures. This new product features a substantial 432GB of HBM4 memory, significantly exceeding the 288GB... -
VELMENI Secures FDA 510(k) Clearance for AI-Powered 3D CBCT Dental Imaging Software ‘VELMENI for DENTISTS (V4D) 3D,’ Expanding Dental AI Portfolio
PR Newswire USA Overview VELMENI announced that its AI-powered 3D Cone Beam Computed Tomography (CBCT) dental imaging software, 'VELMENI for DENTISTS (V4D) 3D,' has received 510(k) clearance from the U.S. Food and Drug Administration (FD... -
Google DeepMind Advances AGI Path with ‘Genie 3,’ a Generative World AI Model Capable of Real-time 3D Environment Generation and Enhanced Reasoning
Wisdomplexus USA Overview Google DeepMind has developed "Genie 3," a generative world AI model designed to deeply understand and simulate physical environments. Genie 3 can generate high-quality 3D environments in real-time from text pro... -
Moonshot AI Releases Kimi K3, a 2.8 Trillion Parameter Multimodal AI Model with a 1 Million Token Context Window, Slated for Open-Weight Release
Build Fast with AI China Overview China's Moonshot AI launched Kimi K3 on July 16, 2026, a 2.8 trillion parameter Mixture-of-Experts (MoE) multimodal AI model. This model features an unprecedented 1 million token context window and suppo... -
Google DeepMind Unveils Gemini 3.5 Flash Cyber AI, a Specialized Model for Vulnerability Hunting, Outperforming DeepSWE Benchmarks for Government and Trusted Partners
Security Affairs USA Overview Google DeepMind has launched "Gemini 3.5 Flash Cyber," a new AI model specialized in software vulnerability discovery and patching. Built on the existing 3.5 Flash architecture, the model demonstrated 49% pe... -
Former OpenAI CTO Mira Murati’s Thinking Machines Lab Unveils ‘Inkling,’ a 975-Billion Parameter Open-Weight Multimodal AI Model
BigGo Finance USA Overview Thinking Machines Lab, founded by former OpenAI CTO Mira Murati, has introduced 'Inkling,' its first open-weight multimodal AI model. Inkling employs a 975-billion parameter Mixture-of-Experts (MoE) architectur... -
EU AI Act: High-Risk AI Systems Granted Grace Period Until 2028, While Generative AI Transparency Obligations Take Effect August 2, 2026
Eversheds Sutherland EU Overview The European Union's AI Act has extended compliance deadlines for high-risk AI systems until December 2027 or August 2028, but transparency obligations for generative AI output and deepfake content will a... -
Nvidia Details Rubin Architecture Optimizations for AI Inference: Tackling Memory Bottlenecks and HBM Shortages Amid Intensifying AI Chip Race
Tom's Hardware USA Overview Nvidia has unveiled comprehensive inference optimization strategies for its next-generation Rubin AI accelerator architecture, focusing on efficient management of Mixture-of-Experts (MoE) models and improved d... -
TSMC Achieves Record $22 Billion Q2 Profit with 77% Surge on AI Chip Demand, Commits $100 Billion Additional Investment in Arizona, and Expands CoWoS Capacity by Over 50%
ServeTheHome Taiwan Overview TSMC, the world's largest contract chipmaker, reported a record net profit of $22 billion in Q2 2026, marking a 77% year-over-year increase, driven by booming AI chip demand. The company raised its full-year ... -
HoneyNaps Receives FDA 510(k) Clearance for AI Software SOMNUM V3.0 for Subtype-Specific Sleep Apnea Classification
Sleep Review USA Overview HoneyNaps, a South Korean company, announced that its AI diagnostic software, SOMNUM V3.0, has received 510(k) clearance from the U.S. Food and Drug Administration (FDA). The software automatically detects and c...
