AI inference
PulseAugur coverage of AI inference — every cluster mentioning AI inference across labs, papers, and developer communities, ranked by signal.
3 day(s) with sentiment data
-
AWS and Qualcomm partner on custom AI inference chips and cost reduction
Amazon Web Services (AWS) and Qualcomm have entered into a multi-year partnership focused on reducing the costs associated with AI model inference. This collaboration involves the development of custom silicon chips des…
-
New MeanField model improves AI inference scheduling on GPUs
Researchers have developed a MeanField surrogate model to address the challenge of scheduling concurrent AI inference workloads on shared GPUs. This new approach predicts model performance based on local configuration a…
-
Agentrys raises $24.5M for AI chip design; agentic AI poised to boost token demand
Agentrys has secured $24.5 million in funding for its agentic chip design platform, which reportedly achieves over 90% on NVIDIA benchmarks. This development coincides with a discussion on how agentic AI could significa…
-
Oxmiq Labs proposes High Bandwidth Flash for AI inference capacity
Oxmiq Labs is proposing High Bandwidth Flash (HBF) as a new capacity tier for AI inference, aiming to offer significantly more storage at a comparable cost to High Bandwidth Memory (HBM). Their presentation at Hot Chips…
-
AI frameworks advance silicon photonics design and energy-harvesting AI concepts
Researchers have developed PICasso, an AI-enabled framework that uses natural language to design, verify, and optimize silicon photonic integrated circuits. This framework couples LLMs with domain-specific knowledge and…
-
Samsung integrates logic into LPDDR5X memory for faster AI inference
Samsung has developed LPDDR5X-PIM, a new memory technology that integrates logic units directly into the memory chips. This Processing-in-Memory (PIM) approach aims to reduce the cost and power consumption associated wi…
-
Intel unveils practical AI accelerator Crescent Island for inference
Intel has revealed new details about its Crescent Island AI accelerator, built on the Xe3P architecture, at the Hot Chips symposium. This accelerator is designed for inference tasks, featuring a 350W air-cooled PCIe car…
-
SK Hynix and SanDisk advance High Bandwidth Flash for AI at 2026 FMS Conference
At the 2026 FMS Conference, SK Hynix and SanDisk showcased advancements in High Bandwidth Flash (HBF) technology, designed to enhance AI inference capabilities. SK Hynix introduced new memory tiers, including G1.5 HBF w…
-
Azure Storage enhances AI inference with prompt caching and KV offload
Microsoft Azure Storage is enhancing its capabilities to accelerate AI inference. The platform is implementing techniques such as prompt caching to improve token efficiency and KV cache offloading to Blob storage. These…
-
Kog's new engine boosts AI inference speed on existing GPUs · 4 sources tracked
French startup Kog has developed a new inference engine designed to significantly accelerate AI model performance on existing datacenter GPUs, such as the AMD MI300X and NVIDIA H200. The company's software-driven approa…
-
AI Voice Agent Uses Function Calling to Avoid Hallucinations
A Python-based AI voice agent has been developed to overcome the issue of AI agents hallucinating answers by integrating function calling capabilities. This agent uses Telnyx for speech recognition and text-to-speech, a…
-
SanDisk, SK Hynix unveil High Bandwidth Flash spec for AI memory
SanDisk and SK Hynix have jointly introduced the High Bandwidth Flash (HBF) specification, a new open standard designed to enhance memory capacity and bandwidth for AI inference systems. This technology aims to bridge t…
-
Edge Computing Best Practices for AI Applications Detailed
Edge computing is best utilized for specific AI application functions such as authentication checks, A/B testing, geolocation routing, and rate limiting. However, it is not suitable for computationally intensive tasks l…
-
AI inference profitability debated amid bubble concerns · 2 sources tracked
The profitability of AI inference is a topic of discussion, with some arguing that it is clearly a profitable endeavor. This perspective suggests that the underlying technology and its applications are generating substa…
-
AI Agents and Inference Powering New Tech Frontiers: $2B for 3D Worlds, $13B for Baseten
General Intuition is investing $2 billion in AI agents designed for 3D environments, aiming to revolutionize simulation, gaming, and digital work. Concurrently, Baseten is seeking a $13 billion valuation, highlighting t…
-
AI Inference Demands Scalable Memory Beyond Compute
The AI industry is shifting its infrastructure focus from model training to inference, which presents new challenges in memory management. Unlike training, which is compute-and-bandwidth intensive, inference requires ef…
-
New methodology tackles AI inference emissions in corporate reporting
A new methodology has been proposed to accurately account for the greenhouse gas emissions generated by AI inference services within corporate sustainability reports. This four-tier framework aims to provide a more prec…
-
Pearl blockchain's 'AI mining' claims debunked by research
A new study has debunked the claims of the Pearl blockchain's "Proof of Useful Work" (PoUW) mechanism, revealing it does not contribute to AI inference as advertised. Despite boasting significant computational power, th…
-
Deploy HIPAA-Compliant AI Inference on Self-Managed Infrastructure
This article provides a guide on deploying AI inference services that comply with HIPAA regulations, emphasizing the use of self-controlled infrastructure. It details how to set up a secure environment, manage data priv…
-
Orbital compute viable only for sovereign cloud, analysis finds
Brandon Karpf has analyzed five potential business models for orbital compute infrastructure, including AI training, AI inference, public cloud, content distribution, and edge compute. His research indicates that only t…