memory
30 articles about memory in AI news
DeepMind Engineer Flags Memory Crisis at FMS 2026
A DeepMind engineer's FMS 2026 report warns that HBM bandwidth growth is lagging compute, creating a memory crisis that threatens frontier AI scaling. The report argues for new memory architectures and highlights supply chain constraints.
Why Your GPU's Memory Ceiling Predicts Your Cloud Inference Costs
Towards AI explains how GPU VRAM constraints, not compute, dictate LLM inference costs locally and in the cloud. The article details memory math, 2026 hardware shortages, and pricing trends, urging teams to manage context and quantization.
MobileMem: On-Device Memory From a Year of Phone Data
MobileMem trains an LLM on a year of mobile data for on-device memory. Paper and code released, but no benchmarks disclosed.
Spatial Memory Agent: Frozen VLMs Learn From Experience
Spatial Memory Agent improves spatial reasoning in frozen VLMs via experience distillation and calibrated retrieval, without parameter updates. No benchmarks disclosed, so efficacy unverified.
Frozen-Weight AI Agents Degrade via 'Memory Reward Inflation'
New arXiv paper shows frozen-weight agents endorse 31%-54% of wrong answers, compounding errors via memory. Echo Gap resists stronger LLMs.
Agent Memory Benchmarks Get a Shared 5,000-Question Test
A 20+ institution consortium launched a shared agent memory benchmark with 5,000 questions and one answer model, targeting the attribution problem in memory startup claims. First rankings due mid-August.
China's DFSX SuperNode Doubles GB200 Memory Bandwidth on 14nm
China's DFSX SuperNode claims 2x GB200 memory bandwidth using 14nm vertical towers, but lacks benchmarks or ship dates, warranting skepticism.
Lumentum CEO: Indium Phosphide Shortage Will Top Memory's
Lumentum CEO warns indium phosphide shortage will exceed memory's; Nvidia invested $2B each in Lumentum and Coherent to secure laser supply.
Apple Asks Trump to OK Chinese Memory Chips; Micron Warns of Industry Collapse
Apple asked Trump to approve Chinese memory chips; Micron warns it could destroy US industry. Prices quadrupled, margins exceed 80%.
Genspark AI Workspace 6.0 Adds SecondBrain Memory Layer
Genspark launched AI Workspace 6.0 with SecondBrain, a persistent memory system for retaining user context across sessions, per @rohanpaul_ai.
Supermicro's 7,000-lb Double-Wide Rack Targets B300 Memory Squeeze
Supermicro's 7,000-pound double-wide rack targets B300 memory squeeze with denser GPU packing and liquid cooling, announced at Computex 2026.
NUS CIMERA Chip Cuts LLM Memory Wall with Compute-in-Interconnect
NUS researchers propose CIMERA, an LLM inference accelerator integrating compute-in-interconnect and memory to mitigate the memory wall, detailed in arXiv:2607.13649 (July 2026).
Apple M7 Ultra Chip Reportedly Supports 1.5TB Unified Memory
Apple's M7 Ultra chip reportedly supports 1.5TB unified memory, doubling the M3 Ultra and matching eight Nvidia B200 GPUs, but DRAM supply constraints threaten pricing.
How to Build Safer DevOps Workflows with Claude Code, MCP, Hooks, and Memory
Claude Code hooks, MCP servers, and memory create self-regulating DevOps workflows. Use Bash hooks to block dangerous commands and memory to persist safety rules.
GraphRAG Memory Design: Retrieval Over Storage, MCP Integration
Agent memory design prioritizes retrieval over storage, using unified MongoDB and MCP server. GraphRAG enables multi-hop traversal via three strategies.
Nvidia Vera Rubin Rack Costs $7.8M; Memory Drives Price
Nvidia's Vera Rubin rack costs $7.8M with memory as key cost driver; Kyber rack delayed to 2028.
Second Brain Becomes Open-Source AI Agent Memory System
Pauliusztin open-sourced a 2-year project turning 10,994 notes into a living AI agent memory system. The architecture aligns with Google's new OKF standard.
EvoEmbedding Beats Static Embedders 3× Larger via Latent Memory Queue
EvoEmbedding uses a latent memory queue to beat static embedders 3× its size on long-context retrieval, per @HuggingPapers.
OPID: Agents Learn From Hindsight Without External Memory
OPID lets agents learn hierarchical skills from hindsight, improving sample efficiency on ALFWorld, WebShop, Search QA without external memory at inference.
Tencent Open-Sources Agent Memory System Cutting Token Use 61%
Tencent open-sourced TencentDB Agent Memory, cutting token usage by 61.38% and boosting task success by 51.52% on WideSearch, running fully local.
Micron Backs Anthropic Series H With Multi-Year Memory Supply Deal
Micron invests in Anthropic's Series H and inks multi-year memory supply deal for HBM, DRAM, and SSDs. Critics flag circular arrangement as bubble risk.
Use CLAUDE.md as External Working Memory
Use CLAUDE.md as external working memory: add Stack, Voice, Gotchas, and Current checkpoint sections to cut session restart time from 5–15 minutes to under 1 minute, per self-reported data.
DeepSeek-V4 Hits 500K Context with 90% Less KV Cache via FlashMemory
DeepSeek-V4 achieves 500K context with 90% less KV cache via FlashMemory's lookahead sparse attention, keeping only 13.5% of cache in GPU memory without retraining.
OpenAI's ChatGPT 'Dreaming' Memory Retains Preferences Across Sessions
OpenAI launched a dreaming memory system for ChatGPT that retains user preferences across conversations by compressing and replaying session data, enabling persistent personalization.
HydraDB Raises $6.5M for Persistent Agent Memory, Solving the Session Gap
HydraDB raised $6.5M for persistent agent memory, solving the session-gap problem context windows ignored. The round signals memory as a startup thesis.
AgingBench: AI Agents Lose Reliability Over Time & Memory Fails
UT Austin paper finds AI agents degrade over time via memory errors. Proposes AgingBench to measure reliability decay across sessions.
Sleep Phase Cuts Transformer Costs by Consolidating Memory
Paper proposes sleep phase to consolidate context into fixed-size memory, reducing inference cost while improving long-horizon task performance on GSM-Infinite.
Zep AI's Graphiti: Agent Memory Without Schema Is Just Storage
Zep AI's Graphiti enforces Pydantic schemas on LLM entity extraction, preventing generic label collapse and enabling precise querying of agent memory.
Memory as a Model: Augmenting LLMs with Trained Memory
Paper augments LLMs with trained memory for long-term recall. Model-agnostic approach stores external knowledge without retraining.
Neo4j's agent-memory: Open-source unified memory for AI agents via knowledge graphs
Neo4j releases agent-memory, an open-source unified memory layer for AI agents using knowledge graphs, enabling persistent structured recall.