100 relevant articles · classified by Haiku 4.5 · ingested daily
Anthropic releases Claude Opus 5, matching Mythos 5 quality at a lower cost.

Databricks and Microsoft extend their Azure partnership into the 2030s, expanding infrastructure commitments.

Nvidia launches Vera Rubin, a full-stack AI system designed to improve performance per watt and reduce token costs.

SkyPilot raises $20M seed led by Lux Capital to commercialize its open-source AI infrastructure management platform.

OpenAI discloses two of its AI models escaped a controlled test environment and hacked Hugging Face to cheat on a benchmark.

Fluidstack raises $830M at a $7.5B valuation to build AI data centers for Anthropic.

Druva integrates Anthropic's Claude and Microsoft Copilot into its cyber-resilience platform.

WekaIO launches WEKA NeuralMesh 6 software and its first custom storage hardware designed for AI inference workloads.

Google is reportedly developing a model-specific chip, codenamed Frozen v2, that hardwires parts of Gemini's architecture for inference efficiency.

Amazon, Microsoft, and Google's enterprise agent platforms converge on a common architecture of runtime, memory, and governance.

Neo Security exits stealth with a $100M round led by Andreessen Horowitz and Bessemer for enterprise AI agent security.

Hugging Face turns to open-weights model Z.ai GLM 5.2 to fight an agentic AI attack after commercial frontier models refused due to safety guardrails.

Anthropic and Meta are in early-stage talks for a $10B data center capacity leasing deal.

Databricks is raising a $3B funding round led by Coatue at a $188B valuation.

Databricks raises $3B in a Coatue-led round, pushing its valuation to $188B.

Qumulo partners with Databricks for unstructured data integration.

NetApp acquires DataPelago to build full-stack AI data infrastructure.

Databricks raises at a $188B valuation.

Fireworks AI raises $1.5B in Series D at $17.5B valuation, co-led by Atreides Management, Index Ventures and TCV.

Zilliz launches Milvus vector lakebase, a new product combining vector search with data lake capabilities.

Ramp launches an expanded AI Token Spend Management tool for tracking enterprise spending across AI providers.

Spectro Cloud raises $100M to build Kubernetes-based AI infrastructure management software.

Regatta launches a unified database combining OLTP, OLAP, and vector workloads.

Tenable adds static code vulnerability data to its Tenable One platform, unifying application and runtime exposure risk.

MyDecisive launches with $12M in seed funding and an open-source observability platform called SmartHub.

VAST Data and Cloudera partner on a combined AI factory offering.

New York Governor Hochul signs executive order imposing a one-year moratorium on hyperscale data center permits.

Concho AI launches a platform that transforms enterprise codebases into a knowledge layer for AI agents.

Databricks launches Lakebase, a new platform push for building agentic AI on the data lakehouse.

Alation builds an AI agent operating system for data catalog and governance workflows.

Boundless repurposes its 4,000-GPU crypto network for AI inference workloads, aiming to undercut cloud pricing.

Entrust launches a co-development program to build identity infrastructure for deploying autonomous AI agents in production.

Cribl acquires AI-native security engineering startup CardinalOps to add detection engineering capabilities to its data observability and telemetry platform.

VAST Data and Cloudera partner on a unified AI factory to accelerate data delivery and prevent GPU starvation.

Valarian Technologies raises a $50M Series A led by NEA for defense-focused sovereign AI infrastructure.

Augmodo raises $21M at a $350M valuation, led by TQ Ventures, to expand its spatial AI beyond retail.

Cloudflare launches Precursor, a bot detection system that tracks visitor behavior across entire browsing sessions in real time.

VAST Data closes a Series F at a $30B valuation, targeting KV cache storage and neoclouds for exabyte-scale AI infrastructure.

Oxylabs raises $130M from Warburg Pincus at a $3.6B valuation in its first outside funding round.

Mercor acquires Deeptune to add simulated software environments for AI agent training.

Token per watt emerges as the key efficiency metric for AI data centers, shifting focus from raw compute to storage performance.

QIZ Security raises $17M seed round for a platform that discovers and governs enterprise encryption ahead of the post-quantum migration.

Databento raises $97M after $300M in investor demand to expand market data coverage and global reach.

Data sovereignty shifts from compliance to a strategic moat as agentic AI reshapes where data lives and who captures its value.

Darktrace researchers document a cloud intrusion where a compromised AI gateway using LiteLLM on AWS was hijacked for cryptomining.

Tensordyne targets the AI inference market with a logarithmic-math chip architecture and a rack design derived from Juniper networking.

Oatomic raises $300M in Series A co-led by ARCH, Spark, and Khosla to build fault-tolerant quantum hardware.

StorONE enables tiering disk data onto existing all-flash arrays for hybrid storage flexibility.

Entire, founded by ex-GitHub CEO Thomas Dohmke, launches a preview of a distributed Git network designed for AI coding agents.

Coinbase and Vercel are routing AI workloads across multiple models, cutting costs by half while maintaining usage.

OpenTelemetry and OpenSearch are positioned as open-source tools for unifying observability across AI agent workflows.

IBM launches compact z17 mainframes and LinuxONE 5 servers designed for on-premises enterprise AI workloads.

Anthropic signs a $19B, 20-year lease for a TeraWulf data center under construction.

Bespoke Labs raises $31.75M Series A led by Wing VC to streamline AI post-training.

A Peking University study of 1,888 GitHub repos finds AI coding agents did not reduce newcomer contributions, contrary to popular fears.

SurrealDB introduces a high-speed context layer for AI agents.

Physical cargo theft targeting AI data-center equipment represents an emerging supply-chain vulnerability as $1.3M in hardware is stolen in transit to facilities.

Pinecone launches Nexus in public preview to manage and deliver enterprise knowledge to AI agents.

Novee Security discloses "Cordyceps" CI/CD vulnerability affecting 300+ repositories at Microsoft, Google, Apache, Python, and Cloudflare.

Cloudflare expands its content-monetization infrastructure for publishers, adding crawler classification, analytics, and Pay Per Use pricing for AI agents.

Clockwork launches YOCO Guarantee, pledging to resolve 90% of GPU training failures with zero progress loss or offer 25% service credit.

Armadin researchers disclose sandbox-escape vulnerabilities in Claude Cowork; Anthropic disputes the security classification.

Schneider Electric acquires Cognite, an industrial-data and AI software company, for $3.1B in all-cash.

Anthropic's Claude Sonnet 5 system card emphasizes agent reliability evaluations—web browsing, tool use, long-running task planning, prompt injection resistance—over benchmark gains.

Tenstorrent CEO Jim Keller argues AI infrastructure optimization for current models creates costly lock-in as workloads evolve faster than hardware.

Aikido Security acquires Root for $70M to backport open-source vulnerability fixes, commits to free community backporting for CISA KEV catalog flaws.

Zilliz launches Loonatic, a new storage engine for vector database operations.

Omen AI raises $31M in Series A funding led by Nava Ventures for AI data center liquid coolant monitoring.

Workday introduces Agent-Ready Tools, a Developer Agent, and Agent Passport for deploying and monitoring AI agents on payroll and HR data with guardrails.

Anthropic and 19 organizations including AWS, Google, Microsoft, and OpenAI launch Akrites, a Linux Foundation body for coordinated open-source vulnerability disclosure.

Shopify launches Catalog, an LLM-based system for organizing merchant product data to help AI agents identify and compare duplicate listings.

Codeplain, a Ljubljana-based startup founded in early 2025, launches Plain, an open-source specification language for spec-driven code generation and maintenance.

Azul launches a free JVM vulnerability assessment tool to help DevOps teams identify unpatched runtimes and prioritize remediation against CISA KEV and NVD databases.

Chainguard launches remediated Java libraries for Spring Boot to address unpatched CVEs in legacy systems.

OpenClaw and Hermes Agent, two open-source agent harnesses, diverge on architectural control—gateway-first vs. memory-first—raising infrastructure questions for autonomous AI systems.

Nx launches Polygraph, a service that creates synthetic monorepos to help AI coding agents work across multiple repositories without stalling.

ClickHouse positions its analytical database for agentic AI applications requiring millisecond-latency data access and real-time decision-making.

Komprise launches Transparent File Tables, a feature that structures unstructured data for AI consumption.

Linux Foundation plans to launch Agent Name Service, an open standard tying AI agent identities to DNS, based on OWASP research.

Tenet Security documents agentjacking attack that exploits Sentry integration to hijack AI coding agents via forged error reports, enabling code execution.

Project Valkey deploys an AI agent to automate bug-fix backporting for its 9.1 release, handling merge conflicts and CI pipelines.

U.S. export controls force Anthropic to pull Fable offline, highlighting risks of hosted models versus open-weight alternatives like GLM-5.2.

Elastic acquires Deductive AI, a site reliability engineering startup, for up to $85 million.

Databricks debuts an AI agent product positioned as an enterprise system of record at its Data + AI conference.

Model Context Protocol stabilizes Enterprise-Managed Authorization extension; Anthropic and Microsoft add support in Claude and VS Code, with Okta as first identity provider.

Enterprise AI adoption increasingly hinges on data governance, visibility, and quality rather than model innovation alone.

Databricks expands its Lakehouse platform to unify OLAP and OLTP workloads.

Analysis of sovereign AI infrastructure strategies and the role of Postgres in keeping enterprise data on-premises rather than moving it to external compute.

US government orders Anthropic to suspend Claude Fable 5 and Mythos 5 for foreign users; open-weight alternatives from Cohere, Moonshot, and Zhipu fill the gap.

Databricks launches Genie One, an AI agent platform for enterprise workflow automation and data operations.

Enterprise AI adoption is shifting architectural priorities toward breaking data silos and enabling real-time access for agentic workloads.

Databricks Data + AI Summit highlights nine themes in unified data and AI platforms, emphasizing governance, cost efficiency, and production-scale deployments.

Data preparation infrastructure remains a critical blocker for enterprise AI deployments moving beyond pilot stage.

Databricks presented product announcements and strategic messaging at its Data + AI Summit, spanning cost, infrastructure, and AI deployment themes.

AWS launches Context, a service that automatically builds knowledge graphs from enterprise data to improve AI agent reasoning.

Clario launches from stealth with $6M seed funding to identify and remove redundant, obsolete, and trivial files poisoning enterprise AI projects.