News

CrowdStrike Brings the Falcon Platform to the Anthropic Claude Marketplace CrowdStrike Protects Customers from Threats Delivered via PHP2Shell Read blog…

Introducing the CrowdStrike Agentic Identity Provider, the Foundation for AI Agent Identity Security CrowdStrike Protects Customers from Threats Delivered…

CrowdStrike and OpenAI Expand Partnership to Secure the Agentic Era CrowdStrike Protects Customers from Threats Delivered via PHP2Shell Read blog…

C3 AI Announces Fiscal First Quarter 2027 Results September 2, 2026 Turnaround on track Bookings increase 73% quarter over quarter REDWOOD CITY, Calif. —…

OpenAI says frontier firms now generate 8.3 times as many output tokens per user as typical firms, up from 2.6 times in January, citing Basis, Clay and Exa.

OpenAI says Astra meets the Critical cyber threshold in its Preparedness Framework, scoring 100% on ExploitBench, and that it delayed parts of the release.

ChatGPT for Healthcare can now pull authorised Epic patient records into chats or sit inside the EHR, plus a plugin for nine official health data sources.

Anthropic's Claude Fable 5.1 is generally available, with Mythos 5.1 on trusted access; it says typical costs fall about 25% on cheaper cache reads.

HiddenLayer has raised a US$100 million Series B led by Delta-v Capital to expand agentic runtime protection, including security for AI coding agents.

Sumsub said its Fraud Prevention Starter Kit returns a single risk level from onboarding signals, citing a 180% rise in sophisticated multi-technique attacks.

CrowdStrike's SafeMind pairs offensive and defensive AI models built on NVIDIA Nemotron; it says its Blue Solano model beat frontier models at 99% lower cost.

SpaceXSI says LatchBio found Grok 4.6 the only model above 50% on both refusing hidden biological hazards and completing routine work, at 59.2% and 64.8%.

OpenAI says ChatGPT Ads hit a $1bn annualised run rate in under 200 days, as self-serve buying opens in India, Europe, the Middle East and North Africa.

LG Uplus builds evaluation datasets first and feeds production traces into Arize AX to improve contact-centre AI agents serving more than 30 million users.

Today, the window between "vulnerability published" and "actively exploited" is shrinking fast - and teams that rely on scheduled scanning are left exposed.