Anthropic

Global – October 24, 2026 -- Agentic intrusions tripled in the second half of 2026 compared with the first half, according to CrowdStrike data, as threat actors hand hacking decisions and command execution to AI agents.

CrowdStrike's report, published on Wednesday, documents four threat actors. Two used sophisticated AI harnesses, the software that connects a large language model to external resources. Two used simpler setups. The firm said the main constraint on scaling such campaigns is "token budgets," not engineering complexity.

Spanish- and English-speaking actor breaches at least 15 Argentine healthcare organizations

The actor used a penetration-testing harness built on the open-source PentestGPT framework, modified with extra hacking scripts. One Anthropic Claude Code agent acted as supervisor and assigned tasks to a second agent. CrowdStrike observed the activity between mid-May and mid-August.

Cambridge – October 19, 2026 -- AI agents left alone on a virtual island for 30 hours outperformed fresh copies on a specific task, according to a new MIT paper that has not yet been peer-reviewed. The agents, copies of Claude Opus 4.7 nicknamed Eko, had no instructions beyond a prompt describing them as "naturally curious."

Researchers placed the agents on a simulated island loosely based on Roblox. Eko wrote instructions that moved its avatar and explored the limits of the world. It set its own goals, including climbing the map's highest point and testing how high it could throw a block.

Global -- Coalition sent just 49 Zero-Day Alerts (ZDAs) between May 1 and August 1, 2026, even as Microsoft's July Patch Tuesday swelled to 569 vulnerabilities, and 94% of its policyholders did not receive a single one.

Microsoft's Patch Tuesday count nearly triples in one month

June's Patch Tuesday set a record with 206 vulnerabilities. July's release nearly tripled that to 569. Before Microsoft adopted multi-model AI for vulnerability discovery, it typically shipped around 100 fixes a month. Windows said in July that it would apply AI across security analysis to find flaws earlier and at greater scale, which Coalition says likely contributed to the jump.

Attackers are also moving faster. Anthropic disclosed earlier this year that Mythos could create working exploits for newly discovered vulnerabilities in hours rather than weeks.

Global – October 17, 2026 -- UKG has announced a headless AI vision that lets employees, managers and leaders start workforce activities from AI experiences including Anthropic, Google Gemini, OpenAI and consumer voice assistants, while its Workforce Operating Platform keeps control of execution.

UKG separates the AI interface from the execution layer

Under the headless approach, the experience where a request begins is decoupled from the platform that carries it out. Through Model Context Protocol tools, pre-built connectors and related technologies, approved AI experiences can reach selected UKG capabilities and trigger workforce workflows.

The UKG platform continues to enforce permissions, policies, approvals and business logic. Customers can build on existing technology investments rather than adopt another standalone application.

CHICAGO – October 16, 2026 -- Eleven vendors are new or returning on Tercera's fifth annual Tercera 30 ranking, the most turnover the growth equity firm has recorded in its Market Anchors category, as Anthropic, OpenAI and Palantir join that group for the first time.

The 2026-2027 list ranks the software and AI vendors and partner ecosystems with the greatest services growth potential for IT services firms. Tercera, an investment and advisory firm for AI and technology service firms, said the list reflects the shift across IT services in the AI era.

Single-vendor focus halves to 5% of IT services firms in Tercera survey

In Tercera's July 2026 survey of 378 technology services leaders, only 5% of firms focused on a single software partner, down from 10% a year earlier. Firms are moving away from single-vendor platform implementations. They now orchestrate data, systems and agents around domain processes and defined business outcomes.

Global – October 15, 2026 -- The main constraint on cyber defense is moving from vulnerability discovery to organizational decision speed, according to CGI's Pierre Le Calvez, as frontier models such as Anthropic's Claude Mythos and OpenAI's GPT-5.6 Sol advance.

Approval delays outpace detection in security operations

In a blog titled "Defending at the speed of frontier AI: Lessons from the Fable shutdown and what to do in the next 60 days," Le Calvez argues that AI-assisted tools now surface vulnerabilities within minutes and clarify their impact within hours. The delays now sit in approval processes, competing business priorities, resource constraints and fragmented ownership rather than inside the security operations center.

Resilient organizations share five operating traits

The blog lists the characteristics of resilient organizations:

Seattle – October 14, 2026 -- Chainguard has launched Chainguard Harness, a cyber harness that runs frontier AI models such as Anthropic's Claude Mythos 5.1 against a live build of a customer's application and delivers only reproduced vulnerabilities, each paired with an engineer-reviewed fix. The product is available through the Claude Marketplace by Anthropic, so enterprises can buy it with their existing Anthropic commitment.

Harness tests running applications in isolated microVMs instead of scanning static code

Harness connects to a team's repository through the Guardener GitHub app, builds the application as it ships, and boots it in an isolated microVM. A frontier model then probes the running instance the way an attacker would. Chainguard says this approach surfaces logic flaws, injection paths and resource exhaustion bugs that appear only at runtime and that traditional static scanners miss.

Singapore – October 14, 2026 -- Protiviti's February 2026 AI Pulse Survey found that 47% of large organizations have incomplete visibility into employees' use of AI tools, a gap Armor Founder and CEO Chris Drake cited at GovWare 2026 while arguing that CISOs should set conditions for AI adoption rather than ban it.

Drake said banning AI does not work and that ignoring it "is how the breach report gets written." In his Tech Talk Spotlights session, he said AI is already arriving through employees' browsers, productivity tools and machines, often without a separate security decision.

Vancouver – October 12, 2026 -- Anthropic committed $250 million across two new initiatives in the past two weeks, while OpenAI rolled out GPT-6 to every ChatGPT user. Most announcements from the two private AI developers targeted large organizations and government customers.

Anthropic launches cheaper Haiku 5.5 and funds two programs worth $250 million

Anthropic introduced Claude Haiku 5.5 on October 7, calling it "the cheapest, fastest, and most capable small model we've ever released." The company said the model costs about 75% less to run on average than Haiku 4.5. It is available on Amazon Web Services, Google Cloud and Microsoft Azure.

Santa Clara – October 12, 2026 -- HCLTech has signed a Global Premier Partnership with Anthropic and will set up a dedicated Anthropic Business Unit with a Center of Excellence (CoE) to speed enterprise deployments of Claude, with a focus on safety, cybersecurity and governance in Agentic AI.

HCLTech's new unit will offer sandbox environments for Claude deployments

The unit will build Agentic AI solutions and provide secure sandbox environments for enterprise Claude deployments through FDEs. HCLTech's services teams will also build industry solutions and offerings powered by Claude, Cowork and Claude Code.

HCLTech will help clients adopt AI-native engineering practices, train their workforces and build the organizational capabilities needed to run Agentic AI safely at scale.