TLDR IT 2026-04-23

Google used Cloud Next to put AI agents at the center of its enterprise strategy, wrapping its tooling into a broader Gemini Enterprise push and pairing that with new TPU chips for training and inference. The bigger takeaway is that Google is trying to sell a full stack for enterprise AI adoption now: models, infrastructure, governance, and deployment.

Anthropic is investigating reported unauthorized access to its unreleased Mythos model through a third-party vendor environment. The story matters less as a one-off breach and more as a reminder that model access controls, vendors, and internal governance are becoming part of the enterprise security surface area.

More than 1,300 internet-exposed Microsoft SharePoint servers remain vulnerable to a spoofing flaw tied to April Patch Tuesday, despite active attacks. It is another sign that patch latency is still one of the most painful operational risks in enterprise IT, especially for older on-prem collaboration infrastructure.

AI integration shifts SaaS gross margins from traditional 70–80% levels toward 52% by 2026. CFOs must isolate AI COGS, align pricing with variable usage, and provide clear margin roadmaps to maintain valuation. Tracking cost per workflow and active user is essential for scaling AI products profitably.

There's a growing tension between flashy AI product rollouts and rising customer concern over trust, accuracy, and security. The mood is shifting from “who has AI features” to “which vendors can prove those features are safe enough and useful enough to run inside real enterprises.”

Most enterprises are overpaying because their workloads are in the wrong place. US Signal, a Microsoft Solutions Partner, created a practical 90-day cloud roadmap to help IT leaders evaluate, optimize, and place workloads across Azure, private cloud, and hybrid environments. Download the free workbook and start building a smarter cloud strategy.

Wiz is expanding its AI-APP platform with the Red Agent in public preview, a new Technology Intel Center, and enhanced Wiz Code features. These updates provide visibility into AI frameworks, secure AI-generated code, and integrate edge security from providers like Cloudflare and Akamai to mitigate systemic risks.

Salesforce and Google Cloud announced deeper integrations so AI agents can take action across both platforms, including Slack, Google services, and Agentforce workflows. This is the kind of partnership that could matter for IT teams because cross-platform action is where governance and identity questions get much more real.

Silobreaker Mimir integrates agentic AI into intelligence workflows to accelerate research while maintaining governance and transparency. The platform uses verifiable source material and an MCP-based integration layer to securely deliver evidence-backed intelligence to enterprise tools, ensuring decision-makers receive auditable, context-rich insights aligned with priority intelligence requirements.

The NVD has significantly reduced its vulnerability enrichment scope, leaving security teams vulnerable as AI tools like Claude Mythos accelerate exploit discovery. Organizations must shift from CVSS-based prioritization to proactive strategies, including stateless service architectures, runtime protection, and third-party intelligence to mitigate risks in a fragmented ecosystem.

Thousands of exposed Apache ActiveMQ servers are reportedly vulnerable amid ongoing exploitation, adding another item to the patching queue for enterprise teams. Messaging infrastructure is not flashy, but when it is exposed and neglected, it becomes exactly the kind of foothold attackers love.

Grafana is widening access to its AI assistant, a move that could put more AI-assisted observability tooling in front of self-managed teams.

Microsoft says Message Center posts will now include more timely summaries and dedicated security and compliance sections for admins.

Open AI-driven security systems help defenders keep up with attackers by enabling faster, shared detection and fixes.

Progress patched multiple high-severity issues, including a firewall bypass bug in MOVEit WAF.