Meta and OpenAI use Arm AI chips as Databricks acquires startups

Key Takeaways

• Arm Holdings is shifting its business model to design and sell its own AI chips for data centers.
• Meta and OpenAI are the first customers for Arm's new AI processor.
• Databricks acquired Antimatter and SiftD.ai to enhance its new AI security product.
• Cisco introduced a new security framework for AI agents, including Duo IAM features and the open-source DefenseClaw.
• AccuKnox launched AI-Security 2.0, a Zero Trust platform with features like a Prompt Firewall.
• Check Point unveiled its AI Defense Plane for enterprise AI, featuring an AI-native security engine.
• RSAC 2026 showcased new AI security products from CrowdStrike, Cyera, and Huntress, which expanded ITDR for Google Workspace.
• Louisiana's health department is considering AI for call centers, potentially saving 25% of its over $40 million annual cost.
• Norway's $2.1 trillion sovereign wealth fund uses AI for investment analysis with human oversight, achieving cost savings.
• Hyatt redesigned Hyatt.com with generative AI to improve sales and customer engagement through intent-based searches.

Cisco creates AI agent security framework for businesses

Cisco has developed a new security framework to help businesses safely use AI agents. The framework addresses three main areas: protecting the world from AI agents, protecting agents from external threats, and detecting AI-related security incidents quickly. New Duo IAM features will help manage agent identities and access, ensuring agents only perform intended tasks. Cisco's AI Defense will also be expanded to test and secure AI agents before they are used. This aims to build trust and prevent security gaps as AI agents become more common in the workplace.

Cisco secures AI agents for business use

Cisco is introducing new security solutions to help businesses confidently use AI agents. These solutions focus on securing AI actions, managing agent identities, and controlling access. The company is extending Zero Trust Access to AI agents, giving them verified identities and limiting their permissions to specific tasks. Cisco is also launching AI Defense: Explorer Edition to help developers test AI agent security. Additionally, a new open-source framework called DefenseClaw will automate security for agents, aiming to speed up secure deployment.

AccuKnox launches AI-Security 2.0 for Zero Trust

AccuKnox has launched AI-Security 2.0, a new platform designed to secure AI models, agents, and data using Zero Trust principles. The system includes features like AI Security Posture Management, AI Detection and Response, and a Prompt Firewall to protect against attacks. It offers continuous discovery of AI systems, real-time analysis of risky AI activity, and automated testing against common AI threats. AccuKnox AI-Security 2.0 provides unified visibility across different cloud environments and supports secure agent workflows.

Check Point introduces AI Defense Plane for enterprise AI

Check Point has unveiled its AI Defense Plane, a system designed to manage and secure enterprise AI systems. This platform offers discovery, governance, and runtime control for AI applications and agentic workflows. It uses an AI-native security engine that analyzes millions of AI interactions to provide protection in under 50 milliseconds. The AI Defense Plane includes modules for securing employee use of AI, managing AI applications and agents, and performing AI red teaming to find weaknesses. This system aims to secure AI behavior in real production environments.

Top AI and Security Products at RSAC 2026

RSAC 2026 is showcasing many new AI and security products, with a major focus on securing AI agents. CrowdStrike introduced Falcon Data Security for AI data protection and Agentic MDR using AI agents for automated security workflows. Cyera launched Browser Shield for AI security and a data lineage product for tracking agent activities. Huntress expanded its ITDR service to support Google Workspace. Torq also unveiled its Agentic Builder to improve security operations team productivity.

Arm Holdings to sell its own AI chips

Arm Holdings, known for licensing its technology, is changing its business model to design and sell its own computer chips. The company announced its first product: a microprocessor specifically for data centers running artificial intelligence tasks. Meta, Facebook's parent company, helped develop the chip and is its first customer, along with OpenAI. This move marks a significant strategic shift for Arm, which has historically focused on selling intellectual property rather than physical products.

Arm launches new AI chip with Meta and OpenAI

Arm, a company owned by SoftBank, is launching its own AI chip, marking a major change in its business strategy. Meta and OpenAI have signed on as the first customers for this new AI processor. This move signifies Arm's shift from licensing its technology to selling its own designed silicon. The company aims to capture a larger share of the growing AI market with this new product.

Ophthalmologist to discuss AI's role in eye care

Dr. Kerry D. Solomon will present the Cornelius D. Binkhorst, MD, Lecture at the ASCRS Annual Meeting, focusing on artificial intelligence in ophthalmology. Dr. Solomon, an early pioneer in using AI for eye care with the Veracity Surgical Planner, will discuss the current state and future of AI in the field. He plans to address how AI can help surgeons manage an increasing patient population and improve efficiency. The lecture will also explore responsible AI implementation and future advancements like robotic AI.

Beehiiv lets creators manage accounts via AI

Beehiiv is introducing a new feature that allows creators to manage their accounts using AI platforms. This update aims to help creators automate tasks like analyzing email lists and managing marketing outreach. The first version of the Beehiiv MCP will support backend functions such as subscriber analysis and SEO optimization. Future plans include expanding the technology to offer more direct writing support. Creators can also set privacy controls to prevent AI chatbots from training on their account data.

Norway wealth fund uses AI for decisions with human oversight

Norway's $2.1 trillion sovereign wealth fund is exploring the use of AI for investment decisions, but with human supervision. Currently, about half of its employees use AI tools, primarily for gathering information and analysis. These tools help monitor companies for risks and prepare for meetings. While the fund anticipates AI agents will eventually make some limited decisions autonomously, human oversight will remain crucial. The fund has seen significant cost savings and benefits from its AI investments.

Hyatt uses AI to improve sales and customer engagement

Hyatt has redesigned its website, Hyatt.com, to use generative AI, allowing travelers to search based on intent rather than just location and dates. This AI-driven approach helps Hyatt capture customer interest earlier in the booking process, leading to better conversion rates and personalized offers. The company is also using AI internally to boost productivity and enhance its World of Hyatt loyalty program. This strategy aims to improve sales and customer engagement in the competitive travel industry.

Databricks buys startups for new AI security product

Databricks has acquired two startups, Antimatter and SiftD.ai, to strengthen its new AI security product. Antimatter, founded by security researcher Andrew Krioukov, was acquired last year, while SiftD.ai was acquired recently. These acquisitions are intended to bolster Databricks' capabilities in the AI security space. The company, known for its cloud data analytics platform, aims to stay ahead of market needs by integrating these new technologies and talent.

Louisiana plans AI for health department services

Louisiana's health department is considering using artificial intelligence to improve services and reduce costs. The department is looking into implementing AI in its call centers, which currently cost over $40 million annually, potentially saving up to 25%. AI chatbots are also being considered to help residents navigate the department's website, which offers information on various services. The state plans to proceed cautiously, ensuring patient privacy and consent are protected when using AI with sensitive health information.

Nevada unemployment system needs policy reform not just AI

This opinion piece argues that Nevada's unemployment system needs policy reform rather than just AI implementation to improve efficiency. While AI can speed up processes, the article suggests that the complex and outdated policies governing unemployment claims are the main issue. The author highlights that even with modern software, the system struggles due to convoluted regulations. True improvement requires updating these underlying policies and regulations, not just adding new technology.

SecurityScorecard uses AI for vendor risk management

SecurityScorecard has launched TITAN AI to automate third-party risk management (TPRM). This AI-driven platform aims to reduce manual work by up to 95%, allowing teams to manage vendor risk more efficiently and scale their programs without increasing staff. TITAN AI helps identify and address risks faster, potentially reducing supply-chain breaches by up to 75%. The system provides continuous visibility, intelligent automation for assessments, and threat-informed remediation, fostering collaboration between organizations and vendors.

Sources

• Cisco builds security framework for safe enterprise adoption of AI agents
• Cisco Reimagines Security for the Agentic Workforce
• AccuKnox Launches AI-Security 2.0 at RSAC 2026, San Francisco, USA
• Check Point unveils AI Defense Plane to govern and secure enterprise AI systems
• 20 Coolest AI And Security Products At RSAC 2026
• Arm Holdings, in Break From Past, Will Sell Its Own Computer Chips
• Arm launches own AI chip in high-stakes strategy shift
• Binkhorst Lecture preview
• Beehiiv to allow creators to manage accounts through AI platforms
• Norway wealth fund moves towards some AI-driven decisions but with humans in control
• Hyatt’s AI Captures Customers Sooner, Improving Sales and Productivity
• Databricks bought two startups to underpin its new AI security product
• Louisiana's health secretary wants AI on the phones and the website
• OPINION: You're mad at policy, not AI
• SecurityScorecard automates third-party risk management with TITAN AI