xAI & Grok

At least nine engineers including two co-founders departed xAI in February 2026, raising questions about stability. Notable exits include Jimmy Ba (co-founder/safety lead) and Tony Wu (co-founder/reasoning lead) amid controversy over safety protocols and recursive self-improvement systems.

In a February 2026 all-hands meeting, xAI positioned its AI development roadmap around interplanetary exploration and space systems support. The company is framing its agent development beyond terrestrial applications, differentiating from competitors focused on earthbound AI.

Databricks research shows 327% growth in multi-agent workflows as enterprises move beyond single chatbots to coordinated AI systems. 59% of organizations now use three or more LLMs in production, with Supervisor Agents being the most popular for directing other agents.

The 2026 State of AI Agents Report shows 57% of organizations deploy multi-step agent workflows and 81% plan expansion into complex agent use cases. However, 46% cite integration with existing systems as their primary challenge, not model capability.

OpenAI Agents

OpenAI announced the launch of OpenAI Frontier, an end-to-end platform designed for enterprises to build and manage AI agents, with an open platform that can manage agents built outside of OpenAI too. OpenAI designed Frontier to work like companies manage human employees, offering an onboarding process for agents and feedback loops to help them improve over time.

Salesforce plans to strengthen its investment in Agentforce Commerce by acquiring Cimulate, a company focused on artificial intelligence-powered product discovery and agentic commerce. Salesforce executive Nitin Mangtani stated that "the future of commerce is agentic, moving beyond simple transactions to intuitive, conversational discovery."

Snowflake and OpenAI announced a multi-year $200 million partnership to make OpenAI models natively available across Snowflake's enterprise data platform. The integration enables organizations to build agents that reason over governed data and deploy multimodal analysis, positioning AI as embedded enterprise infrastructure rather than standalone experimentation.

OpenAI's new Frontier platform helps enterprises build, deploy and manage AI agents, providing these "AI coworkers" with shared context, onboarding, hands-on learning, and clear permissions and boundaries. The company also introduced OpenAI Forward Deployed Engineers (FDEs) who work alongside enterprises' teams to develop best practices and provide direct connection to OpenAI Research.

Anthropic & Claude

Claude Code now features agent teams that can work on tasks together, allowing multiple agents to collaborate on complex projects. Opus 4.6 outperforms GPT-5.2 by 144 Elo points on economically valuable knowledge work tasks.

The most notable addition to Opus 4.6 is agent teams that can split larger tasks into segmented jobs. Instead of one agent working sequentially, multiple agents can own their piece and coordinate directly with others.

Claude Opus 4.6 is now available in Microsoft Foundry, allowing developers to delegate complex tasks end-to-end and trust AI to execute independently in production. The model handles large codebases well and allows senior engineers to delegate work that previously took days.

Xcode 26.3 introduces support for agentic coding, allowing developers to harness agents like Claude Agent and Codex directly in Xcode. Xcode makes its capabilities available through the Model Context Protocol, an open standard for compatible agents.

Microsoft Agent Framework now integrates with the Claude Agent SDK, enabling developers to build AI agents powered by Claude's full agentic capabilities. The integration supports multi-agent workflows with sequential, concurrent, handoff, and group chat patterns.

Anthropic debuted Claude Opus 4.6, an advanced AI model capable of conducting sophisticated professional tasks and coordinating whole teams of AI agents. The agent teams feature allows users to deploy multiple AI agents simultaneously that work in parallel and communicate to coordinate efforts.

Microsoft Copilot

Microsoft outlines six essential capabilities organizations need to successfully scale AI agent adoption in 2026, including governance, security, and operational frameworks. The post details how Copilot Studio enables multi-agent systems and collaborative workflows between business and IT teams.

Microsoft identifies critical security risks in Copilot Studio agent deployments, emphasizing the need for proper authentication, access controls, and governance as agents become integral to business workflows. The blog provides specific recommendations for securing AI agent implementations.

Microsoft officially launched OneDrive AI Agents as generally available on February 3, 2026, allowing users with Copilot licenses to create specialized AI assistants grounded in specific document sets. These agents can analyze up to 20 files and provide context-aware responses for project management and knowledge work.

Microsoft is transforming Copilot from a chat assistant into an operational AI layer with Agent Mode functionality across Word, Excel, and PowerPoint. The update emphasizes multi-step workflows, better grounding through Copilot Notebook, and enhanced governance controls for enterprise deployment.

PwC and Microsoft formed a strategic partnership to accelerate enterprise AI agent adoption across industries, leveraging PwC's Agents Factory framework and Microsoft's Copilot platform. The collaboration focuses on creating autonomous agents embedded within enterprise ecosystems for Fortune 500 clients.

Microsoft expanded partner benefit packages in February 2026 to include additional Copilot licenses, enhanced security features, and AI marketing resources. The update introduces Microsoft Agent 365 as a central control plane for managing AI agents and emphasizes the $300 billion marketplace opportunity by 2030.

Google & Gemini

Google began testing new ad formats on its search AI mode and integrating AI agents into its shopping experience with a new method for standardizing payments and digital identity. The company is working with Walmart, Target and Shopify on a protocol for direct checkout in Google's AI products.

Google outlined its direction for AI-powered digital advertising and commerce, emphasizing deeper AI integration in Search and YouTube creator collaborations. The Universal Commerce Protocol (UCP) now connects businesses to AI agents with secure payments and identity verification rolling out for select US retailers.

Google Cloud released its 2026 AI Agent Trends Report forecasting that AI agents will fundamentally reshape business. The report highlights cross-platform AI agents using the Agent2Agent (A2A) protocol being built by Salesforce and Google Cloud, establishing an interoperable foundation for agentic enterprises.

Google introduced the Universal Commerce Protocol (UCP) for agentic commerce, compatible with existing protocols like Agent2Agent and Agent Payments Protocol. The protocol enables native checkout with buy buttons directly on Google surfaces including AI Mode in Search and Gemini.

Google Workspace Studio launched as a core service for designing and sharing AI agents. The update also introduced Agentic Vision that turns Gemini 3 Flash from a passive observer into an active investigator, available via the Gemini API in Google AI Studio and Vertex AI.

Enterprise ANO Adoption

OpenAI Frontier is an end-to-end platform for enterprises to build and manage AI agents. It's an open platform supporting agents built outside of OpenAI, with Gartner calling agent management platforms the "most valuable real estate in AI."

Anthropic released Claude Opus 4.6 with agent teams functionality, while OpenAI launched GPT-5.3-Codex and the Frontier platform. The focus is shifting from raw model intelligence to infrastructure and tooling for enterprise AI deployment.

Research shows 59% of organizations now use three or more LLMs in production, up from 36% earlier. Multi-agent systems are becoming critical for enterprise AI, with 80% of databases now built by AI agents and 97% of testing done automatically.

ANO Security & Governance

OpenAI announced Frontier on February 5, 2026, an end-to-end platform for enterprises to build and manage AI agents with security controls. The open platform supports agents from multiple vendors and includes identity management, permissions, and governance features for enterprise deployment.

IBM published a security guide on February 10, 2026, revealing 79% of organizations are already deploying AI agents with 88% planning budget increases. The guide emphasizes treating AI agents like new employees requiring identity management, access controls, and security oversight as they gain autonomous capabilities.

A survey of 900+ executives reveals only 21.9% of teams treat AI agents as independent entities with proper identity controls, while 88% of organizations reported confirmed or suspected AI agent security incidents in the past year. Only 14.4% have full security approval for their agent deployments.

The viral open-source AI agent OpenClaw has security experts concerned due to its autonomous capabilities and access to users' personal data including emails and hard drives. Experts warn non-technical users should avoid the software due to potential security risks from uncontrolled agent actions.

ANO Academic Research

Databricks research shows 327% growth in multi-agent workflows on their platform. New specialized databases like 'lakebase' are being designed specifically for AI agents, with 80% of databases now built by AI agents and 97% of testing automated.

NVIDIA launched Nemotron 3 family of open models designed for agentic AI development. Early adopters include Accenture, CrowdStrike, Oracle, Palantir, ServiceNow, and Zoom for multi-agent system deployment across industries.

Training platform launches specialized program for multi-agent systems and autonomous AI agents. Program covers autonomous agents, multi-agent system design, RAG, and LLM orchestration as engineering teams increasingly manage agent-driven workflows.

IBM's Kate Blair predicts 2026 as the year multi-agent systems move into production. IBM contributed BeeAI and Agent Stack to the Linux Foundation's new Agentic AI Foundation, focusing on protocol maturity and convergence.

Multi-year $200 million partnership to deploy AI agents across global enterprises. Partnership brings OpenAI models to Snowflake Cortex AI for building enterprise AI agents with secure governance and data control.

Completed $250 million acquisition of Ask Sage, a generative AI platform for government and regulated industries. Platform supports agentic systems with FedRAMP High authorization and zero-trust security for defense and national security sectors.

ANO Regulation & Policy

OpenAI launched Frontier, an enterprise platform to build, deploy, and manage AI agents that can run other software, such as Salesforce and Workday. Frontier appears to be OpenAI's bid to become the "operating system of the enterprise," offering a unified platform for building agents that can navigate apps, execute workflows, and make decisions, connecting databases, business systems and other internal applications.

OpenAI launched Frontier, a new platform designed for enterprises to build and deploy agents while treating them like human employees. Global research firm Gartner called agent management platforms both the "most valuable real estate in AI" and necessary infrastructure for enterprises to adopt AI, making this launch strategically timed for early 2026.

California's Transparency in Frontier Artificial Intelligence Act and Texas's Responsible Artificial Intelligence Governance Act went into effect on January 1, 2026, but President Trump's December 11, 2025 executive order casts doubt on their enforceability. The Executive Order directs the Attorney General to establish an AI litigation task force to challenge state AI laws and calls for evaluation of burdensome state AI laws by March 11, 2026.

February 2026 marks a critical month in the battle over who controls AI regulation in America, with the question playing out in courtrooms, state legislatures, and federal agencies right now. 2026 is predicted to be "the year of agents as software expands from making humans more productive to automating work itself, delivering on the human-labor displacement value proposition in some areas."