Agentic AI this week (June 12 - June 19, 2026)
The week of June 12 to 19, 2026 delivered a seismic regulatory shock, a flurry of cloud platform announcements, and fresh evidence that the AI gateway has become the decisive control surface in the agentic security market. Anthropic suspended access to Claude Fable 5 and Claude Mythos 5 globally on June 12 after the US government issued an export-control directive citing a jailbreak vulnerability related to cybersecurity tasks, affecting every cloud integration from AWS Bedrock to Google Cloud to Microsoft Foundry simultaneously. AWS held its Summit in New York on June 17, announcing Web Search on Bedrock AgentCore with MCP-native connectors, Kiro Pro Max for high-volume agentic development, and EC2 G7 instances powered by NVIDIA RTX PRO 4500 Blackwell GPUs delivering 4.6x inference performance gains. KPMG and Microsoft announced the rollout of Agent 365 to all 276,000 professionals across 138 countries, making it the largest governed AI agent deployment in enterprise history. Check Point revealed its integration of the AI Defense Plane with Google Cloud's Agent Gateway and Agent Registry, targeting late June availability with prompt-injection detection and runtime behavioral guardrails. Fortinet expanded the FortiGate G Series with the 3500G and 400G appliances running FortiOS 8.0, which adds native MCP and agent-to-agent traffic inspection alongside shadow AI detection dashboards. The MCP ecosystem continued its growth trajectory with SDK downloads reaching 97 million monthly and more than 10,000 public servers indexed, while the release candidate for the July 28 stateless specification entered its ten-week validation window. OpenAI expanded Codex into an enterprise work platform with six role-specific plugins connecting 62 business applications and 5 million weekly users. The EU's Digital Omnibus agreement pushed high-risk AI system obligations from August 2026 to December 2027, while SpaceX SPCX stock swung from an all-time high of $225.64 on June 16 to approximately $185 by week's end as IPO euphoria met reality.
Anthropic Suspends Fable 5 and Mythos 5 After US Export-Control Directive
Anthropic received a US government export-control directive at 5:21 PM ET on June 12, 2026, ordering the suspension of all access to Claude Fable 5 and Claude Mythos 5 by any foreign national, whether inside or outside the United States, including Anthropic's own foreign-national employees. The directive cited national security authorities and centered on a potential method of bypassing safeguards intended to limit Fable 5's use for cybersecurity-related tasks, including identifying software vulnerabilities. Anthropic responded with a universal shutdown across all platforms, simultaneously disabling the models on AWS Bedrock, Google Cloud, Microsoft Foundry, Snowflake, Box, and the direct Claude APIs. The company stated it could not technically filter users by nationality in real time across dozens of global cloud platforms, making a blanket suspension the only legally viable response. Anthropic confirmed that Claude Opus 4.8 and other models remain fully available.
Why it matters. The Fable 5 suspension is the first time a US government export-control directive has forced a live, global model shutdown across every major cloud platform simultaneously. This event reframes the AI gateway discussion from a pure security and governance narrative into a compliance and sovereign-access question. Every enterprise that had built agentic workflows on Fable 5, which Anthropic positioned as a Mythos-class model with 1 million token context and 128K output tokens, discovered on June 12 that their most capable model could be switched off with hours of notice. For AI gateway operators like Palo Alto Networks (through its newly integrated Portkey acquisition), Cloudflare AI Gateway, ServiceNow AI Control Tower, and IBM DataPower Interact Gateway, the Fable 5 suspension validates the architectural premise that enterprises need a centralized control plane capable of instant model failover. Organizations routing all agent traffic through a gateway could, in principle, redirect Fable 5 workloads to Claude Opus 4.8 or a competing model without touching individual application configurations.
The competitive implications extend to how security vendors position model governance. CrowdStrike's Charlotte AI AgentWorks ecosystem and Fortinet's FortiOS 8.0 MCP traffic inspection both provide visibility into which models agents are calling, but neither offers the routing-layer failover that a dedicated AI gateway provides. Check Point's AI Defense Plane, which is integrating with Google Cloud's Agent Gateway for late June availability, includes agent registry and pre-deployment policy enforcement that could theoretically block agents from calling suspended models before runtime. The Fable 5 episode will likely accelerate enterprise adoption of multi-model gateway architectures and strengthen the business case for every gateway vendor in the market.
AWS Summit New York: AgentCore Web Search, Kiro Pro Max, and G7 Instances
AWS held its Summit in New York at the Javits Center on June 17, 2026, with VP of Agentic AI Swami Sivasubramanian delivering the keynote. The headline announcement was Web Search on Amazon Bedrock AgentCore, a fully managed tool that enables agents to ground responses in current, cited web knowledge using a built-in MCP connector. Pricing details were published on June 18, and the tool requires zero data egress from the customer's secured AWS environment. Kiro Pro Max entered the IDE lineup, offering higher usage limits, access to the latest frontier models, and expanded agentic capabilities for professional developers, alongside a native iOS app in gated preview that lets developers monitor and steer Kiro sessions from their phone. AWS also launched EC2 G7 instances as the first major cloud provider to support NVIDIA RTX PRO 4500 Blackwell Server Edition GPUs, delivering up to 4.6x AI inference performance and 2.1x graphics performance over G6. Additional announcements included Amazon Bedrock Managed Knowledge Base for enterprise RAG pipelines, AWS Continuum for AI-native code vulnerability remediation, and SOC 1, 2, and 3 compliance for AgentCore.
Why it matters. AWS is methodically assembling an agentic stack that competes directly with Microsoft's Azure AI Foundry and Google's Gemini Enterprise Agent Platform. The Web Search tool's use of MCP as its connector protocol signals that AWS views MCP as the standard interface for agent-to-tool communication, reinforcing the protocol's momentum ahead of the July 28 specification release. By making AgentCore the gateway through which web search queries flow, AWS positions itself to capture telemetry on how agents access external information, a control surface that parallels what Palo Alto Networks achieves through Portkey's AI gateway integration into Prisma AIRS. The SOC compliance milestone also matters: enterprises evaluating ServiceNow AI Control Tower or IBM DataPower for agent governance now have a direct comparison point in AWS's managed agent runtime.
The Kiro Pro Max launch intensifies the AI-native IDE competition. OpenAI Codex, which expanded to 5 million weekly users with its June 2 enterprise plugins update, now faces a dedicated AWS competitor that integrates natively with Bedrock's model garden and AgentCore's identity and memory services. Google's Agent Development Kit (ADK), available in Python, Go, Java, and TypeScript as part of the Gemini Enterprise Agent Platform, competes on the framework layer. For security teams, the proliferation of agentic IDEs means more autonomous code generation flowing through CI/CD pipelines, increasing the importance of the gateway-level inspection that Fortinet's FortiOS 8.0 and CrowdStrike's Falcon AIDR are designed to provide.
KPMG Deploys Microsoft Agent 365 to 276,000 Staff Across 138 Countries
KPMG and Microsoft announced on June 9, 2026 that KPMG would roll out Microsoft 365 Copilot and Agent 365 to all 276,000 professionals across 138 countries and territories. KPMG will use Agent 365 to enhance its Trusted AI framework and help clients manage, monitor, and secure AI agents across their organizations. Agent 365 functions as an enterprise control plane that registers, maps, secures, and measures every agent in an organization, and it reached general availability on May 1, 2026. Microsoft also confirmed that Agent Mode for Copilot, which enables persistent AI agents inside Word, Excel, Teams, and Outlook, will roll out to Microsoft 365 subscribers in late June. Starting July 1, 2026, Microsoft 365 Business Standard with Copilot and Microsoft 365 Business Premium with Copilot become permanent SKUs.
Why it matters. The KPMG deployment establishes Agent 365 as the first enterprise agent governance platform to achieve six-figure user scale in a single organization. This is a direct competitive challenge to ServiceNow's AI Control Tower, which positions itself as the "agent of agents" and offers its governance capabilities free with every Now Assist and AI Native SKU. ServiceNow's approach embeds agent governance into the ITSM and workflow layer, while Microsoft's approach embeds it into the productivity suite. Both architectures ultimately converge on the same control-plane function: registering, authenticating, and governing AI agents. For enterprises already running Microsoft 365, Agent 365 is a natural extension, while organizations with deep ServiceNow investments in IT operations may prefer AI Control Tower. The question is whether either platform can serve as the universal governance layer or whether enterprises will need both, creating integration complexity that benefits gateway vendors like Cloudflare and IBM DataPower that can sit between multiple agent orchestration platforms.
The KPMG deal also signals that professional services firms, which collectively employ millions of knowledge workers, are moving from AI experimentation to governed production deployment. Gartner forecasts that 40% of enterprise applications will embed task-specific AI agents by end of 2026, up from under 5% in 2025, and the KPMG scale deployment helps validate that projection. For security vendors, the implication is clear: 276,000 users generating agent traffic across 138 countries creates a massive governance and compliance surface. Palo Alto Networks' Prisma AIRS, Check Point's AI Defense Plane, and CrowdStrike's Charlotte AI AgentWorks all need to demonstrate they can secure agent workloads at this scale, across this geographic breadth, and across the regulatory regimes of 138 different jurisdictions.
Check Point Integrates AI Defense Plane with Google Cloud Agent Gateway
Check Point announced its integration with Google Cloud's Gemini Enterprise Agent Platform as a launch partner, combining Check Point's AI Defense Plane with Google Cloud's Agent Gateway and Agent Registry. The integration will be available in late June 2026 and delivers three security layers for enterprise AI agents: agent inventory through the Agent Registry, pre-deployment policy enforcement through governance rules, and real-time runtime guardrails including prompt-injection detection. Check Point's architecture provides contextual intelligence and behavioral protection while Google Cloud's Enterprise Agent Platform serves as the control plane for identity and connectivity. Organizations can register for early access through Check Point's AI security portal.
Why it matters. The Google Cloud partnership gives Check Point a hyperscaler distribution channel for its AI Defense Plane that none of its direct competitors currently match in the Google ecosystem. Palo Alto Networks has anchored its agentic security strategy on the Portkey AI gateway acquisition, which provides a vendor-neutral control plane, but Portkey does not yet have a native Google Cloud Agent Gateway integration. Fortinet's FortiOS 8.0 provides MCP and A2A traffic inspection at the firewall layer but does not integrate with any hyperscaler's agent registry or governance platform. CrowdStrike's Charlotte AI AgentWorks operates within the Falcon platform ecosystem, focusing on agent identity through SGNL and browser security through Seraphic, but it does not position as a hyperscaler agent governance partner. Check Point's move to integrate directly with Google's agent orchestration layer means it can enforce security policies before agents are deployed, not just at runtime, a capability that addresses the Gartner finding that 89% of AI agents never reach production partly due to governance gaps.
The broader strategic pattern is the formation of hyperscaler-security vendor alliances for agent governance. Microsoft has Agent 365 as its own agent control plane, AWS has Bedrock AgentCore, and now Google Cloud is partnering with Check Point for agent security. This alignment creates distinct ecosystem channels: enterprises on Google Cloud may gravitate toward Check Point's AI Defense Plane, enterprises on Azure toward Microsoft's native Agent 365 governance, and enterprises on AWS toward AgentCore's built-in identity and policy features. For ServiceNow and IBM, which position as platform-agnostic governance layers, this hyperscaler-vendor consolidation could either create partnership opportunities or competitive pressure depending on whether enterprises want multi-cloud agent governance or accept per-cloud solutions.
Fortinet Expands FortiGate G Series with Native MCP and A2A Traffic Inspection
Fortinet announced the FortiGate 3500G and FortiGate 400G, new additions to its G Series portfolio running FortiOS 8.0 and powered by Fortinet's NP7 and SP5 custom ASICs. The FortiGate 3500G targets high-density data center environments with 595 Gbps firewall throughput, which Fortinet claims is 3.4x the average of comparable competitors, and supports 179 million concurrent sessions. FortiOS 8.0 introduces native detection and classification of MCP and agent-to-agent traffic, providing real-time visibility into AI data flows without requiring additional tooling. A dedicated FortiView dashboard for the AI attack surface delivers real-time visibility into how AI applications are used across the organization, automatically distinguishing between authorized and unauthorized shadow AI tools. FortiOS 8.0 also includes MCP observability with telemetry into how and why agents communicate, enabling organizations to monitor agent behavior and enforce policy.
Why it matters. Fortinet is taking the only infrastructure-first approach to agentic security among the major vendors, embedding AI traffic inspection directly into the network firewall rather than building or acquiring a separate AI gateway platform. This positions FortiOS 8.0 as the industry's first next-generation firewall with native MCP protocol awareness, a capability that Palo Alto Networks' PA-Series firewalls do not yet offer despite Palo Alto's extensive AI gateway investments through Portkey. The distinction is architectural: Palo Alto routes AI traffic through a dedicated gateway (Portkey integrated into Prisma AIRS), while Fortinet inspects AI traffic inline at the firewall. For enterprises that already deploy FortiGate appliances at data center boundaries, this means MCP and A2A visibility comes as a firmware upgrade rather than a new product deployment, reducing both cost and operational complexity.
The 595 Gbps throughput of the FortiGate 3500G also addresses a practical concern in agentic AI data centers: the volume of model API calls, tool invocations, and agent-to-agent messages can generate substantial east-west traffic that must be inspected at wire speed. Check Point's AI Defense Plane operates at the cloud platform layer through its Google Cloud integration, and CrowdStrike's Falcon AIDR operates at the endpoint and SaaS layer, but neither provides the raw throughput needed for inline data center inspection of AI traffic at scale. IBM's DataPower Interact Gateway can inspect AI interactions at the application layer, but DataPower's deployment model is a software gateway rather than a purpose-built ASIC-accelerated appliance. Fortinet's organic, hardware-accelerated approach to MCP inspection avoids the acquisition costs that have driven Palo Alto past $1.1 billion in 2026 AI-related acquisitions, though it limits Fortinet to the network perimeter rather than the application-layer control point where most agent governance decisions are made.
MCP Ecosystem Approaches July 28 Specification with 97 Million SDK Downloads
The MCP (Model Context Protocol) ecosystem continued its rapid expansion through June 2026, with TypeScript and Python SDKs reaching 97 million monthly downloads, up from approximately 2 million at launch, representing a growth rate of 4,750% in 16 months. The official MCP Registry API counted 9,652 server records as of May 24, with Anthropic's December 2025 ecosystem update citing more than 10,000 active public servers. PulseMCP tracking shows roughly 1,000 new indexed servers per month in Q1-Q2 2026, and 80% of top servers now offer remote deployment options. The release candidate for the July 28 specification delivers a stateless protocol core, the Extensions framework, Tasks for long-running work, MCP Apps for server-rendered UIs, authorization hardening aligned with OAuth and OpenID Connect, and a formal deprecation policy. Stacklok's 2026 software report shows 41% of surveyed organizations in limited or broad production with MCP servers. The A2A (Agent-to-Agent) protocol, hosted by the Linux Foundation, surpassed 150 organizations at its one-year mark, with native integration across Google Cloud, Microsoft Azure AI Foundry, and AWS Bedrock AgentCore.
Why it matters. The convergence of MCP's 97 million monthly SDK downloads, 10,000 public servers, and the approaching July 28 stateless specification creates an inflection point for the AI gateway market. The stateless core is the most significant architectural change since MCP's launch: it eliminates the need for persistent connections between agents and tools, meaning MCP interactions can now scale on ordinary HTTP infrastructure. This directly benefits gateway vendors like Cloudflare AI Gateway, which already processes traffic from 70+ models through its edge network, and Kong AI Gateway, which extends its API management platform with LLM-specific plugins. For security vendors, the protocol's maturation increases both opportunity and urgency. Fortinet's FortiOS 8.0 MCP traffic inspection, Palo Alto's Portkey-powered AI gateway, and ServiceNow's AI Control Tower MCP Server (generally available with every Now Assist SKU) all depend on a stable protocol specification to build reliable inspection and governance capabilities.
The A2A protocol's milestone of 150 supporting organizations, including Google, Microsoft, AWS, Salesforce, SAP, ServiceNow, and IBM, establishes it as the complementary standard to MCP for inter-agent communication. While MCP governs how agents connect to tools and data sources, A2A governs how agents discover and delegate tasks to each other. The introduction of the Agent Payments Protocol (AP2) with more than 60 supporting organizations in payments and financial services signals that the agent protocol stack is expanding beyond computation into transactions. For gateway vendors, this means the control plane must eventually govern not just model calls and tool invocations but also agent-to-agent negotiations and financial transactions, a scope expansion that favors comprehensive platforms like IBM DataPower Interact Gateway and Palo Alto's Prisma AIRS over point solutions.
OpenAI Codex Expands to Enterprise Work Platform with 5 Million Weekly Users
OpenAI announced a major update to Codex on June 2, 2026, introducing Sites (hosted web applications for internal dashboards), Annotations (in-place editing), and six role-specific plugins connecting to 62 business applications with 110 pre-built automated skills. The plugins target data analytics, creative production, sales, product design, equity investing, and investment banking. Non-developers now make up 20% of Codex's 5 million weekly users and are adopting 3x faster than engineers. Enterprise and Edu administrators gained new credit management tools with workspace, group, and user-level monthly limits. Separately, OpenAI retired GPT-5.2 models from ChatGPT on June 12, with conversations automatically migrating to GPT-5.5, and confirmed GPT-4.5 retirement on June 27 following a 30-day sunset. GPT-5.5 Instant personalization improvements rolled out to ChatGPT Go and Free tiers on June 9, and GPT-Realtime-2, a new voice model with configurable reasoning for speech-to-speech agents, entered general availability.
Why it matters. OpenAI's transformation of Codex from a code-generation tool into an enterprise work platform with 5 million weekly users places it in direct competition with Microsoft's Copilot ecosystem and Google's Gemini Enterprise Agent Platform. The 20% non-developer user share is particularly significant: it means Codex is becoming a horizontal agent platform rather than a developer tool, expanding the attack surface that security teams must govern. Every one of those 62 business application integrations represents a potential data exfiltration path, a credential exposure surface, and an authorization boundary that must be secured. Palo Alto Networks' Prisma AIRS with Portkey is designed to sit in front of exactly this type of multi-application agent traffic, providing unified visibility across all model calls and tool invocations.
The model deprecation timeline also carries security implications. The retirement of GPT-5.2 on June 12 and GPT-4.5 on June 27 forces enterprises to migrate agentic workflows to GPT-5.5, a process that requires testing prompt behaviors, verifying tool-call compatibility, and updating rate-limit configurations. Combined with Anthropic's forced Fable 5 suspension on the same day, June 12 became a dual model-disruption event that exposed organizations without gateway-level model routing to simultaneous migration pressures across both providers. Cloudflare AI Gateway's spend-limit controls and unified billing, LiteLLM's open-source proxy wrapping 100+ providers behind an OpenAI-compatible endpoint, and ServiceNow's AI Gateway within AI Control Tower all provide the routing abstraction that insulates enterprises from exactly this type of provider-driven disruption.
EU Digital Omnibus Extends High-Risk AI Deadlines While Scientific Panel Takes Shape
The European Commission appointed a Scientific Panel and an Advisory Forum on June 1, 2026 to support enforcement of the AI Act. The Scientific Panel brings together 60 independent experts focused on general-purpose AI models, systemic risks, model classification, evaluation methodologies, and cross-border market surveillance. On May 7, 2026, negotiators from the Council, Parliament, and Commission reached a provisional agreement on the Digital Omnibus on AI, the first set of amendments to the EU AI Act. The headline change postpones high-risk AI system obligations (Annex III, use-based) from August 2, 2026 to December 2, 2027, providing an additional 16 months of compliance runway. The final Code of Practice on marking and labeling AI-generated content is scheduled for publication in June 2026, establishing standardized methods for generative AI companies to mark and label their outputs.
Why it matters. The 16-month extension for high-risk AI system obligations relieves immediate compliance pressure on enterprises deploying agentic AI in healthcare, critical infrastructure, and law enforcement, the categories most affected by Annex III requirements. However, the appointment of a 60-member Scientific Panel with a mandate covering GPAI model classification and systemic risk evaluation signals that regulatory scrutiny of foundation models is intensifying even as compliance deadlines extend. For AI gateway vendors, the Code of Practice on AI-generated content labeling creates a new technical requirement: gateways will need to inspect and potentially tag AI-generated outputs with provenance metadata, a capability that adds complexity to the control plane. IBM DataPower Interact Gateway, which positions as a governance gateway for AI interactions, and Palo Alto's Prisma AIRS, which provides full AI lifecycle security, are both architected to apply policy transformations to AI traffic that could accommodate content-labeling requirements.
The regulatory timeline extension also affects competitive dynamics among security vendors. Organizations that had been accelerating AI governance investments to meet the August 2026 deadline may now slow procurement, potentially delaying enterprise adoption of platforms like Check Point's AI Defense Plane, CrowdStrike's Falcon AIDR, and ServiceNow's AI Control Tower. Conversely, the extension gives vendors more time to mature their products before the compliance wave hits in December 2027, which benefits vendors pursuing organic development strategies like Fortinet over those that have already made large acquisitions, such as Palo Alto Networks' $1.1 billion in 2026 AI-related deals. The Anthropic Fable 5 suspension demonstrates that regulatory and government actions can reshape the AI landscape overnight, regardless of published compliance timelines, reinforcing the need for gateway-level agility that no regulatory schedule can fully anticipate.
SpaceX SPCX Volatility and xAI Model Developments
SpaceX shares traded under the SPCX ticker on the Nasdaq continued their volatile post-IPO trajectory, surging to an all-time high of $225.64 on June 16 before pulling back sharply to close the week around $185 on June 19. The stock debuted at $135 per share on June 12, raised a record $75 billion, and closed its first day at $160.95 with a 19% gain. The extreme price swings, from $149.34 to $225.64 within a single week, reflect a restricted 4% free float and momentum-driven trading rather than fundamental catalysts. Analysts maintain an average 12-month price target of $187.80, with a wide range from $62 to $310 across 7 analyst ratings. On the AI model front, xAI confirmed that Grok V9-Medium, a 1.5 trillion parameter model three times larger than the current production version, has completed training and is expected for release in mid-June. The larger Grok 5, targeting 6 trillion parameters in a Mixture-of-Experts architecture, continues training on xAI's Colossus 2 supercluster equipped with approximately 550,000 NVIDIA GB200 and GB300 GPUs, though prediction markets give it only a 33% probability of shipping by June 30.
Why it matters. The SpaceX-xAI combination, now sharing a corporate treasury bolstered by $75 billion in IPO capital, represents a unique convergence of compute infrastructure, satellite connectivity, and foundation model development. The Colossus 2 cluster with 550,000 GPUs is the largest single training installation in the world, and the models it produces will eventually flow through the same AI gateways and security control planes that every enterprise uses. For security vendors, xAI's models are already appearing in enterprise environments: Cloudflare AI Gateway recently expanded its partnership with xAI for full Grok integration across its 70+ model gateway, and Palo Alto Networks' Portkey integration into Prisma AIRS is designed to provide model-agnostic governance regardless of whether enterprises run Claude, GPT, Gemini, or Grok workloads.
The SPCX volatility also has implications for the broader AI IPO pipeline. Anthropic and OpenAI both have confidential S-1 filings with the SEC, targeting valuations of $965 billion and $730 billion to $1 trillion respectively. SpaceX's post-IPO price action, which saw a 67% swing from trough to peak in a single week, may cause underwriters to reconsider pricing strategies for subsequent AI company listings. For the security vendor market, this capital formation cycle means that the companies building the agentic AI stack will have unprecedented resources to invest in platform expansion, creating larger and more complex attack surfaces that CrowdStrike, Fortinet, Check Point, and their competitors must secure.
Numbers at a Glance
Anthropic suspended Claude Fable 5 and Mythos 5 across all platforms on June 12 following a US export-control directive received at 5:21 PM ET. SpaceX SPCX stock hit an all-time high of $225.64 on June 16 before settling at approximately $185 by June 19, with a market cap fluctuating around $2.2 trillion. KPMG deployed Microsoft Agent 365 to 276,000 professionals across 138 countries. AWS launched EC2 G7 instances with 4.6x inference performance on NVIDIA RTX PRO 4500 GPUs. Fortinet's FortiGate 3500G delivers 595 Gbps firewall throughput with 179 million concurrent sessions. MCP SDKs reached 97 million monthly downloads with 10,000+ public servers, while the A2A protocol passed 150 supporting organizations. OpenAI Codex hit 5 million weekly users with 20% non-developers. The EU Digital Omnibus extended high-risk AI obligations by 16 months to December 2027. Gartner forecasts 40% of enterprise applications will embed AI agents by end of 2026, up from under 5% in 2025.
References
- Greenberg Traurig: AI Company Anthropic Suspends Access to Claude Fable 5, Claude Mythos 5 Following US Export Control Directive
- AWS: Top Announcements of the AWS Summit in New York, 2026
- AWS: Announcing Web Search on Amazon Bedrock AgentCore
- Microsoft: KPMG and Microsoft Scale Trusted Enterprise AI Agents Globally Through Deployment of Agent 365 and Copilot
- Check Point: AI Defense Plane Integration with Google Cloud
- Fortinet: FortiGate G Series Expansion for AI-Era Security
- Fortinet: FortiOS 8.0 Introduction
- MCP Blog: 2026-07-28 Release Candidate
- Linux Foundation: A2A Protocol Surpasses 150 Organizations
- OpenAI: Codex for Almost Everything
- CNBC: SpaceX IPO Takeaways, SPCX Closes at $161
- EU Digital Strategy: AI Act Enforcement Gets Independent Expert Support
- Inside Privacy: EU AI Act Update, Timeline Relief and Targeted Simplification
- Palo Alto Networks: Completes Acquisition of Portkey
- Effloow: MCP Ecosystem in 2026, From Experiment to 97 Million Installs
- ServiceNow: AI Control Tower Expansion