microsoft, google and anthropic Updates

November 18, 2025, marked a significant day for artificial intelligence, with major announcements from Microsoft, Google, and Anthropic, alongside broader industry discussions on AI's impact and challenges. At Microsoft Ignite 2025, Microsoft showcased a strong focus on AI agents, integrating them deeply into customer experience and security. The company introduced a new Sales Development Agent in Frontier preview to autonomously engage leads and launched Sora 2, also in Frontier, allowing Microsoft 365 Copilot users to create AI-generated video clips. Power Apps gained an agent-powered workspace, enabling app creation through chat, and agents in Microsoft Teams can now connect with third-party tools like Jira and Asana. Microsoft also unveiled Edge for Business as the world's first secure enterprise AI browser, featuring Copilot Mode and Agent Mode for automated tasks on approved websites. Security for AI agents received substantial upgrades with Microsoft Agent 365, a control plane designed to observe, manage, and govern agents at scale, preventing 'agentic sprawl.' Windows 365 for Agents extends these security features to the cloud, and Microsoft Defender received AI-driven enhancements, including new Security Copilot agents now available to Microsoft 365 E5 subscribers at no additional cost. Microsoft is expanding its agent portfolio with over 40 new agents, including 12 Microsoft-built agents in preview and more than 30 partner-built agents in the Microsoft Security Store. Meanwhile, Google launched its next-generation Gemini 3 AI model, integrating it into its search engine and other online services. CEO Sundar Pichai called Gemini 3 'our most intelligent model,' aiming to transform Google Search into a 'thought partner' with an AI 'thinking' feature. This model will first roll out to Gemini Pro and Ultra subscribers in the United States, with Google emphasizing built-in guardrails to prevent misinformation. Alphabet's market value now stands around $3.4 trillion, reflecting its significant AI investment. Anthropic also made headlines, partnering with Coursera to launch two new AI Specializations on November 18, 2025. These courses, 'Building with the Claude API' and 'Real-World AI for Everyone,' aim to expand AI skills and promote responsible AI innovation, addressing a critical skills gap where only one-third of employees currently have AI training. Anthropic is also expanding its collaboration with Microsoft and Nvidia, integrating Claude models like Sonnet 4.5 into Microsoft Azure. Beyond these major announcements, the industry grapples with several challenges. A survey revealed that 73% of manufacturers find AI investment costs a heavy burden, particularly for Small and Medium-sized Enterprises (SMEs), who also face skilled worker shortages. Dell Technologies Vice President Nicole Jefferson highlighted the need for investment in modern data centers, a modern energy grid, and agile regulations to power the AI era. The rapid pace of GPU innovation, with Nvidia pushing for annual releases, poses a significant financial risk for AI hyperscalers due to quick obsolescence and potential depreciation crises. Investors are also scrutinizing AI company valuations more closely, with firms like Rothschild downgrading Microsoft to neutral after its deal with Anthropic, signaling a demand for clearer returns. Velantra AI discussed the crucial partnership between AI algorithms and human traders in financial markets, emphasizing human judgment and oversight for unexpected events. Senator Bernie Sanders voiced concerns about AI's potential to deepen wealth inequality and reshape warfare, calling for more public discussion and legislation. Finally, Jiga secured $12 million in Series A funding to address a bottleneck in AI development by speeding up hardware sourcing through its AI-native manufacturing platform, allowing engineers to order parts in hours instead of weeks.

Key Takeaways

• Microsoft unveiled extensive AI agent capabilities at Ignite 2025, including a Sales Development Agent, Sora 2 for video creation, and agent-powered Power Apps.
• Microsoft significantly enhanced AI agent security with Microsoft Agent 365, Windows 365 for Agents, and Edge for Business as the first secure enterprise AI browser.
• Security Copilot agents are now available to Microsoft 365 E5 subscribers at no extra cost, with over 40 new agents expanding the portfolio for security operations.
• Google launched its Gemini 3 AI model on November 18, 2025, integrating it into Google Search to act as a 'thought partner' with built-in guardrails.
• Anthropic partnered with Coursera to offer two new AI specializations, 'Building with the Claude API' and 'Real-World AI for Everyone,' to address the AI skills gap.
• Anthropic is expanding its collaboration with Microsoft and Nvidia, integrating Claude models like Sonnet 4.5 into Microsoft Azure.
• The rapid pace of GPU innovation, with Nvidia pushing for annual releases, creates significant financial risks and depreciation challenges for AI hyperscalers.
• 73% of manufacturers struggle with high AI investment costs, particularly SMEs, who also face skilled worker shortages and doubts about AI effectiveness.
• Investors are scrutinizing AI company valuations more closely, moving away from 'free passes' and demanding clearer returns, as seen with Rothschild's downgrade of Microsoft.
• Jiga secured $12 million in Series A funding to accelerate AI hardware production, using its AI-native platform to source parts in hours instead of weeks.

Microsoft Ignite 2025 Highlights AI Agents for Customer Experience

Microsoft Ignite 2025, held on November 18, 2025, showed that AI agents are now central to customer experience. Microsoft introduced a new Sales Development Agent in Frontier preview to autonomously research and engage leads, with security built into Agent 365. Sora 2, also in Frontier, now lets Microsoft 365 Copilot users create AI-generated video clips for personalized content like explainer videos. Power Apps gained an agent-powered workspace, allowing users to build apps by chatting with Copilot, speeding up workflow development. Additionally, agents in Microsoft Teams can now connect with third-party tools like Jira and Asana to improve communication and task management.

Windows Boosts Security for AI Agents and Cloud

At Microsoft Ignite on November 18, 2025, Windows announced new security and resiliency features for the AI era. These updates focus on trust and protection for AI agents, building on Windows 11's secure design. Key innovations include distinct agent accounts with limited privileges and a privacy-preserving design. Microsoft introduced Agent workspace, a secure environment for agents to perform tasks, and Windows 365 for Agents, which extends these security features to the cloud. Agent connectors, now in preview, use a Model Context Protocol proxy layer to manage agent interactions securely.

Microsoft Unveils New Autonomous Security for AI Agents

At Microsoft Ignite on November 18, 2025, Microsoft introduced new ambient and autonomous security measures for the AI era. The company launched Microsoft Agent 365, a control plane designed to observe, manage, and govern AI agents at scale, helping to prevent "agentic sprawl." This includes features like registries, access controls, and a unified dashboard for clear visualization of agent activities. Microsoft also launched the Foundry Control Plane for managing agent fleets and a Security Dashboard for AI to provide unified visibility into security risks. Furthermore, Microsoft Purview expanded its data security for Microsoft 365 Copilot, and new integrations were added to Microsoft Defender and Intune to secure platforms and cloud environments.

Microsoft Defender Boosts Security with AI and Cloud Tools

At Ignite 2025, Microsoft Defender received major AI-driven upgrades to enhance security across cloud and multicloud environments. New Security Copilot agents will automate security tasks, rolling out to Microsoft 365 E5 subscribers. Microsoft Defender now features expanded automatic attack disruption, which isolates compromised systems and uses Predictive Shielding to prevent future attacks. It also supports third-party attack disruption for services like AWS and Okta. For AI agents, Microsoft Defender offers unified posture management and threat protection through Microsoft Agent 365, providing a central view and detecting threats like prompt injections. Additionally, Microsoft Defender for Cloud gained new posture management for serverless resources and a unified security view across Azure, AWS, and Google Cloud.

Microsoft Strengthens Autonomous Security for AI Agents

On November 18, 2025, at Microsoft Ignite, Microsoft announced significant updates to enhance autonomous security for AI agents. The company introduced Microsoft Agent 365, a new control plane designed to observe, manage, and govern AI agents at scale, helping to prevent "agentic sprawl." This includes features like registries, access controls, and a unified dashboard for clear visualization of agent activities. Microsoft also launched the Foundry Control Plane for managing AI agent fleets and a Security Dashboard for AI to provide unified visibility into security risks. Furthermore, Microsoft Purview expanded its data security for Microsoft 365 Copilot, and new integrations were added to Microsoft Defender and Intune to secure platforms and cloud environments.

Edge for Business Launches First Secure AI Browser

On November 18, 2025, Microsoft unveiled Edge for Business as the world's first secure enterprise AI browser. This new browser combines the power of Microsoft 365 Copilot with company context from Microsoft Graph and the strong security of Edge for Business. It introduces Copilot Mode, a set of AI-powered features that can be enabled by IT teams to make browsing more intelligent and proactive. Within Copilot Mode, Agent Mode automates complex tasks on approved websites, with clear user control and visual cues. This innovation aims to boost productivity while maintaining enterprise-level security and compliance.

Microsoft 365 E5 Now Includes Security Copilot Agents

At Microsoft Ignite 2025, Microsoft announced that Security Copilot agents are now available to Microsoft 365 E5 subscribers at no additional cost. These AI agents aim to help security teams by automating tasks in Microsoft Defender for security operations, Microsoft Purview for data security, Microsoft Entra for identity management, and Microsoft Intune for endpoint management. Microsoft is also expanding its agent portfolio with over 40 new agents, including 12 Microsoft-built agents in preview and more than 30 partner-built agents available in the Microsoft Security Store. The platform also allows users to create custom agents and offers an interactive agent experience in public preview for focused security chats.

Microsoft Ignite 2025 Unveils New Copilot and AI Agent Features

At Microsoft Ignite 2025 on November 18, Microsoft introduced new Copilot and AI agent capabilities designed for "Frontier Firms." Agent Mode is expanding to Office apps, with Agent Mode in Word now generally available and in Excel and PowerPoint via the Frontier program. Microsoft 365 Copilot now features Work IQ, an intelligence layer that helps Copilot understand user data, preferences, and workflows, and can be used for custom agents. Copilot in Outlook gained voice control for inbox and calendar management, and Copilot Chat will soon include Agent Mode access. Additionally, Sora 2 for next-gen video creation and the secure Edge for Business AI browser were highlighted.

Coursera and Anthropic Launch New Claude AI Courses

On November 18, 2025, Coursera and Anthropic announced a partnership to offer two new AI Specializations. These courses aim to expand access to AI skills and promote responsible AI innovation for workers worldwide. The "Building with the Claude API" specialization teaches developers how to integrate Anthropic's Claude AI into applications, covering topics like prompt engineering and building AI agents. The "Real-World AI for Everyone" specialization, created with Advancing Women in Technology, helps professionals apply AI ethically for writing, analysis, and communication. These programs address a critical AI skills gap, as only one-third of employees currently have AI training. Learners can enroll in these Specializations starting today on Coursera.

Coursera and Anthropic Offer New AI Training Programs

On November 18, 2025, Coursera partnered with Anthropic to launch two new AI training specializations. The "Building with the Claude API" course targets developers, teaching them prompt engineering, model context protocols, and how to build AI applications using Anthropic's tools. The "Real-World AI for Everyone" program, developed with Advancing Women in Technology, focuses on ethical AI use for writing, analysis, and communication for all professionals. These specializations aim to close the AI skills gap, as only a third of employees currently receive AI training. Coursera, with its 191 million learners, continues to expand access to cutting-edge AI education.

Anthropic Launches Claude AI Courses on Coursera

On November 18, 2025, Anthropic partnered with Coursera to launch two new AI courses, "Building with the Claude API" for developers and "Real-World AI for Everyone" for general professionals. These courses aim to boost AI literacy and teach safe, effective use of Anthropic's Claude models. This educational push comes as Anthropic expands its collaboration with Microsoft and Nvidia, integrating Claude models like Sonnet 4.5 into Microsoft Azure. Anthropic, founded in 2021, emphasizes AI safety, even as it recently uncovered an AI-orchestrated cyber espionage campaign using Claude Code. The Coursera collaboration offers early access now, with a full launch in February 2026.

Google Launches Gemini 3 AI to Enhance Search

On Tuesday, November 18, 2025, Google unveiled its next-generation Gemini 3 AI model, integrating it into its search engine and other online services. This new model aims to transform Google Search into a "thought partner" with a new AI "thinking" feature. Gemini 3 will first roll out to Gemini Pro and Ultra subscribers in the United States before expanding globally. Google executives, including Koray Kavukcuoglu and Demis Hassabis, emphasized built-in guardrails to ensure smart, concise, and direct responses, preventing misinformation. The company has significantly increased its AI investment, with Alphabet's market value now around $3.4 trillion, as it competes with other AI models like ChatGPT and Claude.

Google Releases Gemini 3 AI Model for Search

On Tuesday, November 18, 2025, Google launched Gemini 3, its newest artificial intelligence model. CEO Sundar Pichai called it "our most intelligent model." This third generation of Gemini is immediately available in key Google products, including its search engine. The rapid release, just 11 months after the previous version, shows Google's commitment to staying at the forefront of the AI race. This integration aims to enhance user experience by embedding advanced AI capabilities directly into widely used services.

Most Manufacturers Struggle with High AI Costs

A recent survey revealed that 73% of manufacturers find the costs of investing in Artificial Intelligence to be a heavy burden. Small and Medium-sized Enterprises, or SMEs, are especially impacted by these high costs. They also face a shortage of skilled workers and have doubts about how effective AI will truly be in their manufacturing processes. This highlights a significant challenge for businesses looking to adopt AI technology.

Velantra AI Discusses AI and Human Trading Partnership

On November 18, 2025, Velantra AI discussed the roles of AI and human traders in financial markets. While AI algorithms offer advantages like rapid data analysis and consistent execution, they can struggle with unexpected market events. Humans, however, provide crucial judgment, context, and adaptability that machines lack. Velantra AI emphasizes that human oversight is vital, with teams regularly reviewing system behavior. The firm also builds in protective measures, such as a 30% capital protection safeguard and strategy diversification, to manage risks. Velantra AI believes the future of trading lies in a partnership that combines AI's speed with human strategic insight and risk management.

Dell Executive Urges Investment in AI Infrastructure

Nicole Jefferson, Vice President of Dell Technologies Government Affairs, highlights three key areas needed to power the AI era: modern data centers, a modern energy grid, and modern regulations. She notes that demand for data centers is rapidly increasing, requiring streamlined permitting processes across all government levels. AI's significant energy needs mean the US must accelerate transmission, expand energy generation, and improve demand flexibility. Jefferson also stresses the importance of agile, pro-growth policies that keep pace with technology. She believes that government and industry must work together to build secure, efficient data center capacity to ensure America's AI leadership and national security.

Rapid GPU Upgrades Threaten AI Hyperscaler Finances

The rapid pace of innovation in GPUs, with Nvidia pushing for annual releases, poses a significant financial risk for AI hyperscalers. Analysts warn that the quick obsolescence of GPUs could lead to a major depreciation crisis. While traditional servers last 3-5 years, AI "factories" need the latest technology, making older hardware quickly unprofitable. This creates a challenge for companies that invest billions in GPUs, especially as they compete for limited new hardware and struggle to sell off older models. The AI industry's financing often relies on GPUs as collateral, and a sudden drop in their value could impact cash flow and profitability.

Bernie Sanders Warns AI Risks Inequality and War

On November 18, 2025, Senator Bernie Sanders warned that artificial intelligence poses serious risks, including deepening wealth inequality and reshaping warfare. He expressed concern about the rise of "robot warriors" and how this could change decisions about war and peace. Sanders highlighted that billionaires are investing heavily in AI, which could further concentrate wealth and power, potentially leading to a decline in living standards for working people. He also raised concerns about young people relying on AI for companionship. Sanders called for more public discussion, legislation, and potentially breaking up large AI companies to ensure AI benefits everyone, not just a few wealthy individuals.

Investors Scrutinize AI Company Valuations More Closely

Jack Janasiewicz, a lead portfolio strategist at Natixis Investment Managers, stated that investors are now scrutinizing AI companies more closely. This means they are no longer giving AI names a "free pass" in terms of valuation. His comments followed Rothschild's decision to downgrade Microsoft to neutral after its deal with Anthropic. This shift suggests that the market is becoming more cautious and demanding clearer returns from AI investments.

Jiga Raises $12M to Speed Up AI Hardware Production

On November 18, 2025, Jiga announced it secured $12 million in Series A funding to solve a major bottleneck in AI development: slow hardware sourcing. While AI software moves quickly, getting physical parts can take weeks, hindering progress. Jiga's AI-native manufacturing platform allows engineers to upload designs, get matched with manufacturers, and order parts in hours instead of weeks, significantly reducing costs. The platform uses AI to manage requirements, detect risks, and streamline communication. Founded by Adar Hay, Yonatan Wolowelsky, and Assaf Geuz, Jiga is already profitable and will use the new capital to expand its production capabilities and quality assurance for the rapidly growing robotics and space industries.

Sources

• 5 Biggest CX Takeaways from Microsoft Ignite 2025
• Preparing for what’s next: Windows security and resiliency innovations help organizations mitigate risks, recover faster and prepare for the era of AI
• Microsoft Ignite: ​​Ambient and autonomous security for the agentic era​​ 
• Microsoft Defender Adds AI, Cloud, Multicloud Security Tools
• Microsoft leans hard on autonomous security for AI agents
• Edge for Business presents: the world’s first secure enterprise AI browser
• Agents built into your workflow: Get Security Copilot with Microsoft 365 E5
• Microsoft Ignite 2025: Copilot and agents built to power the Frontier Firm
• Coursera (NYSE: COUR) and Anthropic debut Claude AI Specializations for global workers
• Coursera partners with Anthropic to launch AI training specializations
• Anthropic’s AI Education Push: Claude Courses Hit Coursera Amid Tech Giant Deals
• Google Unveils Gemini's Next Generation, Aiming to Turn Its Search Engine Into a 'Thought Partner'
• Google launches Gemini 3, embeds AI model into search immediately
• 73% of Manufacturers Find AI Investment Costs Burdensome
• Velantra AI on AI vs Human Traders
• Powering the AI Era—Reinvent Data Centers, Rewire the Grid and Unleash Innovation
• GPU depreciation could be the next big crisis coming for AI hyperscalers — after spending billions on buildouts, next-gen upgrades may amplify cashflow quirks
• Bernie Sanders warns AI could deepen inequality and reshape war
• Investors are no longer willing to give AI names a ‘free pass’, says strategist
• Jiga Secures $12M Series A to Eliminate Hardware Sourcing Bottleneck Threatening AI Scale