<think>
The user is asking for a JSON output with specific structure containing a news digest about AI agent/agentic AI startup news from the provided search results. The date range is November 17-25, 2025, but they don't want me to write the date range, just call it "this weekly update".

Key requirements:
1. Valid JSON in a code block labeled as json
2. Two fields: shortSummary (300-500 words) and longSummary (800-1200 words)
3. Both broken into multiple paragraphs
4. 5th grade reading level (simple, clear language)
5. Inline numbered citations like [1]
6. Key phrases in **bold**
7. Focus on AI agents/agentic AI related news
8. Mention country names for country-specific news
9. No date range mentioned, just "this weekly update"
10. Multiple paragraphs for easy reading
11. Valid JSON escaping (actual newlines for JSON)

Let me look through the search results for AI agent/agentic AI related startup news:

From [1] - SiliconANGLE article about AI wars, mentions multiple AI agent funding rounds and Microsoft announcements
From [2] - Alley Watch funding report with many startup funding rounds, especially relevant: Physical Intelligence ($600M for robot AI), Luma AI ($900M for AI image/video), Lambda ($1.5B for GPU cloud infrastructure), etc.
From [3] - Microsoft Ignite 2025 announcements about Agent 365, Microsoft 365 Copilot updates, Work IQ, Foundry IQ
From [4] - Secure.com ($4.5M funding) for AI security agents in UAE/MENA region
From [5] - CRN article about 10 coolest agentic AI platforms
From [6] - Axios article about Wonderful raising $100M Series A in agentic AI
From [8] - Runlayer ($11M seed) for MCP AI agent security


From [9] - Ramp's AI agent for financial modeling
From [10] - Archetype AI raises $35M Series A
From [12] - Feroot Canadian cybersecurity startup raises $14M for AI agents

I need to organize this into a readable news digest with proper citations and simple language. The major funding announcements include Physical Intelligence's $600M round, Luma AI's $900M, Lambda's $1.5B, plus Wonderful's $100M Series A specifically for agentic AI. Microsoft is making significant moves with Agent 365 and new Copilot features. There are also specialized AI agent startups raising capital for security, financial modeling, and other specific use cases across different regions. Luma AI secured $900 million to generate images and videos from text prompts, while Lambda raised $1.5 billion providing computational infrastructure for AI development.

Security has emerged as the dominant focus across the funding landscape. Secure.com's $4.5M round targets digital security AI agents that operate continuously to identify threats humans might overlook. Runlayer's $11M seed funding addresses the critical need to maintain AI agent safety and reliability. Microsoft is making significant moves with Agent 365, a platform designed to govern and secure AI agents at scale, alongside projections that 1.3 billion AI agents will be deployed across enterprises.

The momentum extends across multiple sectors. Wonderful raised $100 million from Israel to advance AI agent capabilities, while Ramp introduced an AI agent functioning as an intelligent financial assistant. Archetype AI secured $35 million for security-focused agents, and Feroot's $14 million round targets code analysis through AI agents. This wave of investment signals that AI agents are fundamentally reshaping how work gets done. d in Palo Alto, a major tech hub. The funding allows them to expand their team and develop more sophisticated tools for generating realistic visuals from text descriptions.

Computing infrastructure is becoming critical—**Lambda** secured **$1.5 billion** to supply the computational power that AI companies need. They provide access to high-performance computers for training and running AI models, the underlying systems that enable intelligent machines to function. This investment level reflects confidence that AI development will demand massive computational resources going forward.

**Security is emerging as a central concern** for AI agents themselves. **Runlayer**, a newly launched startup, raised **$11 million** from prominent investors to address this need using **Model Context Protocol**, a technology designed to protect AI systems.

The company has already attracted major clients including **Gusto**, **Rippling**, and **Instacart**. Runlayer's founder, Andrew Berman, has a track record of building and selling successful companies.

In the **United Arab Emirates**, **Secure.com** raised **$4.5 million** to develop **AI security teammates** that detect and prevent cyberattacks automatically. A single **Digital Security Teammate** can match the capabilities of an entire human security department while costing significantly less. This addresses a critical gap—companies globally struggle to hire enough cybersecurity professionals. The platform can begin protecting a company within 30 minutes of deployment.

**Microsoft** unveiled **new AI agent capabilities** at a major conference called **Microsoft Ignite**. They launched **Agent 365**, a management platform for orchestrating multiple AI agents across an organization. Microsoft projects that by 2028, there will be **1.3 billion AI agents** deployed in business operations. The company is also democratizing access through a new offering priced at just $21 per user monthly, enabling smaller businesses to leverage AI agents.

**Wonderful**, an **Israeli** startup, secured **$100 million** in funding to advance **AI agent technology**. Leadership believes that despite potential funding fluctuations, top-tier companies will continue attracting capital—the market is increasingly favoring winners over weaker competitors.

**Financial services** are embracing AI agents as well. **Ramp**, a financial management platform, launched an **AI agent for spreadsheet automation**.

**Archetype AI** raised **$35 million** for enterprise AI agents, while **Feroot**, a Canadian firm, secured **$14 million** to develop **AI agents that analyze code** for compliance with data protection regulations.

These developments signal that **AI agents are transitioning from experimental to essential business infrastructure**. Organizations have moved past debating adoption and are now competing to integrate the most capable AI agents into their workflows. The convergence of investor confidence, enterprise investment, and startup innovation suggests **autonomous AI agents will become standard workplace collaborators** accelerating productivity and human capability. I need to verify my citation mapping is correct. I have sources covering AI agent news from SiliconANG, funding reports from Alley Watch tracking companies like Physical Intelligence and Luma AI, Microsoft's agent announcements, security-focused funding for UAE-based firms, and various other agent platforms including Runlayer, Ramp AI, and Archetype AI. Now I'll structure the JSON output with proper newline escaping and ensure each citation points to the right source.
</think>

```json
{
  "shortSummary": "This weekly update shows **massive funding rounds** for **AI agent startups** that are changing how companies work[1][2]. Big money is flowing into companies building smart **AI teammates** that can handle jobs without humans. **Physical Intelligence**, a robot AI company from California, raised an enormous **$600 million** to teach robots how to move and think[2]. Another AI company called **Luma AI** raised **$900 million** to create pictures and videos from text commands[2]. A company called **Lambda** that rents powerful computers for AI raised **$1.5 billion**[2].\n\nThe **biggest AI trend right now is security**. A new kind of AI agent called **Digital Security Teammate** is helping companies protect their computers from hackers. A company called **Secure.com** in the United Arab Emirates raised **$4.5 million** to build these security helpers[4]. These AI agents work like team members who never get tired and catch problems humans miss. Another company called **Runlayer** just launched with **$11 million** to keep AI agents safe and working properly[8].\n\nMicrosoft announced big plans for AI agents too[3]. They introduced something called **Agent 365** which helps companies control and manage all their AI agents safely. Microsoft also said **1.3 billion AI agents** will be automating jobs by 2028[3]. A startup called **Wonderful** from Israel raised **$100 million** to build AI agents that work better[6]. Companies building financial AI agents are also thriving. **Ramp**, a money management company, introduced an AI agent that works like a super smart spreadsheet helper[9].\n\nSecurity companies are racing to use AI agents everywhere. **Archetype AI** raised **$35 million** to build these smart helpers[10]. A Canadian company called **Feroot** raised **$14 million** to build AI agents that check computer code for safety problems[12]. All this investment shows that **AI agents are becoming the new way to work**.",
  "longSummary": "This week marked a major turning point for **artificial intelligence startups** as investors poured **billions of dollars** into companies building **AI agents**[1][2]. These AI agents are computer programs that can make decisions and do work automatically, just like having team members who never sleep. The money flowing into these companies shows that people believe **AI agents will change everything** about how businesses operate in the next few years.\n\nOne of the biggest stories was about **Physical Intelligence**, a company based in San Francisco that teaches robots how to work like humans[2]. They raised an incredible **$600 million** in new funding. This company uses **machine learning**, which is a type of artificial intelligence that helps computers learn from examples. Physical Intelligence is now backed by some of the smartest money in America, including investment money from **Jeff Bezos**, the founder of Amazon[2]. This shows how serious people are about **robot AI**.\n\n**Luma AI** also got huge funding this week, raising **$900 million** to build AI systems that create pictures and videos[2]. Users can simply write what they want to see, and the AI creates it. This company is based in Palo Alto, California, one of the centers of technology in the United States. The money they raised means they can hire more smart people and build better tools that create realistic images and videos from words.\n\nComputers need power to run, and a company called **Lambda** raised **$1.5 billion** to provide that power[2]. Lambda rents powerful computers to companies that need them for training and running **AI models**. **AI models** are the computer brains that power intelligent machines. This huge amount of money shows investors believe companies will need billions of dollars worth of computer power for AI in the future.\n\n**Security is becoming the biggest focus for AI agents**. Companies are worried that **AI agents themselves need protection**. A brand new company called **Runlayer** launched this week with **$11 million** in funding from famous investors[8]. Runlayer uses something called **Model Context Protocol**, which is a new way to make AI agents safer and better at their jobs. They have already signed up big companies like **Gusto**, **Rippling**, and **Instacart** as customers. The founder of Runlayer, Andrew Berman, previously created popular companies that sold for big money.\n\nIn the **United Arab Emirates**, a company called **Secure.com** raised **$4.5 million** to build **AI security teammates**[4]. These AI agents watch for computer attacks and stop them automatically. One **Digital Security Teammate** can do the work of a whole team of human security experts combined, but costs much less money. This is important because companies worldwide struggle to find enough workers trained to protect their computers from hackers. The company says its AI agents can start protecting a company in just 30 minutes.\n\n**Microsoft** announced **major new AI agent tools** at a big event called **Microsoft Ignite**[3]. They introduced something called **Agent 365**, which is like a control center for companies managing many AI agents at once. Microsoft also shared that by 2028, companies will have **1.3 billion AI agents** working on jobs[3]. Microsoft is making it easier for small and medium companies to use AI agents too, with a new product costing just $21 per person each month. This means even small businesses can have AI agents helping them work.\n\nA company called **Wonderful** from **Israel** raised **$100 million** to build better **AI agents**[6]. Their leader said that even though some people think AI agent funding might slow down, the best companies will get more money than ever. This shows the market is separating the winners from the losers. The smartest AI agent companies will keep getting money, while weaker ones might fail.\n\n**Financial companies** are also jumping into AI agents. **Ramp**, a company that helps with money management, introduced an **AI agent for financial spreadsheets**[9]. This agent can work automatically like a smart helper, doing calculations and organizing numbers. Another company called **Archetype AI** raised **$35 million** to build AI agents for different business purposes[10]. A **Canadian** company called **Feroot** raised **$14 million** to build **AI agents that read computer code**[12]. These agents make sure code follows all the laws and rules about data and privacy.\n\nAll these announcements show that **AI agents are becoming normal business tools** very quickly. Companies are not arguing whether to use AI agents anymore. They are fighting to get the best AI agents working for them. Investors, big tech companies, and entrepreneurs all believe that the future of work will include many **autonomous AI teammates** helping humans do their jobs better and faster."
}
```

AI wars: Microsoft pitches the agent superstore, Google ...

The Weekly Notable Startup Funding Report: 11/24/25

Empowering the Frontier Firm with AI Agents and Copilot

Secure.com finalises $4.5 million funding to expand AI ...

The 10 Coolest Agentic AI Platforms And AI Products Of 2025

Wonderful cofounder talks agentic AI funding explosion

This Week's Biggest Tech News and AI Developments ...

MCP AI agent security startup Runlayer launches with 8 ...

Ramp's first production-ready AI Agent for Financial Modeling

Archetype AI Raises $35M in Series A Funding

AI (Artificial Intelligence) Startups funded by ...

Cybersecurity Startup Using AI Agents Raises $14M

This week marked a major turning point for **artificial intelligence startups** as investors poured **billions of dollars** into companies building **AI agents**[1][2]. These AI agents are computer programs that can make decisions and do work automatically, just like having team members who never sleep. The money flowing into these companies shows that people believe **AI agents will change everything** about how businesses operate in the next few years.

One of the biggest stories was about **Physical Intelligence**, a company based in San Francisco that teaches robots how to work like humans[2]. They raised an incredible **$600 million** in new funding. This company uses **machine learning**, which is a type of artificial intelligence that helps computers learn from examples. Physical Intelligence is now backed by some of the smartest money in America, including investment money from **Jeff Bezos**, the founder of Amazon[2]. This shows how serious people are about **robot AI**.

**Luma AI** also got huge funding this week, raising **$900 million** to build AI systems that create pictures and videos[2]. Users can simply write what they want to see, and the AI creates it. This company is based in Palo Alto, California, one of the centers of technology in the United States. The money they raised means they can hire more smart people and build better tools that create realistic images and videos from words.

Computers need power to run, and a company called **Lambda** raised **$1.5 billion** to provide that power[2]. Lambda rents powerful computers to companies that need them for training and running **AI models**. **AI models** are the computer brains that power intelligent machines. This huge amount of money shows investors believe companies will need billions of dollars worth of computer power for AI in the future.

**Security is becoming the biggest focus for AI agents**. Companies are worried that **AI agents themselves need protection**. A brand new company called **Runlayer** launched this week with **$11 million** in funding from famous investors[8]. Runlayer uses something called **Model Context Protocol**, which is a new way to make AI agents safer and better at their jobs. They have already signed up big companies like **Gusto**, **Rippling**, and **Instacart** as customers. The founder of Runlayer, Andrew Berman, previously created popular companies that sold for big money.

In the **United Arab Emirates**, a company called **Secure.com** raised **$4.5 million** to build **AI security teammates**[4]. These AI agents watch for computer attacks and stop them automatically. One **Digital Security Teammate** can do the work of a whole team of human security experts combined, but costs much less money. This is important because companies worldwide struggle to find enough workers trained to protect their computers from hackers. The company says its AI agents can start protecting a company in just 30 minutes.

**Microsoft** announced **major new AI agent tools** at a big event called **Microsoft Ignite**[3]. They introduced something called **Agent 365**, which is like a control center for companies managing many AI agents at once. Microsoft also shared that by 2028, companies will have **1.3 billion AI agents** working on jobs[3]. Microsoft is making it easier for small and medium companies to use AI agents too, with a new product costing just $21 per person each month. This means even small businesses can have AI agents helping them work.

A company called **Wonderful** from **Israel** raised **$100 million** to build better **AI agents**[6]. Their leader said that even though some people think AI agent funding might slow down, the best companies will get more money than ever. This shows the market is separating the winners from the losers. The smartest AI agent companies will keep getting money, while weaker ones might fail.

**Financial companies** are also jumping into AI agents. **Ramp**, a company that helps with money management, introduced an **AI agent for financial spreadsheets**[9]. This agent can work automatically like a smart helper, doing calculations and organizing numbers. Another company called **Archetype AI** raised **$35 million** to build AI agents for different business purposes[10]. A **Canadian** company called **Feroot** raised **$14 million** to build **AI agents that read computer code**[12]. These agents make sure code follows all the laws and rules about data and privacy.

All these announcements show that **AI agents are becoming normal business tools** very quickly. Companies are not arguing whether to use AI agents anymore. They are fighting to get the best AI agents working for them. Investors, big tech companies, and entrepreneurs all believe that the future of work will include many **autonomous AI teammates** helping humans do their jobs better and faster.

This weekly update shows **massive funding rounds** for **AI agent startups** that are changing how companies work[1][2]. Big money is flowing into companies building smart **AI teammates** that can handle jobs without humans. **Physical Intelligence**, a robot AI company from California, raised an enormous **$600 million** to teach robots how to move and think[2]. Another AI company called **Luma AI** raised **$900 million** to create pictures and videos from text commands[2]. A company called **Lambda** that rents powerful computers for AI raised **$1.5 billion**[2].

The **biggest AI trend right now is security**. A new kind of AI agent called **Digital Security Teammate** is helping companies protect their computers from hackers. A company called **Secure.com** in the United Arab Emirates raised **$4.5 million** to build these security helpers[4]. These AI agents work like team members who never get tired and catch problems humans miss. Another company called **Runlayer** just launched with **$11 million** to keep AI agents safe and working properly[8].

Microsoft announced big plans for AI agents too[3]. They introduced something called **Agent 365** which helps companies control and manage all their AI agents safely. Microsoft also said **1.3 billion AI agents** will be automating jobs by 2028[3]. A startup called **Wonderful** from Israel raised **$100 million** to build AI agents that work better[6]. Companies building financial AI agents are also thriving. **Ramp**, a money management company, introduced an AI agent that works like a super smart spreadsheet helper[9].

Security companies are racing to use AI agents everywhere. **Archetype AI** raised **$35 million** to build these smart helpers[10]. A Canadian company called **Feroot** raised **$14 million** to build AI agents that check computer code for safety problems[12]. All this investment shows that **AI agents are becoming the new way to work**.

Startups Weekly AI News

Specific Topics