Amazon Launches AI Factories While Nevis Secures $40 Million

AWS made significant announcements at its re:Invent conference on December 2, 2025, unveiling its new AI Factories. This service provides a fully managed AI solution designed for businesses and governments needing to keep sensitive data on-premises. Customers supply the physical data center and power, while AWS handles the installation and management of its AI hardware and software, including Amazon Bedrock and SageMaker. This setup allows organizations to leverage advanced AI tools like Trainium3 and Nvidia GPUs, specifically the Blackwell architecture HGX B300, while adhering to strict security and data sovereignty regulations. AWS CEO Matt Garman noted that the concept originated from building an "AI Zone" for Saudi Arabia. The launch of AI Factories further deepens a 15-year partnership between AWS and Nvidia, with AWS supporting Nvidia's NVLink Fusion and incorporating its own upcoming Trainium4 and Graviton chips. This collaboration also extends to integrating Nvidia software such as Nemotron Nano 2 for generative AI and Nvidia cuVS for Amazon OpenSearch Service, alongside advancements in physical AI for robotics using Cosmos world foundation models on Amazon EKS. Concurrently, the focus on AI security is growing, with Skyflow introducing its Runtime AI Data Security platform for AWS AgentCore to protect sensitive customer data like PII, PHI, and PCI within AI workflows. Cisco also launched its AI Defense solution, which scans the AI supply chain for vulnerabilities and provides runtime protections for AI agents to prevent data leakage and harmful actions. Beyond infrastructure and security, AI is finding diverse applications. Lee Health in Fort Myers, Florida, began using AI in all mammograms in September, enhancing breast cancer screenings for better accuracy and speed. Wisconsin counties, La Crosse and Waukesha, are deploying AI platforms like "Ava" to manage non-emergency calls, freeing up dispatchers for urgent 911 situations. In finance, AI and machine learning are transforming trading by processing vast information and managing risk in real-time. Nevis, a new AI platform for wealth management, recently secured $40 million in funding from investors including Sequoia, ICONIQ, and Ribbit Capital. Nevis aims to help financial advisors, whose current customers oversee over $50 billion in assets, dedicate more time to clients by automating administrative tasks. However, the surge in AI demand is also impacting hardware costs, as Raspberry Pi CEO Eben Upton announced price increases of $5 to $25 for its Raspberry Pi 4 and 5 modules, attributing the hike to strong competition for memory from the AI industry.

Key Takeaways

• AWS launched AI Factories on December 2, 2025, offering fully managed AI infrastructure for private data centers, including Amazon Bedrock, SageMaker, Trainium, and Nvidia GPUs, addressing data sovereignty needs.
• The partnership between AWS and Nvidia expanded significantly, integrating Nvidia Blackwell architecture GPUs (HGX B300), NVLink Fusion, and software like Nemotron Nano 2, alongside AWS's Trainium4 and Graviton chips.
• AWS CEO Matt Garman indicated the AI Factories concept was inspired by building an "AI Zone" for Saudi Arabia.
• Skyflow introduced its Runtime AI Data Security platform for AWS AgentCore to protect sensitive data (PII, PHI, PCI) in AI agent workflows, ensuring compliance with regulations like GDPR and HIPAA.
• Cisco launched its AI Defense solution, providing AI supply chain scanning and runtime protections for AI agents to prevent vulnerabilities and data leakage.
• Nevis, an AI platform for wealth management, raised $40 million in funding from Sequoia, ICONIQ, and Ribbit Capital, aiming to help advisors manage over $50 billion in assets more efficiently.
• Raspberry Pi CEO Eben Upton announced price increases of $5 to $25 for Raspberry Pi 4 and 5 modules due to strong competition for memory from the AI industry, with 16GB Compute Module 5 variants now starting at $140.
• AI is being adopted across various sectors, including Lee Health for faster and more accurate breast cancer screenings, and Wisconsin counties (La Crosse and Waukesha) for managing non-emergency calls.
• The use of AI in strategic foresight and financial trading is growing, enabling faster data analysis, complex pattern detection, and real-time risk management, but also introducing new security and market stability risks.
• Securing AI vector embeddings is critical, as they are information-dense and susceptible to data reconstruction, necessitating secure storage like MySQL's VECTOR Data Type, strict access controls, and encryption.

AWS launches AI Factories for private data centers

AWS announced its new AI Factories at the re:Invent conference on Tuesday, December 2, 2025. This service offers a fully managed AI solution for businesses and governments that need to keep sensitive data on-premises. Customers provide the physical datacenter and power, while AWS installs and manages its own AI hardware and software, including Amazon Bedrock and SageMaker. This allows organizations to use advanced AI tools like Trainium3 and Nvidia GPUs while meeting strict security and data sovereignty rules. AWS CEO Matt Garman said the idea came from building an "AI Zone" for Saudi Arabia.

AWS AI Factories bring advanced AI to customer data centers

AWS introduced its new AI Factories, a service that puts dedicated AWS AI infrastructure directly into customer data centers. This helps businesses and governments quickly build and use AI applications while meeting strict data rules. Customers use their own data center space and power, and AWS manages the advanced hardware and software. The factories include NVIDIA and AWS Trainium chips, high-speed networking, and services like Amazon SageMaker and Amazon Bedrock. This expansion builds on a 15-year partnership between AWS and NVIDIA to offer top-tier AI solutions.

Nvidia and AWS boost AI with new chips and software

Nvidia and AWS announced an expanded partnership at the re:Invent conference on December 2, 2025, focusing on new AI hardware and software. AWS will support Nvidia's NVLink Fusion and use its own upcoming Trainium4 and Graviton chips. The collaboration also brings Nvidia Blackwell architecture GPUs, like the HGX B300, to AWS AI Factories. This aims to provide secure AI infrastructure globally and for the public sector. The partnership also includes integrating Nvidia software like Nemotron Nano 2 for generative AI and Nvidia cuVS for Amazon OpenSearch Service. They are also advancing physical AI for robotics with Cosmos world foundation models on Amazon EKS.

NVIDIA and AWS deepen partnership for advanced AI

NVIDIA and AWS announced a major expansion of their partnership at AWS re:Invent, focusing on advanced AI computing. AWS will support NVIDIA NVLink Fusion for its custom AI infrastructure, including next-generation Trainium4 chips and Graviton CPUs. This collaboration aims to boost performance and speed up the release of new cloud AI features. AWS is also adding NVIDIA Blackwell architecture GPUs, like the HGX B300, to its offerings. Both companies are committed to deploying sovereign AI clouds worldwide, with AWS AI Factories providing secure, high-performance AI infrastructure in customer data centers.

Skyflow secures sensitive data in AI workflows

On December 2, 2025, Skyflow launched its Runtime AI Data Security platform for AWS AgentCore. This platform helps businesses protect sensitive customer data like PII, PHI, and PCI when using AI agents. Unlike traditional tools that only block data, Skyflow offers fine-grained controls to keep data secure even while agents process it. The solution integrates with Amazon Quick Suite and Amazon Bedrock AgentCore, ensuring compliance with rules like GDPR and HIPAA. It also makes sure agents only access the necessary data and provides detailed logs for auditing.

Cisco AI Defense protects AI agents and supply chain

Cisco introduced its AI Defense solution to secure AI agents and the AI supply chain. Many organizations plan to use AI agents, but they face security challenges. AI Defense now includes AI supply chain scanning to find vulnerabilities in third-party models and tools, like a malicious MCP server found by Koi Security. It also offers special runtime protections for AI agents, acting as an MCP gateway. This helps prevent harmful actions, data leakage, and other risks as AI systems become more complex and autonomous.

Kraken uses AI to boost energy grid resilience

Kraken is using its "utility grade AI" to improve energy grid resilience and lower costs. Assaf Biderman, Kraken's Chief AI Officer, explains that a unified data model brings all energy data together, allowing AI to make better correlations. This system also has strong security and creates client-specific knowledge bases to meet regulations. Kraken combines its deep energy industry knowledge with AI, providing important context for its solutions. The company also uses a "human-in-the-loop" approach, empowering universal agents to provide better customer service and increase satisfaction.

Experts use AI to predict the future responsibly

AI is now a key tool in strategic foresight, helping governments and businesses plan for the future. A mid-2025 survey by the OECD and World Economic Forum found that two-thirds of foresight experts already use AI. They use AI at three levels: first, for basic tasks like data synthesis; second, as a creative partner for ideas and testing; and third, fully integrated into their workflow with AI agents. While AI offers benefits like time efficiency and better data analysis, experts also worry about risks like bias. The goal is to use AI ethically to improve future planning.

AI helps Lee Health fight breast cancer

Lee Health in Fort Myers, Florida, began using artificial intelligence in all mammograms in September. The AI system first reads the mammograms, then doctors review the images and use the AI's findings. Radiologist Dr. Priyanka Handa states that this new technology makes breast cancer screenings better, faster, and more accurate. This integration gives doctors a major boost in the fight against breast cancer.

Secure AI vector embeddings in MySQL

AI applications use vector embeddings for search and recommendations, but these vectors create new security and privacy risks. Attackers can partially reconstruct original data or generate similar data from stolen vectors. This happens because embeddings are information-dense and not cryptographically secure. To protect this data, businesses should use secure storage like MySQL's VECTOR Data Type, implement strict access controls, and encrypt data. Best practices include limiting access, enabling auditing, and regularly reviewing security to prevent data leakage and ensure compliance.

Wisconsin counties use AI for non-emergency calls

Two Wisconsin counties, La Crosse and Waukesha, are now using AI to handle non-emergency calls. La Crosse County started using its AI platform, "Ava," in May, handling over 40,000 calls. This system helps dispatchers focus on urgent 911 calls by managing simple requests like parking questions. If an emergency caller dials the non-emergency line, "Ava" recognizes "hot words" and redirects them to 911. Waukesha County plans to adopt a similar system by the end of the year to help its understaffed communications center. The AI can also send texts with information like parade routes.

AI and machine learning reshape financial trading

Artificial intelligence and machine learning are changing how financial trading works, moving beyond human intuition. Algorithms process vast amounts of information in milliseconds, allowing for faster trading decisions. These models can track hundreds of variables, finding complex patterns that humans might miss, even if the reasons behind some patterns are unclear. AI also makes risk management more complex, adjusting based on real-time market conditions and detecting unusual behavior faster than humans. While AI offers significant advantages, it also introduces new risks, such as models not adapting quickly to changing markets or even causing market instability.

Nevis AI platform helps wealth managers save time

Nevis, a new AI platform, aims to transform wealth management by giving financial advisors more time for clients. Founded by Mark Swan, Philipp Burda, and Ivan Chalov, Nevis helps advisors who spend 80% of their time on administrative tasks. The New York City-based company believes AI will enhance, not replace, human advisors. Nevis recently raised $40 million in funding from Sequoia, ICONIQ, and Ribbit Capital. Its platform helps wealth managers serve more clients and build deeper relationships, with current customers overseeing over $50 billion in assets.

Raspberry Pi prices increase due to AI demand

Raspberry Pi is raising prices for its Raspberry Pi 4 and 5 modules immediately, with increases ranging from $5 to $25. CEO Eben Upton stated that the price hike is due to strong competition for memory from the AI industry. For example, 16GB Compute Module 5 variants now start at $140, up by $20. Upton hopes to reverse these increases once memory prices stabilize. In some good news, the company also announced a new $45 model featuring a quad-core 2.4GHz Arm Cortex-A76 processor.

Sources

• AWS AI Factories: AI-in-a-box for enterprise datacenters
• New AWS AI Factories transform customers’ existing infrastructure into high-performance AI environments
• Nvidia and AWS expand partnership with specialized AI hardware and software
• NVIDIA and AWS Expand Full-Stack Partnership, Providing the Secure, High-Performance Compute Platform Vital for Future Innovation
• Skyflow delivers Runtime AI Data Security for protecting sensitive data in agentic workflows
• Securing Agents & AI Supply Chain with Cisco AI Defense
• Kraken’s utility grade AI
• How to responsibly integrate AI into strategic foresight
• Video: Artificial Intelligence joins the battle against breast cancer
• Protecting AI Vector Embeddings in MySQL: Security Risks, Database Protection, and Best Practices
• These 2 Wisconsin counties are using AI to handle non-emergency calls
• AI in finance: how machine learning is transforming trading
• Nevis: Bringing AI to Wealth Management
• Raspberry Pi raises prices, thanks to AI