CoreWeave Launches Zero Trust AI Framework at NVIDIA GTC Berlin

CoreWeave launched its Zero Trust AI Framework at NVIDIA GTC Berlin, giving customers granular control over access to models, data, and runtime environments. The framework relies on six layers of protection, including Zero Key Access, Zero Node Access, and Zero Trust Infrastructure, and supports confidential computing with network encryption through IEEE 802.1AE MACsec.

At the same event, Lambda outlined its collaboration plans with NVIDIA, including support for Confidential Bare Metal and the NVIDIA OpenShell Reference Architecture for agent safety. Lambda is also deploying NVIDIA Minerva for infrastructure monitoring and testing NVIDIA DSX MaxLPS to improve token throughput for frontier model builders.

In the enterprise AI space, Patra appointed Yousef Hashimi as Head of Products and Client AI Transformation. With nearly 30 years of experience from Lüt, IBM, and EY, Hashimi previously helped commercialize IBM Watson and will now expand Patra's agentic AI services for the insurance industry, starting with post-bind processing.

Mach9 Robotics rebranded to Mach9 and raised a $34 million Series A round led by First Round Capital and Bessemer Venture Partners. The company is building AI-powered CAD software for civil engineers and has already reached nearly 100 customers and over 2,500 projects during its private beta.

New research shows that AI-assisted shopping is growing, with nearly 61 million U.S. consumers using AI as their primary starting point for online shopping. Average AI-assisted spending rose 17 percent between June and August surveys, yet 60 percent of respondents completed their most recent AI-assisted purchase on Amazon, highlighting the platform's dominance in this emerging behavior.

Walmart and Amazon are diverging in their AI shopping strategies. Walmart is embracing AI shopping agents and investing heavily in the technology, while Amazon is taking a more cautious approach and has blocked its Muse platform. The contrasting paths reflect the broader competition to capture AI-driven consumer behavior.

On the infrastructure front, T-Mobile and Ericsson are running live AI-RAN network trials in New York City, using AI-managed beamforming on live traffic to improve signal strength and reduce latency. Meanwhile, Naver reported that its third-quarter operating profit is expected to fall 12.4 percent to 500 billion won as AI infrastructure costs, including GPU depreciation, exceed 200 billion won per quarter.

Key Takeaways

  • CoreWeave launched a Zero Trust AI Framework with six security layers at NVIDIA GTC Berlin, giving customers control over model and data access.
  • Lambda announced multiple NVIDIA collaborations at GTC Berlin 2026, including Confidential Bare Metal support and the OpenShell Reference Architecture for agent safety.
  • Patra hired Yousef Hashimi, a former IBM Watson commercializer, to lead AI product development and expand agentic AI services for insurance clients.
  • Mach9 raised $34 million in a Series A round led by First Round Capital and Bessemer Venture Partners to fund its AI-powered CAD software for construction.
  • Nearly 61 million U.S. consumers use AI as their main shopping research tool, with 60 percent completing purchases on Amazon despite rising AI-assisted spending.
  • Walmart is embracing AI shopping agents while Amazon has blocked its Muse platform, showing divergent strategies in the AI retail race.
  • T-Mobile and Ericsson are testing AI-RAN in New York City using AI-managed beamforming to improve 5G network performance and reduce latency.
  • Naver's Q3 operating profit is expected to drop 12.4 percent to 500 billion won due to AI infrastructure costs exceeding 200 billion won per quarter.
  • S-Agent 2.0 improved spatial reasoning for coding agents, boosting Qwen3.8-27B by up to 20.5 points and outperforming the specialized SpatialClaw system.
  • Researchers improved drone delivery route planning using reinforcement learning, achieving up to 3.2 percent improvement across seen and unseen problem sizes.

CoreWeave launches Zero Trust AI Framework with six security layers

CoreWeave introduced the Zero Trust AI Framework at NVIDIA GTC Berlin. The framework gives customers control over access to their models, data, and runtime environments. It uses six layers of protection including Zero Key Access, Zero Node Access, and Zero Trust Infrastructure. Customers can use their existing security systems instead of rebuilding around a provider-defined model. The framework also includes confidential computing support and network encryption through IEEE 802.1AE MACsec.

Cisco and Fortanix bring verifiable confidential AI to Kubernetes

Cisco, Fortanix, and VAST Data partnered to bring confidential AI to Kubernetes and data platforms. Fortanix Confidential AI now supports CNCF Confidential Containers. The system verifies containers before releasing encryption keys, using composite attestation for CPUs and GPUs. Keys stay in a FIPS 140-2 Level 3 HSM and are only released into verified environments. This allows regulated organizations to run AI on sensitive data while maintaining security.

Lambda shares NVIDIA collaboration plans at GTC Berlin 2026

Lambda announced several initiatives with NVIDIA at GTC Berlin 2026. The company is supporting Confidential Bare Metal with NVIDIA Confidential Computing for frontier model builders. Lambda is also advancing agent safety through the NVIDIA OpenShell Reference Architecture. Additionally, Lambda plans to deploy NVIDIA Minerva for infrastructure monitoring and is testing NVIDIA DSX MaxLPS to increase token throughput. Lambda is a Gold sponsor at the event and has a booth for demonstrations.

Patra hires Yousef Hashimi to lead AI product development

Patra appointed Yousef Hashimi as Head of Products and Client AI Transformation. Hashimi brings nearly 30 years of experience from roles at Lüt, IBM, and EY. He previously helped commercialize IBM Watson and led applied AI for insurance at EY. Patra CEO Pratap Sarker said Hashimi will ensure products tie directly to client outcomes. Hashimi will focus on expanding Patra's agentic AI services for the insurance industry.

Patra names Yousef Hashimi to expand AI managed services

Yousef Hashimi joined Patra to lead AI products and client transformation efforts. He previously served as Chief Product and Operating Officer at Lüt, where he launched three payment products. He also held senior roles at IBM and EY working on enterprise AI and automation. Patra is expanding its managed services model to combine agentic AI with insurance process expertise. The company's initial offering focuses on post-bind processing for insurance policies.

AI Shopping Growth Favors Amazon Over Direct Purchases

Nearly 61 million U.S. consumers now use AI as their main starting point for online shopping research. Average AI-assisted spending rose 17 percent between June and August surveys. However, 60 percent of respondents completed their most recent AI-assisted purchase on Amazon. The PYMNTS Intelligence study surveyed 2,191 U.S. consumers in August 2026.

Reinforcement Learning Improves Drone Delivery Route Planning

Researchers developed a method to help Pointer Networks handle drone delivery routes with varying numbers of clients. The study uses the Capacitated Vehicle Routing Problem as a test case for drone payload constraints. A learned approach selector chooses the best model for different training and test size pairs. The selector improved results by up to 3.2 percent across seen and unseen problem sizes.

AI Spending Weighs on Naver Quarterly Profits

Naver's third-quarter operating profit is expected to fall 12.4 percent to 500 billion won due to rising AI infrastructure costs. Infrastructure expenses including GPU depreciation have exceeded 200 billion won per quarter. Revenue is projected to rise 12 percent to 3.51 trillion won. AI service revenue growth remains slow, with new ad placements not expected to expand in earnest until 2027.

S-Agent 2.0 Boosts Spatial Reasoning With Coding Tools

S-Agent 2.0 adds a command-line perception toolkit to general coding agents for spatial reasoning tasks. The toolkit improved Qwen3.8-27B by 13.9 and 20.5 points on MMSI-Bench and ReVSI benchmarks. With Gemini-3.8-Flash, gains reached 11.2 and 9.6 points. S-Agent 2.0 outperformed the specialized SpatialClaw system on the same model backbones.

TikTok Users Post AI Deepfake Videos of Police Misconduct

Engagement farmers on TikTok are using AI to alter real arrest footage and show police officers kissing or groping arrestees. Copyleaks discovered the deepfakes, which were based on footage from reality TV shows like TLC's D.U.I. The manipulated videos were spliced in front of longer arrest footage to keep viewers watching. TikTok removed flagged examples after being contacted.

Mach9 lands $34M to power AI CAD software for construction

Mach9 Robotics changed its name to Mach9 and shifted to building AI-powered CAD software. The company reached nearly 100 customers and over 2,500 projects in its private beta. On Wednesday, it announced a $34 million Series A round led by First Round Capital and Bessemer Venture Partners. Prudence and Quiet Capital also joined the round. The software helps civil engineers turn large amounts of data into useful design and asset management information.

Walmart embraces AI shopping agents as Amazon blocks Muse

Walmart and Amazon are taking different paths in the AI shopping wars. Walmart is welcoming AI agents and investing heavily in AI technology. Amazon is taking a more cautious approach and blocking Muse. The moves show how the two retail giants are competing to stay ahead in the changing retail landscape.

T-Mobile and Ericsson test AI-RAN in New York City

T-Mobile and Ericsson are running live AI-RAN network trials in New York City. The tests use AI-managed beamforming on live network traffic. Beamforming focuses network energy to improve signal strength and reduce interference. The work aims to improve network performance and reduce latency. It marks an important step for 5G and AI in network management.

Sources

NOTE:

This news brief was generated using AI technology (including, but not limited to, Google Gemini API, Llama, Grok, and Mistral) from aggregated news articles, with minimal to no human editing/review. It is provided for informational purposes only and may contain inaccuracies or biases. This is not financial, investment, or professional advice. If you have any questions or concerns, please verify all information with the linked original articles in the Sources section below.

AI Products AI Technologies Zero Trust AI Framework Security Layers NVIDIA GTC Berlin Lambda Collaboration Confidential Computing AI Shopping Consumer Behavior AI Infrastructure AI-RAN Naver AI Costs AI Retail AI Agents AI Software AI Improvements

Comments

Loading...