📊 All Intelligence Feed

Anthropic Vendor Strategy May 20, 2026

Anthropic Engages Diverse Wisdom Traditions to Explore AI Moral Formation

Anthropic initiates a long-term research project, engaging in dialogues with scholars, clergy, and ethicists from over 15 religious, philosophical, and cultural groups. The aim is to draw on diverse human wisdom to inform the moral formation of AI systems like Claude and the development of its 'constitution'.
Impact: ImportantStrength: MediumConf: 85%AI Infra
NVIDIA Ecosystem Restructuring May 20, 2026

NVIDIA and Google Cloud Deepen Developer Ecosystem Integration, Advancing AI Infrastructure and Application Stack

NVIDIA and Google Cloud's joint developer community surpasses 100k members, offering full-stack learning paths from JAX optimization and NVIDIA Dynamo inference tuning to AI watermarking (SynthID). This move aims to accelerate enterprise AI application deployment from prototype to production by integrating underlying hardware (Blackwell/Rubin GPU), cloud platforms (GKE, AI Hypercomputer), and software frameworks (Nemotron, Gemma).
Impact: ImportantStrength: HighConf: 85%AI Infra
Microsoft Architecture Shift May 19, 2026

Microsoft Launches New Surface for Business Line, Emphasizing On-Device AI and Security Integration

Microsoft introduces new Surface Pro and Laptop for Business models with Intel Core Ultra Series 3 and upcoming Snapdragon X2 processors. Key focus is on-device AI inference, security-by-design, and full-stack Microsoft management. Devices serve as reference hardware for Windows AI APIs and the Foundry platform, positioning Surface as the hardware foundation for enterprise hybrid AI strategies.
Impact: ImportantStrength: HighConf: 85%AI Infra
Google Vendor Strategy May 19, 2026

Google Public Sector Showcases Blueprint for AI Agent Deployment at Scale

Google Public Sector outlines its strategy for driving government agencies from AI pilots to full-scale 'agentic' transformation, using case studies from the U.S. DOT, FDA, and City of Los Angeles. The approach centers on an integrated AI stack and emphasizes leadership, scale, and human-centered adoption.
Impact: ImportantStrength: HighConf: 85%AI Infra
Anthropic Architecture Shift May 19, 2026

Anthropic and KPMG Form Global Alliance, Embedding Claude into Core Business Platform

KPMG and Anthropic have formed a global strategic alliance, embedding Claude into KPMG's core business platform, Digital Gateway, and providing access to over 276,000 employees worldwide. The alliance will co-develop AI products for industries like private equity and apply Claude to critical business areas such as cybersecurity vulnerability detection.
Impact: ImportantStrength: HighConf: 90%AI Infra
NVIDIA Architecture Shift May 19, 2026

NVIDIA and Dell Launch Full-Stack AI Factory for Enterprise Agentic AI Deployment

NVIDIA and Dell have deepened their partnership, launching an updated Dell AI Factory with NVIDIA to provide an end-to-end platform for enterprise Agentic AI inference and deployment, from workstations to data centers. The platform integrates NVIDIA Vera Rubin GPUs, Vera CPUs, Confidential Computing, and Nemotron models, emphasizing secure, high-performance on-premises AI infrastructure to meet surging inference demand.
Impact: MajorStrength: HighConf: 90%AI Infra
Amazon Architecture Shift May 19, 2026

AWS Deepens AI Agent and Multicloud Integration, Strengthening Enterprise Modernization and Security

AWS announced multiple updates, highlighting the native integration of Claude Platform into AWS accounts, the launch of more powerful EC2 M3 Ultra Mac instances, and the expansion of AWS Transform AI agent modernization service to platforms like Kiro and Claude. Additionally, AWS Security Agent added full repository code scanning, and AWS Interconnect extended multicloud connectivity to Oracle Cloud Infrastructure.
Impact: ImportantStrength: HighConf: 85%AI Infra
Google Architecture Shift May 19, 2026

Google Launches Antigravity Platform to Accelerate AI Agent Development and Deployment

At I/O 2026, Google launched the Antigravity 2.0 desktop app and ecosystem, platformizing AI agent development. It integrates a Managed Agents API, aiming to eliminate infrastructure friction from AI app ideation to production deployment.
Impact: MajorStrength: HighConf: 85%AI Infra
Google Architecture Shift May 19, 2026

Google Launches Gemini 3.5 Series, Defining New Agent-Centric AI Infrastructure Paradigm

Google launches the Gemini 3.5 model series, starting with 3.5 Flash, which is positioned as an "agent-first" engine. Combined with the Antigravity platform, it is designed to handle enterprise-scale, long-horizon, multi-step workflows, signaling AI's shift from a tool to a productive system for executing complex tasks.
Impact: MajorStrength: HighConf: 85%AI Infra
Cloudflare Architecture Shift May 19, 2026

Cloudflare Partners with Anthropic to Provide Cloud-Native Execution for Claude Agents

Cloudflare partners with Anthropic to decouple the execution layer (“hands”) of Claude Managed Agents from the reasoning layer (“brain”) and integrate it into the Cloudflare Developer Platform. This enables enterprises to securely run AI agent code and tools at scale within Cloudflare's sandbox, VPC, and proxy network.
Impact: ImportantStrength: HighConf: 85%AI Infra
Microsoft Architecture Shift May 18, 2026

Microsoft Open Sources Conductor: Deterministic AI Agent Orchestration with Zero Token Cost

Microsoft introduced Conductor at the Open Source Summit, an open-source orchestration tool for multi-agent AI workflows. Its key feature is defining workflows in YAML for deterministic routing between agents, using Jinja2 templates for conditional logic, with the orchestration layer consuming zero LLM tokens.
Impact: ImportantStrength: HighConf: 85%AI Infra
Google Architecture Shift May 18, 2026

Google Outlines Five-Layer Architecture for Evolving Enterprise Data to AI Agents

Google's technical blog outlines five data architecture evolution scenarios, from static APIs to autonomous workflows based on the Model Context Protocol (MCP), aiming to build an "agentic data layer" for enterprises. This signals a shift in data access patterns from manual development to AI-driven, standardized dynamic interactions.
Impact: ImportantStrength: HighConf: 85%AI Infra
Google Architecture Shift May 18, 2026

Google Shares Methodology for Large-Scale A/B Experimentation on Data Center Infrastructure

Google details its four-pillar methodology for conducting large-scale A/B experimentation at the data center infrastructure level, covering machine-level testing, balanced setups, binary hermeticity, and performance metrics, aiming to safely validate system-wide micro-optimizations.
Impact: ImportantStrength: HighConf: 90%AI Infra
Cisco Architecture Shift May 15, 2026

Cisco Partners with SūmerSports to Deploy AI Inference Infrastructure On-Premises

Cisco, via its AI POD solution, partnered with sports analytics platform SūmerSports to deploy a complete on-premises AI infrastructure within an NFL team. This move addresses the industry's core concerns over data sovereignty, low latency, and integration complexity by bringing AI inference capabilities directly to where the data resides.
Impact: ImportantStrength: HighConf: 85%AI Infra
Google Vendor Strategy May 15, 2026

Google Drives Multimodal AI Agent Ecosystem via Developer Challenge

Google announced the results of its Gemini Live Agent Challenge, showcasing next-gen multimodal AI agent applications built on the Gemini Live API and Agent Development Kit. Winning projects span surgical assistance, hardware control, and desktop navigation, highlighting Google's strategy to accelerate the shift from text-based to real-time, multimodal AI interaction through its developer ecosystem.
Impact: ImportantStrength: HighConf: 85%AI Infra
Anthropic Architecture Shift May 15, 2026

PwC and Anthropic Deepen Alliance to Build Enterprise AI Agentic Operating Models with Claude

PwC and Anthropic expanded their strategic alliance, integrating Claude across PwC's global operations. The partnership establishes a joint Center of Excellence, trains tens of thousands of consultants, and focuses on building 'AI-native' agentic technology, deal execution, and enterprise function reinvention using Claude Code and Cowork. This signals a shift from AI pilots to scaled production deployment by major consultancies.
Impact: ImportantStrength: HighConf: 85%AI Infra
Amazon Architecture Shift May 15, 2026

Amazon Bedrock Launches Advanced Prompt Optimization and Model Migration Tool

Amazon introduces an advanced prompt optimization tool within Bedrock, enabling users to automatically optimize prompts through a metric-driven feedback loop and test/migrate across up to 5 models simultaneously. It integrates multiple evaluation methods including Lambda functions, LLM-as-a-Judge, and natural language steering criteria.
Impact: ImportantStrength: HighConf: 90%AI Infra
Cisco Architecture Shift May 14, 2026

Cisco Advocates for Service Providers to Transform Edge Infrastructure into AI Service Platform

Cisco outlines a new edge opportunity for service providers driven by AI workloads, which involves leveraging their large-scale, distributed network infrastructure to deliver enterprise services including AI inference and localized data processing. The Cisco Unified Edge platform is designed to address the challenges of automated, consistent management across thousands of sites.
Impact: ImportantStrength: HighConf: 85%AI Infra
Google Architecture Shift May 14, 2026

Google Introduces Application Design Center, Shifting Compliance & Governance Left

At Cloud Next '26, Google Cloud introduced Application Design Center and enhanced App Hub/Topology. These capabilities embed compliance and governance guardrails into development via architectural templates, Terraform generation, and a unified semantic graph, shifting control points left to address the operational bottleneck of AI-accelerated development.
Impact: ImportantStrength: HighConf: 85%AI Infra
Microsoft Architecture Shift May 14, 2026

Microsoft Strengthens Windows Platform Control via Driver Quality Initiative

Microsoft launched the Driver Quality Initiative at WinHEC 2026, aiming to systematically improve driver reliability, security, and performance through four pillars: architecture, trust, lifecycle, and quality measures. This move signals Microsoft's intent to tighten technical governance and control over the Windows hardware ecosystem to enhance end-user experience.
Impact: ImportantStrength: HighConf: 85%AI Infra
Cisco Vendor Strategy May 14, 2026

Cisco Announces Strategic Restructuring and Layoffs, Focusing Investments in Silicon, Optics, Security, and AI

Following strong Q3 FY26 earnings, Cisco announced a workforce reduction of approximately 4,000 roles. The company simultaneously signaled a clear strategic pivot, directing investments towards silicon, optics, security, and internal AI adoption. This move reflects difficult choices to optimize cost structure and concentrate on areas of long-term value creation amidst intensifying competition in the AI era.
Impact: ImportantStrength: HighConf: 90%AI Infra
Cloudflare Technology Integration May 14, 2026

Cloudflare Optimizes ClickHouse Partitioning, Reveals Hidden Bottlenecks in Massive-Scale Data Architecture

Cloudflare addressed a critical performance degradation in its billing pipeline caused by a partitioning change in its petabyte-scale ClickHouse analytics platform. Through deep performance profiling, they identified lock contention and vector copying bottlenecks in the query planner. The company contributed three key optimization patches upstream, significantly improving query performance in high-concurrency, high-partition-count scenarios.
Impact: ImportantStrength: MediumConf: 95%AI Infra
NVIDIA Architecture Shift May 13, 2026

NVIDIA and Ineffable Intelligence Co-Design Reinforcement Learning Infrastructure

NVIDIA has entered an engineering-level collaboration with Ineffable Intelligence, founded by AlphaGo architect David Silver, to co-design infrastructure for large-scale reinforcement learning (RL). The partnership will explore RL training pipelines on the Grace Blackwell platform and plan for the upcoming Vera Rubin platform, addressing RL's unique demands on interconnect, memory bandwidth, and real-time serving.
Impact: ImportantStrength: HighConf: 85%AI Infra
NVIDIA Architecture Shift May 13, 2026

NVIDIA Advances On-Device AI Agent Infrastructure with Hermes and Qwen 3.6

NVIDIA promotes the open-source AI agent framework Hermes from Nous Research and optimizes it with Alibaba's Qwen 3.6 models, aiming to establish a reliable, on-device AI agent runtime centered on RTX PCs and DGX Spark. This extends the deployment frontier of high-performance AI agents from the cloud to the enterprise edge and personal devices.
Impact: ImportantStrength: HighConf: 85%AI Infra
Microsoft Architecture Shift May 13, 2026

Microsoft and SAP Deepen AI Integration with "Microsoft IQ" Intelligence Layer and Cross-System Agent Collaboration

Microsoft and SAP announced a deepened partnership, introducing "Microsoft IQ" as a shared intelligence layer for enterprise AI and enabling agent-to-agent integration between Microsoft Copilot and SAP Joule. This move aims to deeply embed AI into core business processes and build a unified data foundation, signaling a shift of enterprise AI from the application layer to the core operational architecture layer.
Impact: MajorStrength: HighConf: 85%AI Infra
Amazon Architecture Shift May 13, 2026

AWS Launches Graviton-based Redshift RG Instances with Integrated Data Lake Query Engine

AWS introduces the Amazon Redshift RG instance family powered by its in-house Graviton processors, delivering up to 2.4x performance gains and 30% lower cost. The instances feature an integrated data lake query engine, unifying analytics across data warehouses and S3 data lakes, while eliminating Spectrum scanning fees.
Impact: ImportantStrength: HighConf: 95%AI Infra
Cloudflare Architecture Shift May 13, 2026

Cloudflare Migrates Browser Run to Containers, Boosting AI Agent Web Interaction

Cloudflare has migrated its Browser Run service from shared Browser Isolation infrastructure to its own Cloudflare Containers platform, achieving performance gains and scalability. This move optimizes the experience for AI Agents interacting with the web and demonstrates its 'Customer Zero' strategy of driving platform evolution through internal product use.
Impact: ImportantStrength: HighConf: 90%AI Infra
Cisco Architecture Shift May 12, 2026

Cisco and Red Hat Deepen AI Infrastructure Integration for Core-to-Edge Intelligent Platform

Cisco showcased deep integration with Red Hat's ecosystem at Red Hat Summit, covering AI PODs, Unified Edge, Network-as-Code, and Secure AI Factory. By embedding Ansible, Splunk, and Isovalent's eBPF capabilities into the OpenShift platform, it aims to provide enterprises with a unified, programmable, and secure AI infrastructure control plane from core to edge.
Impact: ImportantStrength: HighConf: 85%AI Infra
AMD Product Launch May 12, 2026

AMD Unveils Spartan UltraScale+ FPGA, Emphasizing Cost Optimization and Supply Chain Stability

AMD launches the Spartan UltraScale+ FPGA series, targeting the cost-optimized market. By comparing with Intel's Agilex 3, it highlights advantages in performance per watt, package size, and long-term supply assurance. The product aims to meet edge application needs in industrial and machine vision sectors.
Impact: ImportantStrength: MediumConf: 85%AI Infra
HPE Architecture Shift May 12, 2026

HPE Consolidates Private Cloud and Data Platforms for AI Data Readiness

HPE announced updates to its GreenLake platform, aiming to help enterprises modernize infrastructure and accelerate AI data readiness through unified private cloud, storage, and data protection solutions. Key actions include integrating Kubernetes management, unifying file and object storage, and introducing agentic AI capabilities across storage and data protection products.
Impact: ImportantStrength: MediumConf: 80%AI Infra
NVIDIA Architecture Shift May 12, 2026

NVIDIA and SAP Embed OpenShell into Business AI Platform, Providing Runtime Security for AI Agents.

NVIDIA and SAP have deepened their collaboration by embedding NVIDIA's open-source AI Agent runtime security framework, OpenShell, into the SAP Business AI Platform. This serves as a secure execution layer for all AI Agents, aiming to address trust and governance challenges in enterprise deployment through infrastructure-level isolation, policy enforcement, and audit trails.
Impact: ImportantStrength: HighConf: 85%AI Infra
Microsoft Architecture Shift May 12, 2026

Microsoft Unveils Copilot Design System, Defining AI-First Product Interaction Paradigm

Microsoft publicly details its Copilot Design System, aiming to build a unified, human-centric product interaction and behavior model for the AI-first era. Through core architectural elements like the Dynamic Action Button, Chat, and On-Canvas integration, it enables seamless, context-aware collaboration across applications, transforming AI from a tool into a thought partner.
Impact: ImportantStrength: HighConf: 85%AI Infra
Google Technology Integration May 12, 2026

Google Cloud G4 VMs Power Imgix's Real-Time Image Processing Performance Leap

Google Cloud's G4 VM instances, powered by NVIDIA Blackwell GPUs within its AI Hypercomputer infrastructure, enabled Imgix's image processing platform to achieve a 50% reduction in median latency and a 6x increase in throughput per node without core application code changes. This demonstrates the transformative impact of cloud-based AI inference infrastructure on real-time media processing workloads.
Impact: ImportantStrength: HighConf: 85%AI Infra
Microsoft Architecture Shift May 12, 2026

Microsoft and Red Hat Deepen Azure OpenShift Integration for Enterprise AI Production and Platform Modernization

At Red Hat Summit, Microsoft and Red Hat showcased Azure Red Hat OpenShift (ARO) as a unified platform for enterprise AI production. By integrating Azure identity, security, and governance services, ARO enables large institutions like Banco Bradesco to transition over 200 AI pilot projects into production systems, meeting stringent regulatory requirements.
Impact: ImportantStrength: HighConf: 85%AI Infra
AMD Technology Integration May 12, 2026

AMD Partners with Tsinghua on Open-Source Multi-Agent AI Education, Showcasing Edge-Cloud Deployment

AMD collaborates with Tsinghua's OpenMAIC team to deploy a multi-agent interactive AI classroom framework on its ROCm software stack. The solution uses Instinct GPUs for cloud-based course generation and Ryzen AI PCs with the Lemonade local server for real-time, low-latency classroom interaction, demonstrating an edge-cloud architecture on a unified software stack.
Impact: ImportantStrength: HighConf: 85%AI Infra
Amazon Architecture Shift May 12, 2026

AWS Launches AgentCore Payments and Agent Toolkit, Advancing Autonomous AI Agent Operations

AWS previews AgentCore payments, enabling AI agents to autonomously access and pay for APIs, MCP servers, and other services. It also launches the Agent Toolkit for AWS, a production-ready suite for AI coding agents, and makes the AWS MCP Server generally available.
Impact: ImportantStrength: HighConf: 85%AI Infra
Microsoft Architecture Shift May 12, 2026

Microsoft and BNY Demonstrate AI-Driven Enterprise Organizational Reshaping

The Microsoft-BNY collaboration case reveals how a large financial institution is reshaping workflows and organizational structure through 'digital employees' and an AI platform. BNY has established a comprehensive AI system covering governance, training, and operations, with its 'diamond-shaped' organizational model signaling AI's evolution from a tool to a core productivity architecture.
Impact: ImportantStrength: HighConf: 90%AI Infra
Google Architecture Shift May 12, 2026

Google Launches Googlebook, Defining AI-Native PC as a New Category

Google announced Googlebook, a new category of laptops designed from the ground up for Gemini AI. It merges Android and ChromeOS, featuring a 'Magic Pointer' for contextual AI suggestions and AI-generated widgets, aiming to deeply integrate AI into the user workflow and seamlessly connect with the Android ecosystem.
Impact: ImportantStrength: HighConf: 85%AI Infra
Microsoft Architecture Shift May 12, 2026

Microsoft Copilot Studio Update: Enhanced AI Agent Governance and Intelligent Workflows

Microsoft's Copilot Studio updates focus on strengthening centralized governance, cost visibility, and intelligent workflow capabilities for AI agents. Features like the Agent 365 control plane, agent nodes within workflows, and business app integration aim to transform isolated automation into trusted, scalable intelligent systems.
Impact: ImportantStrength: HighConf: 85%AI Infra
Google Vendor Strategy May 11, 2026

Google Public Sector Outlines AI Infrastructure, Data, and Security Architecture for the Agentic Era

Google Public Sector argues that moving from AI pilots to organization-wide agentic transformation requires a resilient, scalable, and secure foundation. Its architecture centers on three pillars: AI Hypercomputer, the agentic data cloud, and agentic defense, emphasizing high-performance hardware, AI-native data architecture, and the integration of Wiz's Cloud and AI Security Platform.
Impact: ImportantStrength: HighConf: 85%AI Infra
AMD Technology Integration May 08, 2026

AMD EPYC CPUs Gain Support in AWS RDS for SQL Server, Boosting Cloud Database Price-Performance

AWS has introduced instance options powered by 5th Gen AMD EPYC processors for Amazon RDS for SQL Server. This move provides a new, cost-effective compute choice for mission-critical database workloads and may shift the price-performance baseline for relational databases in the cloud.
Impact: ImportantStrength: HighConf: 85%AI Infra
Google Vendor Strategy May 08, 2026

Google Launches Gemini CLI DevOps Extension to Control Cloud Deployment Flow via AI Agents

Google launched the Gemini CLI DevOps Extension, enabling developers to use natural language commands to complete the entire process from code analysis and security checks to deployment on Google Cloud via AI agents (supporting Gemini CLI, Claude Code, Antigravity). The tool aims to bridge the efficiency gap between local development and production deployment.
Impact: ImportantStrength: HighConf: 85%AI Infra
NVIDIA Architecture Shift May 08, 2026

NVIDIA Collaborates with Slurm to Optimize GB200 NVL72 Cluster Scheduling for Rack-Scale AI Compute

NVIDIA, in collaboration with the Slurm community, introduced the topology/block scheduling plugin for GB200 NVL72 rack-scale GPU clusters. This approach treats NVLink domains as hard scheduling boundaries, using parameters like `--segment` to fine-tune job placement to mitigate severe performance drops across domains. It signals a shift in AI infrastructure orchestration from network optimization to compute-domain awareness.
Impact: ImportantStrength: HighConf: 90%AI Infra
Cloudflare Vendor Strategy May 08, 2026

Cloudflare Reorganizes with Layoffs, Reimagining Operations for Agentic AI Era

Cloudflare announced a global workforce reduction of over 1,100 employees. The core driver is a 600% surge in internal AI usage over the past three months, with thousands of employees relying on AI agents daily. The company is fundamentally reimagining all internal processes, teams, and roles to adapt to and lead in the agentic AI era, not as a cost-cutting measure.
Impact: ImportantStrength: HighConf: 85%AI Infra
NVIDIA Vendor Strategy May 08, 2026

NVIDIA Deepens AI-for-Science Partnership with US DOE on Genesis Mission

NVIDIA and the U.S. Department of Energy outlined the Genesis Mission at the SCSP AI+ Expo, applying AI to scientific discovery. The partnership includes building two AI supercomputers at Argonne National Lab and developing specialized AI agents to accelerate research in energy, materials, and grid optimization.
Impact: ImportantStrength: HighConf: 90%AI Infra
Microsoft Architecture Shift May 08, 2026

Microsoft Integrates GPT 5.5 Instant into M365 Copilot, Accelerating Multi-Model Platform Strategy

Microsoft CEO announced the integration of OpenAI's GPT 5.5 Instant model into Microsoft 365 Copilot for faster responses. This marks a shift for Copilot from a single-model assistant to a backend platform supporting multiple models like OpenAI and Anthropic, pushing model choice down to the user and task level.
Impact: ImportantStrength: HighConf: 90%AI Infra
NVIDIA Technology Integration May 08, 2026

NVIDIA Adds Prometheus Real-Time Monitoring to NCCL, Enhancing AI Training Observability

NVIDIA's NCCL 2.30 introduces Prometheus mode, converting GPU-to-GPU communication metrics into time-series data. This enables AI training teams to monitor and debug distributed training performance issues in real-time via Grafana dashboards, particularly for bottlenecks in mixed network and NVLink communication scenarios.
Impact: ImportantStrength: HighConf: 90%AI Infra
AMD Architecture Shift May 07, 2026

AMD Proposes Agentic AI Driving Separation of CPU and GPU Architecture in Data Centers

AMD SVP Dan McNamara states in an official blog that Agentic AI is fundamentally altering data center infrastructure architecture. It's not just about adding more CPUs to GPU servers, but necessitates building a separate, dedicated CPU compute layer for orchestration and tool execution, forming a distributed system alongside the dense GPU compute layer.
Impact: MajorStrength: HighConf: 85%AI Infra
ARM Architecture Shift May 07, 2026

Arm Reports Record Results, AGI CPU Emerges as New AI Infrastructure Focal Point

Arm reported record FY2026 results with $4.92B revenue and over 20% growth for three consecutive years. The core highlight is the Arm AGI CPU designed for agentic AI, securing over $2B in customer demand and backing from Meta, AWS, Google, and others.
Impact: MajorStrength: HighConf: 90%AI Infra
AMD Vendor Strategy May 07, 2026

AMD Backs SPEC CPU 2026 Benchmark, Emphasizing Open, Trusted Performance Measurement

AMD published a blog endorsing the upcoming SPEC CPU 2026 industry benchmark, emphasizing the critical role of open, reproducible CPU performance standards for customer infrastructure decisions in the AI era. The new benchmark updates its application suite and strengthens support for bare-metal cloud environments and parallel computing.
Impact: ImportantStrength: MediumConf: 85%AI Infra