JAMS Brings AI Agents to Enterprise Job Scheduling — and Lets You Choose Your Own Model

Why it matters: For 37 years, JAMS has been the quiet backbone of enterprise IT automation — scheduling batch jobs, orchestrating data pipelines, keeping the lights on. Now they’re adding an AI agent called JAX that lets ops teams ask plain-language questions like “why did that payroll job fail last night?” — and get an answer grounded in actual system logs and documentation.

The smart move: JAMS isn’t locking customers into one AI provider. You can point JAX at OpenAI, Anthropic, or a local model running entirely inside your own network. Every action runs as the signed-in user — no elevated AI account, no data leaving your perimeter unless you want it to. Write actions require explicit human approval.

For your business: This is the blueprint for practical AI agent deployment in regulated industries. If you run finance, healthcare, or manufacturing operations with scheduled automation, JAX shows how to layer AI on top of existing infrastructure without ripping anything out — and without sacrificing security or audit trails. Available at no extra cost to existing JAMS Web customers.

📎 Read the full announcement →

Audi Deploys Multi-Agent AI System to Power Its B2B Used Car Platform

Comsysto Reply built a multi-agent AI system for Audi’s B2B used car platform — a collaboration dating back to 2021. The system coordinates multiple specialized AI agents handling different parts of the vehicle resale workflow. It’s a real-world example of agent orchestration moving beyond chatbots into complex, multi-step business processes at enterprise scale. Bottom line: If a legacy manufacturer like Audi is betting on multi-agent architectures, the pattern is going mainstream.

📎 Read more →

The Next AI Agent Breach Won’t Come From the Model — It’ll Come From the API Path

F5 security expert Vishnu Gatla warns that the real threat to AI agents isn’t a “rogue model” scenario — it’s the mundane reality of over-permissioned API access. When an AI agent becomes an API client with a valid token, your SIEM, WAF, and identity systems all see legitimate traffic. Gatla’s prescription: assign every production agent a clear identity, scope access aggressively, and make agent traffic distinguishable from human traffic. If you’re deploying AI agents that touch APIs, this is your Monday morning reading.

📎 Read the full analysis →

AMD Drops $5B on Anthropic and Launches Helios — The AI Chip War Just Went Nuclear

The news: At its Advancing AI 2026 event in San Francisco, AMD dropped a series of announcements that fundamentally reshape the AI infrastructure landscape. The headline: a $5 billion investment in Anthropic, paired with Anthropic’s commitment to buy up to 2 gigawatts of AMD’s upcoming Instinct MI450 chips — with the first gigawatt deployed by H1 2027.

AMD also unveiled Helios, its first fully integrated rack-scale AI system: liquid-cooled, 72 Instinct MI455X GPUs, 18 EPYC “Venice” CPUs, delivering ~2.9 exaflops of AI inference at $5M–$5.5M per rack. Helios claims up to 30% higher inference output per dollar versus comparable Nvidia racks. OpenAI, Meta, and Oracle are already lined up as customers. Microsoft will deploy Helios in Azure starting H2 2026.

The bigger picture: AMD also announced a partnership with Cerebras to combine Helios racks with Cerebras’ Wafer-Scale Engine, targeting 5x better tokens-per-second-per-watt for ultra-low-latency inference — the kind needed for real-time coding copilots and agentic workflows. Separately, AMD launched its Robotics Partner Network and the ROCm.ai software stack, which promises 3.3x inference improvements over the previous version.

What it means for you: Nvidia’s CUDA moat is still real, but it’s shrinking. AMD now has marquee AI customers, competitive hardware, and a credible software story. For businesses evaluating AI infrastructure, a real second source for AI compute means pricing leverage, supply chain resilience, and optionality. The duopoly is here — and it’s good for everyone except Nvidia’s margins.

📎 AMD-Anthropic deal details →  |  Forbes analysis →  |  CRN full event recap →

Intel Posts Strongest Revenue Growth in 15 Years — AI Compute Demand Is the Engine

Intel reported Q2 2026 revenue of $16.1 billion, up 25% year-over-year — its best growth in over 15 years. The Data Center and AI segment surged 59% to $6.3 billion. Intel Foundry revenue jumped 31%. CEO Lip-Bu Tan credited AI demand for lifting the entire portfolio, from CPUs to ASICs to advanced packaging. Q3 guidance came in above expectations. While Nvidia and AMD grab headlines, Intel’s turnaround is quietly gaining momentum — and the AI compute tide is lifting all boats.

📎 Full earnings breakdown →

Uber Cuts 10% of Customer Service Staff in Second AI-Linked Layoff in Two Months

Uber laid off roughly 10% of its customer service workforce this week, explicitly linking the cuts to AI restructuring. This follows June’s 23% reduction in the People and Places division. A VP memo said the organization had become “too complex and siloed” and needed streamlining before expanding AI further. Uber joins Oracle (25,000+ positions eliminated), Monday.com (~20% of staff), and Amazon in a growing list of companies where AI adoption is directly reshaping headcount. Cognizant now estimates 93% of jobs could be affected by AI — and says the pace of change arrived six years earlier than expected.

📎 Read the full story →