AIToday
Large Language ModelsAI Business & IndustrySiliconANGLE AIPublished: Nov 18, 2026, 06:01 JST

ScaleOps: AI agents crash 80% more often than microservices

Send on LINE
ScaleOps: AI agents crash 80% more often than microservices

3 Key Points

  1. What happened

    ScaleOps Inc. founder and CEO Yodar Shafrir said at KubeCon + CloudNativeCon NA that agents crash due to insufficient resources 80% more often than traditional microservices, and that ScaleOps adjusts allocations in real time.

  2. Why it matters

    Static resource settings either waste money or fail under agent spikes, so autonomous orchestration that responds in real time is presented as the workable model for keeping agent workloads reliable.

Not sure about something? Ask the AI

Questions and answers are published on this page.

Summaries like this, every morning.

Context & Analysis

Shafrir's argument at KubeCon + CloudNativeCon NA hinges on how agents differ from traditional microservices. An agent waits for a human or another agent, checks with a large language model, then runs its task in a burst, so teams must provision for the highest spike while the agent holds its memory the whole time. When resources fall short, a crashed step has to rerun and burns tokens all over again.

ScaleOps' response is to make resource decisions in real time, adjusting allocations, replicas and node placement for every application, agent and Model Context Protocol server. The company has also added a capability that lets other workloads use an agent's resources while it sits idle, and it uses memory swap so an agent gets memory only when it needs it.

Shafrir frames the shift as a break from past practice: humans and traditional scaling capabilities are too static and require manual configuration for today's dynamic workloads. Adobe Inc. and Salesforce Inc. are cited as customers that build trust by first running the platform in read-only mode on their hardest environments.

FAQ
How much does ScaleOps say its platform cuts cloud costs?
ScaleOps says it cuts cloud costs by 70% to 80% by adjusting allocations, replicas and node placement in real time.
How does ScaleOps make running agents more cost-efficient?
It lets other workloads use an agent's resources while it sits idle and uses memory swap so an agent gets memory only when needed, which the company says makes running agents 90% more cost-efficient without moving off Kubernetes.
Which customers are named?
Adobe Inc. and Salesforce Inc. are named as customers that first run the platform in read-only mode on their hardest environments.
SiliconANGLE AIRead Original Article
Send on LINE

AI news that matters for your work, delivered every morning.

Pick your industry and the AI tools you use, and get news related to your work every day.

Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →

Ask AI

Ask AI anything about this article. The AI reads this article, earlier AIToday articles, and Wikipedia, and cites its sources. Q&As are published on this page for other readers too.

Questions and answers are published on this page.

Related Articles

Next articleJensen Huang exit seen shifting HBM demand beyond Nvidia