← Back
NVIDIA Blog

Vera Delivers the Best of Both Worlds: Highest Throughput and Fastest Agents

•4 min read•
#agents#deployment
Level:Advanced
For:AI Engineers
✦TL;DR

NVIDIA Vera CPU delivers the highest throughput and fastest performance for AI agents, achieving up to 2× faster agent speed and 1.6× more agents per watt compared to the fastest x86 CPUs. Benchmarks show Vera outperforms both the fastest and most energy‑efficient x86 processors in a 100‑MW data‑center scenario, with partners reporting 1.5‑2.2× speed gains on real‑world agent workloads. This enables AI infrastructure providers to run more concurrent agents on a single CPU while staying within power limits, simplifying hardware stacks for production AI factories.

⚡ Key Takeaways

  • Vera provides 1.6× more agents per watt versus the fastest x86 CPU.
  • Architecture delivers 2× faster agent speed compared with the most energy‑efficient x86 CPU.
  • Partners report up to 2.2× speed improvement on agentic orchestration (DeepInfra) and 1.5× faster coding tasks (Perplexity).
  • Vera scales near‑linearly across 176 cores, topping the ClickBench hardware leaderboard with ClickHouse’s standard binary.
  • Deployment of Vera enables 1,000 agent sandboxes to launch in 10.8 s and sustain 12.7 completed jobs per second at 2,000 concurrent sandboxes (Daytona), far outpacing x86 (88.4 s launch).
💡 Why It Matters

Engineers can consolidate diverse agent workloads onto a single high‑performance, power‑efficient CPU, reducing hardware complexity and operational cost while delivering lower latency and higher concurrency in production AI systems.

✅ Practical Steps

  1. Evaluate NVIDIA Vera CPU specifications against current x86 servers to assess potential 1.6× agent‑per‑watt gains.
  2. Benchmark your agent workloads (e.g., coding, orchestration) on Vera using existing binaries to verify 1.5‑2.2× speed improvements.
  3. Deploy Vera in your AI factory and monitor sandbox launch times and throughput to achieve near‑linear scaling across cores.

Want the full story? Read the original article.

Read on NVIDIA Blog ↗

More like this

New in AWS Quick: More capable flows, an updated web experience, and agents that work together

AWS ML Blog•#agents

Three insights you may have missed from theCUBE’s coverage of the Dell Technologies New York Symposium

SiliconANGLE AI•#deployment

How Google Duplex evolved to also call businesses for you in Gemini

Google Research•#agents

Welcome Inkling 2 Mini, Cutting-edge Performance for your Local Setup

Hugging Face Blog•#llm

EXPLORE AI NEWS

Daily hand-picked stories on LLMs, RAG, agents and production AI — curated for engineers who ship.

BROWSE NEWS

GET THE WEEKLY DIGEST

Join engineers getting the Monday signal-over-noise AI breakdown. No spam, unsubscribe anytime.

LEARN AI ENGINEERING

Curated courses, research papers, repos and tutorials built for engineers leveling up in AI.

START LEARNING