Nvidia's AI Systems Design Chip Circuits, Verify Designs, and Test New Layouts
Nvidia chief scientist Bill Dally described the company's use of AI across five stages of chip design at GTC 2025, including NVCell (a RL+genetic algorithm system that redesigns ~2,500-3,000 layout cells overnight vs. 10 engineer-months), PrefixRL (RL-designed arithmetic circuits 20-30% better than human designs), and ChipNeMo/BugNeMo (LLaMA 2-based LLMs fine-tuned on internal GPU documentation). The systems demonstrate measurable improvements over human and industry-standard designs, though Dally acknowledged that fully autonomous GPU design from a prompt remains a distant goal. The piece also references a 2025 Verkoran paper describing an agentic system that autonomously designed a RISC-V CPU from a 219-word specification.
Related guides (4)
Related events (8)
How NVIDIA Engineers and Researchers Build with Codex
OpenAI published a case study describing how NVIDIA teams use Codex powered by GPT-5.5 to ship production systems and accelerate research experimentation. The piece highlights enterprise adoption of Codex as a coding agent in a major hardware/AI lab context. It signals continued real-world deployment of OpenAI's agentic coding tools at scale.
Data Points: Perplexity Computer expands, Google Aletheia math agent, DeepSeek chip strategy, Nvidia retrieval pipeline, Stargate cancellation
The Batch's weekly data points roundup covers five significant AI developments: Perplexity expanded its Computer agentic platform to desktop, mobile, and enterprise with new APIs and financial data tools; Google released Aletheia, a Gemini-based math research agent achieving 95.1% on IMO-Proof Bench Advanced (up from 65.7%); DeepSeek withheld pre-release access to its V4 model from Nvidia and AMD while giving domestic Chinese chipmakers early access; Nvidia's NeMo Retriever topped the ViDoRe v3 leaderboard using a ReACT-based agentic retrieval loop; and OpenAI and Oracle cancelled plans to expand the Abilene Stargate campus from 1.2 GW to 2.0 GW due to financing and reliability issues.
NVIDIA Cosmos 3, Nemotron 3 Ultra, and RTX Spark
A Latent Space AI news digest covers three NVIDIA announcements: Cosmos 3 (a world model/simulation platform), Nemotron 3 Ultra (a large language model), and RTX Spark (likely a new hardware or inference product). The piece frames these as a significant win for Jensen Huang and NVIDIA's AI portfolio. Coverage is commentary-tier aggregation rather than primary technical reporting.
Data Points: Nvidia Ising Models for Quantum Computing, Meta Muse Spark, GitHub Rubber Duck, Anthropic Claude Managed Agents, GPT-5.4-Cyber
Nvidia released Ising, a family of open AI models targeting quantum processor calibration and error correction, achieving 2.5x faster and 3x more accurate decoding than pyMatching, with adoption by Fermilab, Harvard, and others. Meta announced Muse Spark, a small multimodal model powering a new AI assistant series for its apps and glasses. GitHub introduced Rubber Duck, a cross-model review feature pairing Claude with GPT-5.4 for two-pass coding agent validation. Anthropic launched Claude Managed Agents, a managed infrastructure platform for enterprise autonomous AI deployment, while OpenAI expanded its Trusted Access for Cyber program with GPT-5.4-Cyber, a fine-tuned defensive cybersecurity model.
NVIDIA brings agents to life with DGX Spark and Reachy Mini
NVIDIA is integrating its DGX Spark computing platform with the Reachy Mini robot to enable embodied AI agents. The collaboration, highlighted on the Hugging Face blog, demonstrates running agent workloads on edge hardware for robotics applications. This represents a convergence of NVIDIA's inference infrastructure with open robotics platforms.
Custom CUDA Kernels for All from Codex and Claude
A Hugging Face blog post describes using AI coding agents (Codex and Claude) to automatically generate custom CUDA kernels, lowering the barrier to GPU kernel development. The piece demonstrates agent-assisted GPU programming as a practical workflow for ML practitioners. This represents a concrete application of AI coding tools to the specialized domain of CUDA/GPU optimization.
NVIDIA's GTC 2025 Announcement for Physical AI Developers: New Open Models and Datasets
NVIDIA announced new open models and datasets for physical AI development at GTC 2025, covered via the Hugging Face blog. The release targets robotics and embodied AI developers with open-weights resources. This represents NVIDIA's continued push into the physical AI ecosystem alongside its hardware dominance.
OpenAI and Broadcom Announce Strategic Collaboration to Deploy 10 GW of OpenAI-Designed AI Accelerators
OpenAI and Broadcom have announced a multi-year strategic partnership targeting deployment of 10 gigawatts of OpenAI-designed AI accelerators by 2029. The collaboration involves co-developing next-generation AI accelerator systems and Ethernet networking solutions aimed at scalable, energy-efficient AI infrastructure. This represents OpenAI's continued push into custom silicon, reducing dependence on third-party chip suppliers like NVIDIA.