Yorkville Ives Starts NVIDIA at Outperform, $300 Price Target

Yorkville Ives analyst Dan Ives initiates coverage on NVIDIA (NASDAQ: NVDA) with an Outperform rating and a price target of $300.

“We are initiating coverage of NVIDIA (NVDA) with an OUTPERFORM rating and $300 PT.

We view NVIDIA as the foundational compute platform for the AI economy, with its competitive advantage extending beyond GPUs to the architecture that defines how AI infrastructure is built and operated.

NVIDIA’s CUDA software ecosystem gives developers a common programming environment and extensive library of tools optimized for its hardware, creating an entrenched developer base that is harder to displace as applications enter production.

The agreed $12.9 billion acquisition of Hugging Face, announced September 3, extends that ecosystem to the place where developers discover, fine-tune, and deploy models.

Hugging Face hosts more than 3 million models, 500,000 datasets, and 1 million applications for over 18 million developers and roughly 200,000 enterprises, and NVIDIA was already its largest contributor of open models and datasets before the deal.

Management has committed that the platform will remain open and that NVIDIA compute will not be required because the value lies in proximity to the open-weight community whose models are increasingly optimized on NVIDIA infrastructure.

The transaction is expected to close in the first half of 2027 subject to regulatory approval, and is immaterial to the balance sheet against free cash flow approaching $200 billion this fiscal year.

Around that software foundation, the company has expanded from selling accelerators into complete AI-factory systems spanning GPUs, Vera CPUs, NVLink scale-up networking, Spectrum-X Ethernet, BlueField data processing, storage infrastructure, and rack-scale systems.

This matters as customers optimize around the economics of producing intelligence (cost per token, throughput, power consumption, and utilization) rather than individual chip specifications.

NVIDIA can improve those economics by designing the entire system together, while capturing a greater share of the infrastructure spending behind each deployment.

Vera Rubin represents the next step in that strategy, with the platform purpose-built for reasoning and agentic AI workloads that can require significantly more compute than traditional chatbot interactions.

NVIDIA has also extended Rubin with Groq 3 LPX for low-latency inference and introduced a Vera CPU designed for agentic systems.

We believe the transition from generative AI toward reasoning, autonomous agents, and physical AI meaningfully expands the amount of compute required for each user interaction rather than simply shifting demand from training to inference.

NVIDIA is positioned across both, while its presence in robotics through Jetson, Isaac, and Cosmos provides another path for the same accelerated-computing architecture to move beyond the data center.

In our view, the company’s advantage is increasingly defined by the breadth of the ecosystem surrounding the GPU: as AI workloads become more complex, customers can adopt more NVIDIA compute, networking, and software within the same architecture rather than assembling those layers independently.”

This research note is reproduced verbatim from the issuing firm. Price Target never edits, paraphrases or alters analysts’ words — we only republish them in one place.