NVIDIA (NVDA) shares slipped on Friday as the company advanced its Nemotron 3.5 AI model expansion and a series of related developments spanning open-model releases, a Taiwan-backed Texas supply project, and financing pressure on a major cloud customer. The stock traded at $217.09 on October 16, down $0.14, or 0.06%, intraday.

Stock Moves Lower During Volatile Trading

NVIDIA shares moved between roughly $215.30 and $218.50 during a volatile Friday morning session. The price recovered from its earlier morning low but faded before the afternoon, and by the latest reading had returned to trade near $217 — slightly below the previous reference price.

NVIDIA trades on the Nasdaq under the ticker NVDA and supplies chips for AI computing systems used by business and research customers. Beyond hardware, its business spans software, networking equipment, and technology for large data centers. Its operations extend into the infrastructure companies rely on to develop and operate AI workloads. Because those chips anchor a large share of the industry's computing buildout, the company's share movements are widely followed as a read on sentiment across the AI sector.

The latest share movement followed a run of company and industry announcements. NVIDIA advanced its model portfolio, while other disclosures concerned manufacturing capacity and cloud financing. The reported developments do not, by themselves, establish why shares declined. Further updates on the model rollout, the Texas project, and CoreWeave's financing could add context in upcoming sessions.

Nemotron 3.5 Expansion Adds Open Model Options

NVIDIA expanded the Nemotron 3.5 lineup with several additional open models, targeting greater efficiency, stronger performance, and faster inference speed across the family. The additions complement the hardware that trains and runs AI systems, and they connect to the company's software and computing platforms. Open-weight releases have become a prominent strand of the AI industry, offering an alternative to closed models that are accessible only through a provider's own interfaces.

The company also partnered with Mindbeam to tune Nemotron 3.5 models for open-source applications. The work focuses on improving inference — the process through which trained models generate responses or predictions when deployed in applications. By tying model optimization directly to NVIDIA hardware, the partnership aligns open-source model development with the company's chip platform. Inference is also where deployed models serve end users, making its speed and cost a recurring focus as AI applications scale.

NVIDIA uses its CUDA software platform to help developers run computing tasks on compatible processors. Open models give developers access to model components for adaptation and deployment within different computing systems. Together, Nemotron development and the Mindbeam agreement add software activity to NVIDIA's established chip business — a pattern which hardware makers build software ecosystems that make their processors more useful to developers.

Texas Supply Project and CoreWeave Financing Pressure

Taiwan backed a $4 billion industrial park project in Texas connected with NVIDIA's American AI supply network. The project adds a manufacturing and supply-chain development alongside the company's software announcements, and it expands planned industrial activity in the United States. The announcement lands amid a broader U.S. push to expand domestic chip and AI manufacturing capacity, in which Taiwanese-linked industrial investment has played a growing role.

NVIDIA relies on a broad network of suppliers and customers across the AI infrastructure market. Large computing deployments combine processors, networking equipment, software, power infrastructure, and facilities, making supply projects an important support for data center construction and expansion across the United States.

Separately, CoreWeave, an important NVIDIA cloud customer, faces pressure linked to its debt obligations. CoreWeave, which went public on the Nasdaq in March 2025, provides cloud computing capacity for AI workloads and purchases infrastructure for its services. Debt-financed expansion is common among AI cloud providers, whose GPU purchases and data center buildouts require substantial upfront capital. The debt pressure has raised concerns about funding for AI cloud expansion, and it represents a separate financial issue within NVIDIA's wider AI customer network, distinct from the company's own model and manufacturing announcements.