Hark announces multi-year NVIDIA collaboration aimed at gigawatt-scale support for personalized agentic AI

Hark said it has entered a multi-year strategic partnership with NVIDIA to advance what it describes as personalized agentic AI, a type of software that can pursue goals on a user or organization’s behalf rather than only answering questions. The company also said the effort includes a “deep technical collaboration” with NVIDIA, along with access to “gigawatt-scale compute capacity,” a phrase the announcement uses to announcement very large power and infrastructure requirements for next-generation AI workloads.

The announcement, published Aug. 27 and carried by Yahoo Finance, frames the work as a step toward more tailored AI assistance. In this context, “personalized” generally refers to using information about a specific user, organization, or environment to shape outputs and actions, while “agentic” implies the system is designed to take multi-step actions toward objectives.

Beyond the partnership headline, the posting does not provide additional specifics about the technical stack, model architectures, or which NVIDIA platforms will be involved. It also does not detail the expected timeline for milestones, any target performance benchmarks, or whether the compute capacity is intended for training, inference, or both.

Hark did not disclose any financial terms in the announcement as presented in the Yahoo Finance post. That includes the size of any planned spend, the structure of any capacity arrangements, or whether the collaboration includes funding, resale, or long-term service commitments tied to compute availability.

The announcement also does not specify whether the gigawatt-scale capacity is being pursued through owned infrastructure, co-located sites, third-party data center partners, or NVIDIA-linked capacity programs. Because those details are not included in the information surfaced by the market outlet, investors and industry watchers will likely need to wait for further documentation or additional statements from one or both companies.

NVIDIA’s role, as described here, is both technical and capacity-oriented, suggesting a partnership model that goes beyond selling chips. Over the past year, the broader industry trend has been toward arrangements that pair hardware availability with software and systems integration to reduce time-to-deployment for advanced AI applications.

Still, key elements remain unclear. The announcement does not name specific products, software frameworks, or deployment targets, and it does not say how the compute capacity will be allocated among Hark’s use cases. It also does not include any public performance claims or third-party validation that could help gauge the expected capability gains from “agentic” and personalized designs.

What to watch next is whether Hark provides additional technical details, such as which NVIDIA computing platforms and enterprise software components will be used, and whether the company offers clearer milestones for prototype releases or customer deployments. For NVIDIA, the next step is likely more detail on how this collaboration fits into its broader data center and AI infrastructure strategy, particularly around very large-scale compute planning.