PayPal said Tuesday (Oct. 28) that it accelerated the time-to-market of its artificial intelligence solutions by five times after implementing Nvidia Nemotron open models and Nvidia NIM inference microservices.

Within weeks of implementing the Nvidia solutions, PayPal achieved a 50% increase in speed and a boost in developer productivity, according to a Tuesday press release.

“This dramatic improvement stems from two key technological advantages: the intuitive open architecture that streamlines development, and our ability to directly fine-tune models rather than relying on less efficient prompt engineering approaches,” PayPal said in the release.

PayPal is using Nvidia Nemotron open models to build large language model-powered agentic commerce experiences, according to the release.

By using open models, PayPal can fine-tune AI systems for its specific domains and gain “unprecedented control” over how it deploys AI to address business challenges, the release said.

With this control, PayPal can develop intelligent agentic commerce experiences, scale them globally, and maintain its trust and security, per the release.

Advertisement: Scroll to Continue

“The move toward open models isn’t just a technological update, it’s a strategic step toward gaining more control of our AI stack, from model selection and training to deployment and optimization—enabling us to deliver AI experiences with greater precision, lower latency and more reliability than ever before,” PayPal said in the release.

PYMNTS reported Tuesday that PayPal is turning to AI as its account growth and consumer spending slow.

For example, the company is integrating its wallet into OpenAI’s ChatGPT. PayPal CEO Alex Chriss said Tuesday during an earnings call that this partnership will “expand payments and commerce in ChatGPT, including adding PayPal Holdings Inc. branded checkout for shoppers and payment processing for merchants using instant checkout.”

PayPal also announced its own agentic commerce services, which Chriss said will help merchants sell through multiple AI platforms, including OpenAI, Google and Perplexity.

When Nvidia unveiled a family of large language models called Nvidia Llama Nemotron in January, Nvidia CEO Jensen Huang said they can help developers create and deploy AI agents across various applications.

“AI agents are the new, digital workforce working for and with us,” Huang said.

For all PYMNTS AI coverage, subscribe to the daily AI Newsletter.