Deepl uses new NVIDIA chips to translate the entire Internet after 18 days

Deepl on Wednesday said he had deployed one of the latest Nvidia Systems that would allow the German startup to translate the entire Internet in just 18 days.

This has been sharply down from 194 days before.

Deepl is a startup that has developed its own AI models for and compete with Google Translate.

In the meantime, NVIDIA is looking to expand the customer base for its chips – which are designed to feed artificial intelligence applications – outside of hypersallers such as Microsoft and Amazon.

It also emphasizes how startup companies use high -end NVIDIA products to build AI applications that are viewed as the next step after founding models, such as those designed by Openai.

The Cologne -based company has a NVIDIA system known as the DGX SuperPod. Each of the DGX Superpod Server Racks contains 36 B200 Grace Blackwell Superchips, one of the company’s most products on the market. NVIDIA chips are required to train and manage huge AI models, such as those designed by Deepl.

“The idea, of course, is to provide a much more clarifying power to our research scientists to build even more advanced models,” Stefan Mesque, a major Deepl scientist, told CNBC.

Mesken said upgraded infrastructure will help improve current products such as Clisify that the company launches this year. Clarification is a tool that asks users questions to make sure that the context is included in the translation.

“This was simply not technically feasible until recently with the progress we have achieved in our efforts from the next generation. This has become possible. So these are the types of achievements that we continue to hunt,” Meskin said.