June 12 (Bloomberg) -- CNBC reports that German startup DeepL announced Wednesday that it has enabledNvidiaThe latest DGX SuperPOD system, which enables theTranslate the entire Internet in just 18 days -- a much faster pace than the previous 194 days. DeepL has developed its own generative AI models in direct competition with Google Translate.
For its part, NVIDIA is working to bring AI chipThe customer base has expanded from "mega" tech companies such as Microsoft and Amazon to a wider range of usage scenarios.
According to 1AI, DeepL is headquartered in Cologne, Germany, and this deployment of the DGX SuperPOD system36 B200 Grace Blackwell Superchips per rack, is one of NVIDIA's latest flagship chip releases. These chips provide key underpinnings for companies like DeepL to train and run large-scale AI models.
In an interview with CNBC, Stefan Mesken, chief scientist at DeepL, said the company wants to leverage more computing power to push its R&D team to build more advanced models.
He mentioned that the upgraded hardware will enhance the capabilities of existing products such as Clarify, a new tool introduced this year that improves the quality of translations by asking users questions to more accurately understand the context.
Mesken says that in the past, such features were difficult to achieve due to the state of the art, "and now that it's finally possible as we move forward with the next generation of technology, we're constantly pursuing similar breakthroughs."