AI-based language translation startup DeepL has spent millions of dollars it has raised on hiring, growing its headcount by about seven times in three years, in a bid to maintain its edge over well-funded tech giants.
The German startup, which recently raised $300 million at a $2 billion valuation, is pioneering AI technology that's helping to create a world where language barriers are less of a problem. The company announced Wednesday that it's adding 165 countries to the market for its translation products, aimed at businesses in Asia, Africa, Europe and the Americas.
However, Apple, Amazon, Google, Microsoft, and Meta are all investing in similar technologies and dominating the consumer market. Products such as mobile phones, earphones, and smart speakers could become the translation interfaces of the future. As hardware and software improve, real-time translation on these devices could soon become a reality.
In an interview with Semafor, Jarek Kutylowski, co-founder and CEO of DeepL, said the only answer is to continually improve and get faster.
“The key is speed of innovation,” he said. “We have been able to convince our customers of our superior quality.”
Kutzyrowski said the company is dedicating up to 100 of its roughly 1,000 employees to research, working to improve and expand DeepL's machine translation capabilities. “That's a pretty significant part of what we're doing given our size, maturity and revenue,” he said. The company's clients include Nikkei, Deutsche Bahn and Zendesk.
Katharina Wilhelm, a partner at Index Ventures, which led DeepL's most recent investment round, said that while consumers love DeepL, the real value is with businesses. “You're probably not going to be willing to pay $50 a month for a personal translation,” she said. “But if you're a journalist or a business person, you might pay because accuracy and security are really important to you.”
Kutyrowski declined to say whether any major tech companies have approached DeepL about partnering or buying its technology, and he said DeepL has no plans to develop its own hardware devices to compete with those companies.
Index Ventures is betting on a big win, not a sale to a big tech company. “Of course, if there was a very attractive acquisition offer, we'd consider it,” Wilhelm said. “We wouldn't have invested if it was just a short-term acquisition by a big platform. This is going to be a generational company.”
Kutilowski said the best translation models are too large to run on-device and therefore require access to the cloud, meaning hardware devices that perform real-time translation are still a long way away.
Rather than focusing on the consumer market, DeepL has turned its attention to the corporate market, selling seven-figure accounts to large corporations such as international law firms and businesses in need of diplomatic translations.
For example, one client, which he declined to name, is a Japanese automaker that has its research and development division in Japan and its sales division in the U.S. “Communication between business divisions is extremely important, and if it doesn't work well, a lot of problems can occur,” he said.
The company has also branched out into literature, providing its products to publishers: “We have a book in our office that was translated using DeepL,” he says, “and we also have a book on machine learning neural networks.”
