What If DeepL Goes Public?

DeepL is reportedly aiming for a market capitalization of up to USD 5bn, a significant jump from its USD 2bn valuation following a USD 300m funding round in March 2025. The rumor mill has since been active on social media about the German AI unicorn eyeing an initial public offering (IPO) in the US.

Bloomberg called the language AI firm a “Google Translate Rival” when reporting about its potential listing (sources remaining anonymous, though). The comparisons do not end there: some in the mainstream media put the company as a rival to OpenAI and other AI hyperscalers — and not just in terms of translation. DeepL actually offers a suite of AI content and voice products, agents, apps, and integrations.

At the time of writing, DeepL is promising “big announcements” it will make in early November 2025. With all the buzz, we were curious about whether readers would buy DeepL shares after an IPO, and close to half of respondents said they would, of course (47.3%). A significant 34.5% voted “probably not,” and the rest (18.2%) were not sure.

For now, we will continue to monitor the regulator’s website for hints of an S-1 filing.

Bizarro World AI

A report prepared by consultancy Europe Economics for the UK’s Department for Energy Security & Net Zero (DESNZ) ignited criticism after proclaiming that AI translation is a thousand times more energy-efficient than human translation.

The report, which assesses how digital services and their supporting data centers affect UK electricity demand, included a controversial case study comparing cloud-based AI translation to a professional translator working in a physical office.

The authors estimated that AI uses less than 0.05 kWh for a 50,000-word document, while a human translator performing the same task in an office environment consumes between 29 kWh and over 121 kWh. The report concludes that such efficiencies demonstrate how digitalization can displace “energy-intensive physical activities” and help achieve net-zero goals.

Critics argue that the methodology is deeply flawed and based on an “asymmetrical” comparison. For example, the human scenario assigns a large share of an entire office building’s energy — for heating, lighting, and air conditioning — to a single translator over two weeks. 

The report also overlooks industry standards, where translators use technology for post-editing and can reach 1,000 words per hour or more, potentially halving the modeled energy use.

Questionable research methods aside, we asked readers if they are concerned about AI’s energy use, and an overwhelming majority (73.8%) are definitely so. Some respondents think more nuclear power plants are in order (15.4%), and about one in ten (10.8%) are not really worried. 

Perplexingly Unoriginal or Just Resourceful?

AI scale-up Perplexity followed OpenAI in signaling that even AI needs localization management by posting a job ad in October 2025 for a “Localization Program Manager.” The role is assigned an annual salary between USD 180k and 250k, to be based in San Francisco, just like OpenAI, in late August 2025.

And also just like OpenAI, the core responsibilities for the Perplexity role include piloting “AI-first translation workflows with human-in-the-loop review,” managing linguist vendors, and leading quality assurance processes. All in all, Perplexity’s job ad is a substantial calque of OpenAI’s ad for a Localization Manager.

Despite the highly similar language used in the two job postings, a key difference points to varying organizational views on localization: the Perplexity candidate will report to the Product organization, while the OpenAI role is housed under Marketing, the primary focus of its L10N manager role, at least initially.

We asked readers what they think about a company copy-pasting a job ad, and a little less than half (47.4%) are of the mind that HR should try a little harder. Two equal cohorts (26.3% each) think either why not, or that the AI is responsible for the calque.

AI Translation is a UI Feature

AI Translation-as-a-Feature (TaaF) has become a pervasive feature across numerous industries. In a typical week, news in the language industry is now dominated by a continuous stream of AI translation launches, the use cases and modalities highly diverse and increasingly specialized.

For example, a Michigan, US, police force integrated real-time AI translation into their body cams for communication with citizens, as more and more law enforcement institutions are doing.

Media channel Arab News launched a service during an October 2025 event in Madrid, Spain, to make its content available in 50 languages. Even the B2B travel technology sector is leveraging AI, with HBX Group deploying real-time AI translation via a customer service AI agent called Olivia.

Beyond TaaF, the widespread adoption of AI is attracting new players, evidenced by cases like that of global consulting firm Capgemini announcing its new, secure AI speech translation platform, and BabelSpeak, specifically positioned as sovereign AI for the Nordic region, to name but a couple.

Readers shared their views on whether the TaaF boom is a net positive or negative for the language industry. For over half (53.2%), it is a net negative. One in four (25.5%) take a neutral position on the matter, and the rest (21.3%) consider it a net positive development.