Deutsche Telekom Expects AI to Save €1.1 Billion by 2027
Deutsche Telekom stated that artificial intelligence and automation are expected to generate approximately €1.1 billion in gross savings outside the U.S. by 2027, using 2023 as the baseline. By 2030, indirect cost savings are estimated to be around €2.5 billion.
This €1.1 billion does not account for the token costs associated with using AI itself. The company plans to reinvest part of the additional savings from 2027 back into Germany's digital infrastructure and fiber networks. CEO Tim Hoettges mentioned before an investor event on Monday that AI is fundamentally transforming the company, making networks smarter, services more personalized, and opening up new business opportunities.
Savings will occur in areas with repetitive tasks: network operations, customer service, software development, and management processes. The company cited examples where AI can detect traffic peaks in cellular networks earlier and assist customer service representatives to resolve issues on the spot. Additionally, they are developing a platform for businesses to automate recurring tasks. Voice assistants will be able to translate calls in real-time, answer questions, and summarize conversations, with plans to expand services and pricing options based on this.
Revenue targets are calculated separately. By 2026, AI-related revenue outside the U.S. is expected to be around €250 million, rising to about €800 million by 2030. The company views the development of sovereign AI in Europe as a major business opportunity, as businesses and public institutions want to use AI without relinquishing control of their data. Full-year guidance and mid-term targets remain unchanged.
The U.S. business is not included in these savings figures. Deutsche Telekom holds significant assets in the U.S. through T-Mobile, while the cost and revenue estimates pertain to its European operations. The event was held in Bonn for AI Investor Day, with the statement released in advance.
This is a cost guidance, not a layoff or contract announcement. Buyers are European companies and public institutions seeking sovereign computing power and automated processes, while the seller is Deutsche Telekom, which is transferring customer service, network, and management processes to models. Savings shift from manual repetitive operations to token expenditures and fiber reinvestment. The beneficiaries are the groups that can capture gross savings and reinvest part back into the network, while those under pressure are customer service and process roles covered by automation, as well as profit margins not yet offset by token costs.
Source: Public Information
ABAB AI Insight
Tim Hoettges joined the Deutsche Telekom board in 2006, took charge of European operations in 2009, and became CEO in 2014. Under his leadership, T-Mobile in the U.S. was transformed from a loss-making asset into a major value for the group, while continuing to expand fiber networks in Europe. In November 2025, Deutsche Telekom announced a partnership with Nvidia worth approximately $1.15 billion to build a large AI factory in Europe. This Investor Day connects the same line to savings and revenue figures.
The operational costs outside the U.S. are being addressed. The €1.1 billion gross savings uses 2023 as the baseline, with indirect cost savings of €2.5 billion by 2030, but gross savings do not include token costs. Part of the savings from 2027 is earmarked for reinvestment in Germany's fiber and digital infrastructure. On the revenue side, AI-related revenue outside the U.S. is projected to be around €250 million in 2026 and about €800 million by 2030, relying on business task platforms and real-time translation assistants sold to clients unwilling to relinquish data in Europe.
In comparison, Vodafone uses AI for network energy savings and customer service, while Orange integrates sovereign cloud into public sector orders. All three face competition from U.S. or global cloud providers in front of enterprise clients. Deutsche Telekom is in a cost-harvesting phase of expansion: automating networks and customer service first, while sovereign computing factories are still turning Nvidia hardware into sellable European capacity. T-Mobile's exclusion from the savings figures indicates that the group is not mixing the U.S. wireless cost story into Europe.
This represents technological substitution. The roles being replaced are repetitive tasks in customer service, work orders, and management, not spectrum and fiber assets. The mechanism is that models first take over scriptable processes, the saved labor is partially consumed by token costs, and the remaining gross amount returns to network capital expenditures. Sovereign AI turns data export restrictions into a reason for European telecoms to charge. Pricing power shifts from call minutes to who can run models locally and prove savings.
ABAB News · Cognitive Law
Gross savings do not account for tokens; net savings contribute to profits.
Repetitive processes are first taken over by models, while assets remain on the books.
Data export restrictions create a pricing structure for local operators.