Deutsche Telekom targets €2.5 billion in AI savings by 2030
The figures were presented at the company's first investor day dedicated to AI, in Bonn. But the headline number is gross — token costs are not deducted — and Telekom is not saying how many jobs the automation will cost.
Deutsche Telekom held its first-ever investor day dedicated to artificial intelligence on Monday, October 5, 2026, in Bonn. There, the group promised around €2.5 billion in indirect-cost savings by 2030, measured against 2023, as a result of AI and automation (press release). At the same time, the company confirmed its 2026 outlook and its medium-term targets (Reuters).
The message to the market is twofold: AI is meant not only to cut costs, but also to generate new revenue — with so-called sovereign AI infrastructure for European business as the selling point.
The numbers the company presented
The core of the message consists of three metrics, all from the company itself:
- Around €2.5 billion in indirect-cost savings by 2030, measured against 2023, as a result of AI and automation.
- Around €1.1 billion in gross savings outside the US by 2027. This figure needs to be read carefully: according to Trending Topics, the €1.1 billion is a gross figure that "does not account for the cost of using AI itself, the so-called token costs." The net saving is therefore unknown.
- Revenue targets: Around €250 million in AI-related revenue from business customers outside the US by 2026, with a goal of around €800 million by 2030 (press release).
Unresolved discrepancy between sources
There is a significant unresolved disagreement about the 2027 figure. While Reuters and Trending Topics report around €1.1 billion in gross savings outside the US by 2027, dpa writes that Telekom "expects AI and automation savings of €100 to 150 million between 2023 and 2027," and that part of the savings is to be reinvested in digital transformation and fiber network expansion in Germany. None of the sources explain how the figures are to be reconciled — whether the dpa figure is net of token costs, or a separate interim figure. Readers should therefore treat the 2027 level as unresolved.
What the company claims AI has already delivered
The press release should throughout be understood as the company's own claims, not independently verified results:
- The chatbot Frag Magenta is said to have handled around 2.6 million customer service inquiries in the first half of 2026 alone.
- In the US, customer service inquiries are said to have fallen by 55 percent, and AI agents are said to now handle 40 percent of customer contacts.
- The RAN Guardian Agent, a network monitoring agent, is said to have cut response times for network incidents from several hours to around one minute.
These figures are part of the sales pitch: the purpose of the investor day was, as the company itself put it in the press release, to show how artificial intelligence makes a measurable contribution to the business.
Höttges' strategic framing
CEO Tim Höttges framed AI as a sweeping transformation: "AI is fundamentally changing Deutsche Telekom. It makes our networks even better and smarter, our service more personal, and opens up new business models: from AI assistants during phone calls to a sovereign infrastructure for European industry," he said in the press release.
The quote points to the second half of the strategy: beyond cost cuts, Telekom intends to sell AI services to business customers, with "sovereign infrastructure" — that is, European-controlled AI infrastructure — as a contrast to American hyperscalers. The details of what this infrastructure concretely involves are not specified in the available sources.
What the figures do not contain, and the questions that remain
Several caveats should accompany the metrics:
- Gross, not net. The €2.5 billion figure is a savings target from the company, not a verified result. Token costs — that is, what the AI usage itself costs — are excluded from the 2027 figure, according to Trending Topics. The net financial effect is therefore unknown.
- Unresolved discrepancy. As described above, €1.1 billion (gross, outside the US, by 2027) stands against dpa's information of €100–150 million in the period 2023–2027, without any source explaining the relationship.
- The jobs. Telekom has not stated how many jobs the automation will cost. At the same time, subsidiary T-Mobile US is cutting around 4,700 positions in a parallel restructuring — but the company attributes this to leaner structures and the removal of duplicated roles, not to AI specifically. The connection between the automation program and staffing therefore remains unresolved.
The investor day also confirmed the 2026 outlook and the medium-term targets — a signal that the AI message should be read as an addition to, not a change in, the financial course.
What is expected now
The concrete milestones are 2026 (around €250 million in AI revenue outside the US), 2027 (the savings figure the sources describe differently) and 2030 (€2.5 billion in savings and €800 million in revenue). Whether the figures hold can only be assessed when the company reports results — and when it potentially quantifies the employment impact, something it has so far kept out of the accounts.