Deutsche Telekom's Buyback Is Real Money — Its AI Targets Are Still Just Paper

Deutsche Telekom's buyback passed 25 million shares since August 10, but its AI targets and fiber orders still await the Q3 2026 results due November 5.

Deutsche Telekom kept its foot on the accelerator in its own stock this week, disclosing the purchase of 1,562,000 treasury shares on Monday as part of an ongoing repurchase program. The filing adds to a running total that has already climbed past the 25 million mark since the program's August 10 start, with 4,272,881 shares picked up between October 5 and October 9 alone.

That is a tangible corporate action, and it is the kind of signal that tends to draw attention. But a buyback answers a narrow question — whether management thinks the equity is cheap — and leaves the harder ones untouched. For shareholders weighing the stock today, the more consequential issue is whether the operating business can carry the confirmed guidance that underpins the whole investment case.

Two Sets of Numbers, Two Different Clocks

At its AI investor day on October 5, the Bonn-based group reaffirmed both its full-year outlook and its medium-term objectives. The headline figures attached to the strategy are substantial: roughly EUR 2.5 billion in indirect cost savings from artificial intelligence and automation by 2030, measured against a 2023 baseline, plus around EUR 250 million in AI-related enterprise revenue outside the US targeted for 2026, scaling to about EUR 800 million by 2030.

Those are long-dated ambitions, and the gap between announcing them and delivering them is where the investment debate actually lives. A confirmed forecast describes what management expects. It does not substitute for savings already banked or revenue already invoiced. Treating the full EUR 2.5 billion as though it were incremental profit today would be a category error, just as booking the EUR 800 million revenue goal as secured would be.

The efficiency lever is the one with the clearest price tag, which makes cost development the single most useful yardstick for tracking whether the strategy is working. Progress on the revenue side alone would be welcome but insufficient — the plan rests on two distinct tasks, building new business and making existing operations leaner, and only both together validate its economic logic.

Fiber Orders Fill the Pipeline, Not the Ledger

Complicating the picture is the capital that has to go out the door alongside the savings that are supposed to come in. On October 6, Deutsche Telekom announced orders from the real estate sector for fiber connections covering seven million housing units, backed by EUR 800 million in additional investment over the next three years to expand the network.

Those contracts represent demand for infrastructure, not cash already collected. They sit in the order book rather than the income statement, and the spending they require is precisely why the AI savings cannot be evaluated in isolation. What matters is how growth, outlay and efficiency fit together across the whole business. In an unfavorable scenario, the expected AI relief would surface more slowly than hoped while the fiber buildout proceeds on schedule — and a reaffirmed long-term target would not, on its own, be proof of anything better.

A Spectrum Deal Redraws the US Risk Map

The other variable is competitive, and it arrived with little warning. SpaceX's spectrum agreement, struck about a week ago, triggered concerns about fresh competition in the US mobile market, and Reuters identified the deal as the trigger for the share price losses at Deutsche Telekom. Its American subsidiary T-Mobile US sits squarely in the path of those worries.

Here the distinction between a risk and a result matters enormously. A spectrum purchase does not by itself establish any demonstrated harm to Deutsche Telekom's business; the exposure is to a possible future rival, not to a present earnings decline. Equally, it would be too glib to treat the buyback as a sufficient answer to that threat. Should the competitive landscape genuinely shift against T-Mobile US, the operating assumptions would need to be revisited — and repurchasing shares would not resolve uncertainty about the business outlook.

AlphaValue Trims, Keeps Its Call

Analysts are not treating the situation as a reason to retreat. AlphaValue/Baader Europe on Monday maintained its buy recommendation on the stock while nudging its price target down to EUR 34.90 from EUR 35.10 — a modest trim rather than a change of heart. That view is an assessment, not evidence that the company's objectives will be met.

What would strengthen the bullish reading is not a fresh round of targets but visible execution. Repeated indications of delivery would offer more orientation than restating long-term sums. If the next set of figures supports the confirmed guidance and shows traceable efficiency gains at the same time, the buyback can be read as a complement to a dependable operating performance rather than a substitute for one — and the case for the equity would rest on more than the company's appetite for its own shares.

November 5 Is the Next Real Test

Until then, investors are left choosing between two readings: confidence in the operating foundation beneath the repurchase, or greater caution toward potential competitive pressure. The tiebreaker arrives on November 5, when Deutsche Telekom publishes its financial results for Q3 2026. That report offers the next opportunity to compare confirmed expectations against reported numbers — and it will show how the current business is developing, even if it cannot settle whether the long-term AI ambitions will pay off.

The framework for interpreting whatever lands is straightforward. Forecasts describe the ambition; financial statements describe the achieved position. The closer those two converge, the more durable the investment thesis becomes. If execution lags expectations, even large long-term headline figures lose their persuasive force. The buyback has sent its signal — whether the business confirms its meaning is a question only the numbers can answer.