Deutsche Telekom's AI Push Cuts Both Ways: 500 Projects, a 2028 Chip Bet and a Staff Strain Warning

Deutsche Telekom has placed artificial intelligence at the center of its equity story, but the case for owning the stock now rests less on how many projects the group can announce than on whether it can convert them into durable operating gains without exhausting the workforce that has to deliver them.

Chief executive Timotheus Höttges on Thursday named AI and Elon Musk as the company's biggest challenges, according to Reuters. At the same time, employee representatives are flagging serious strain from the pace of change — a combination that puts execution, not ambition, at the heart of the investment debate.

The scale is hard to ignore. More than 500 AI projects are running across Telekom's European operations, and Höttges says he personally spends 30% of his time on such initiatives. That signals management attention in abundance. It does not, on its own, prove economic success.

A 2028 Chip Horizon and a Bank Deal on the Application Side

The group's infrastructure ambitions gained a concrete anchor this week. T-Systems, its enterprise IT arm, agreed a partnership with Dresden-based AI firm Sparxity to develop energy-efficient AI computing hardware for Telekom data centers, with the next chip generation slated for introduction in 2028. Telekom frames the tie-up as a boost to European AI infrastructure, with energy efficiency as the focal point.

That timeline matters for how investors should read the announcement. The partnership is a real step; the chip rollout remains a future project. A development agreement is not the same as booked business, and the two should be kept apart.

T-Systems is working the application layer in parallel. Under an agreement dated October 28, Banco Sabadell plans to expand agent-based AI, drawing among other things on T-Systems' "T-AI Agentic Hub." The two deals address different links in the same chain — computing hardware for data centers on one side, AI agents in customer use on the other — and together sketch a strategy spanning both the provision and the application of AI in the enterprise business. Neither, however, demonstrates commercial success for the new chip generation. They remain separate ventures.

The Net Effect Is the Real Test

The decisive question is the net economic impact of all this AI activity. Telekom has announced billions in savings from AI, according to media reports, while already cutting a large number of jobs in Germany this year. Labor representatives criticize the sheer pace, which they say has produced extreme pressure situations.

For shareholders, that criticism carries financial weight. If the strain hampers delivery, expected benefits could arrive later or prove smaller. A workable restructuring, by contrast, would show that staff changes and new workflows can be reconciled. Both are scenarios, not demonstrated outcomes.

Announced savings therefore should not be treated as realized earnings contributions. More telling would be tangible implementation progress and signals of how Telekom intends to limit the burden on its employees. The question is not how many applications emerge, but whether they permanently improve the way the business runs.

What the Bull Case Requires

The favorable path assumes AI genuinely makes tasks more efficient and the organization absorbs the change. On that reading, the announced savings could underpin earnings power, and the CEO's deep engagement would be more than a strategic gesture — it could help drive initiatives into daily operations.

A large project count could also turn into an advantage if successful applications are rolled out more widely. That does not mean every project will deliver a measurable contribution. What would count is whether Telekom can point to robust results and make their economic significance understandable.

Investors would not need a further expansion of the project tally for this scenario to hold. A shift from many individual initiatives to demonstrable improvements would be more convincing — evidence that the AI strategy is economically viable rather than merely expectation-generating.

Musk, Boost and a Second Risk Layer

The bear case starts with execution. The warnings from employee representatives are a signal that the overhaul could hit organizational limits. If new processes add to staff workloads instead of relieving them, the expected benefits could shrink, and the headline savings figure alone would not offset that risk.

An external challenge may be building at the same time. Höttges considers a Musk entry into the mobile market conceivable. According to a Washington Post report picked up by Reuters, SpaceX wants to acquire Boost or use its network. Höttges sees a global communications platform as a possible goal for Musk.

No such competitive move should be treated as fact. Still, speculation about Musk's interest has already weighed on shares of Verizon, AT&T and T-Mobile US. For Telekom, that adds a valuation risk: internal efficiency gains could convince less if worries about a shifting competitive landscape intensify at the same time.

November 5 Is the Next Hard Checkpoint

JPMorgan reaffirmed its "Overweight" rating on Deutsche Telekom with a price target of 38 euros on Thursday. The call predates the Sparxity announcement and should not be read as an assessment of the new partnership.

As long as Telekom can turn its AI efforts into economically viable improvements, the efficiency scenario stays plausible. If workload problems harden, investors would need to judge implementation speed more cautiously. Should Musk's mobile plans take firmer shape on top of that, a second layer of uncertainty would come into play.

The next concrete checkpoint is the third-quarter 2026 results, expected on November 5. Investors can watch for commentary on AI deployment, savings and the restructuring. No such detail has been confirmed so far. What would matter is whether management makes the link between strategic ambition and economic delivery more tangible.