SK Telecom has announced plans to unveil a robot-learning platform in the second half of 2026, with full commercialisation targeted for 2027. The move signals the company's formal entry into the physical AI market — the fast-growing field concerned with giving robots the ability to perceive and act in the real world — and reflects a strategic pivot away from a core telecoms business that is running out of room to grow.
From connectivity to robotics: the logic of escape
South Korea's mobile market has reached saturation. Growth in 5G subscribers has slowed every year since 2022, according to the Ministry of Science and ICT, and average revenue per user (ARPU) has stagnated for several years running. All three of the country's major carriers — SK Telecom, KT, and LG Uplus — are hunting for revenue beyond their legacy businesses. SK Telecom's answer is robot AI.
The platform it is developing would serve as training infrastructure for robots: collecting, processing, and feeding data to machines so they can learn to perform real-world tasks autonomously. Think of it as a boot camp for robot intelligence. The intended business model is business-to-business, supplying robotic solutions across manufacturing, logistics, healthcare, and services.
A market worth trillions
SK Telecom's timing is deliberate. Goldman Sachs forecasts the global humanoid-robot market will reach roughly $154 billion by 2035. At this year's CES, Nvidia's chief executive Jensen Huang declared physical AI "the next revolution." Tesla's Optimus robot, Figure AI, and Boston Dynamics are all absorbing vast investment. The scramble is on.
What underpins this race is the recognition that training data and simulation environments are as strategically important as the robots themselves. Google DeepMind has released Open X-Embodiment, an open dataset for robot learning. Nvidia has pushed its Isaac Sim simulation platform to the fore. SK Telecom's planned platform is a bid to claim a slice of this foundational layer of the emerging robotics ecosystem.
Strengths, and the gaps that could prove fatal
SK Telecom's potential advantages are real. Its nationwide 5G and AI infrastructure, substantial data-processing capacity, and a portfolio of technology subsidiaries — including SAPEON, an AI chip start-up — could generate meaningful synergies. Its ability to transmit data in real time via its telecoms network is particularly relevant to "cloud robotics," a model in which robots learn and receive instructions remotely rather than processing everything on-board.
The weaknesses are equally plain. The company has no hardware manufacturing capability, which limits its role to that of a platform provider. Its technology gap with Google, Nvidia, and Microsoft is wide. And South Korea's domestic market alone is too small to generate the economies of scale that platform businesses require. Industry observers are blunt: "The success of a platform business ultimately depends on how many robot manufacturers and industry partners you can attract into your ecosystem."
A crowded domestic field
SK Telecom is not alone in this pursuit. KT is developing remote robot-control technology in partnership with Hyundai Motor Group. LG Uplus is expanding its robot-integrated services platform. Samsung Electronics has reinforced its dedicated robotics division and is targeting both consumer and industrial markets. Hyundai, which acquired Boston Dynamics, has already established itself as a global player.
The rush of South Korean conglomerates into robotics AI is intensifying competition for scarce talent. The Korea Institute for Robot Industry Advancement projects that demand for specialist robotics engineers will significantly outstrip supply over the next five years — making talent acquisition as critical a variable as platform development.
Can the timetable hold?
Industry experts are cautiously sceptical about the timelines. "A robot-learning platform is not a straightforward software project," said one AI robotics researcher. "It requires complex integration between the physical and digital worlds. The 2026 commercialisation target could shift substantially depending on how quickly partnerships and data can be secured."
Precedents are sobering. SoftBank's attempt to build a platform ecosystem around its Pepper robot ended in failure. Amazon Robotics, by contrast, succeeded by staying tightly focused on logistics. The lesson for SK Telecom is pointed: the company must decide whether to pursue a broad, general-purpose platform across all industries, or concentrate on a specific vertical where it can establish defensible depth. Trying to be everything to everyone has been a reliable recipe for achieving nothing in particular.
Winner takes most
In platform markets, early movers who establish a standard tend to dominate for decades. The parallel with smartphones is instructive: Google's Android and Apple's iOS effectively divided the mobile ecosystem between them, and no serious challenger emerged. A similar dynamic is likely to unfold in robot-learning platforms, with a small number of players capturing the bulk of the market.
SK Telecom's message in launching this initiative is unambiguous: it intends to combine its telecoms infrastructure with AI and robotics to construct an entirely new business. The declaration is bold. Whether execution can match it — and whether SK Telecom can offer something sufficiently differentiated from the global technology giants already in the field — will determine whether this bet pays off.