SK Group chairman Chey Tae-won wants the conglomerate to move beyond chips, infrastructure and energy into AI models and services. The proposed “AI flywheel” would connect its businesses, but the examples so far are mostly applications of AI within existing operations, not evidence that SK has already become a leading AI-services provider.
Watch Desk analysis
What happened
At a seminar for subsidiary leaders in Icheon, Chey called on SK companies to redefine themselves as AI businesses and find roles in what he described as a “token economy”, where AI computing and services are traded as resources. The Korea Times reports that he also called for cooperation with global partners, saying SK could not do everything alone.
The report identifies several existing or developing applications across the group. SK hynix is upgrading factories that combine AI and digital-twin technology; SK Innovation uses equipment and operating data to improve yields; and SK Telecom applies AI to network and service automation. SK Gas uses AI to analyse market conditions and turn operational knowledge into reusable assets.
Why it matters
This is a broad strategic push from a conglomerate that already spans chips, telecoms, energy and infrastructure. Chey’s argument is that those businesses can reinforce one another: chip capabilities support AI factories, while energy and services help sustain demand and investment. If that links up, SK’s AI ambitions would reach beyond supplying the hardware behind other companies’ models.
The immediate evidence is more grounded than the grand strategy. The reported examples concern factory efficiency, network automation, yields and market analysis. Turning those operational uses into connected, saleable AI services is the bigger ambition, and the report does not say that transformation has happened yet.
Our read
The interesting question is whether SK can make its sprawling portfolio work as one AI business, rather than simply add AI labels to several existing ones. The smart-factory and operations examples give the strategy something tangible to build on. Now comes the less glamorous test: whether the connections create useful services and returns, not just a tidy diagram of a flywheel.
What to watch
- Whether SK announces specific AI models or services, rather than broader ambitions.
- How the group connects its chip, factory, telecoms and energy businesses in practice.
- Whether future updates show measurable results from the cited operational AI uses.
Discussion spark: Should a conglomerate such as SK build AI services around the businesses it already has, or focus its investment on becoming a model developer?
Sources and evidence
- Source update (15 November 2026, 03:46 UTC)
Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.