SK Telecom's 67% Profit Jump Shows AI May Be Saving a Telecom Story

SK Telecom's Q2 beat improved margins, not the top line

SK Telecom's second quarter looks good enough to attract buyers, but not clean enough to confirm a full reset. The company posted a 67.3% jump in operating profit to 566 billion won, while net profit reached 466 billion won. Revenue, by contrast, rose only 0.5%. That gap matters because it shows how much of the beat came from cost control, base-effect relief, and the early contribution of data-center activity rather than broad-based demand.

That context matters. Management cited tighter cost control, stronger data center activity, and a favorable base effect from one-off expenses booked a year earlier. The quarter is still solid, but the profit surge is not a pure read-through for the next year.

Bulls see room for rerating if AI data centers become a durable second engine. Bears see a higher bar: if that growth pillar slows, investors may find they are paying for more than a mature telecom just as the new business is still proving itself.

A further complication is timing. Before earnings, SK Telecom shares had surged as much as 15.86% intraday and finished the session up 12.30%, after Hana Securities raised its target price from 55,000 won to 80,000 won. So part of the good news was already in the stock. The real question now is whether AI infrastructure can scale quickly enough to support that optimism.

AI data centers are the main valuation pivot

The part of the business that could matter most is AI data center revenue of 136.2 billion won, up 92.5% year over year. That is the only segment with both the scale and the growth rate needed to change how investors think about the company.

Why AI infrastructure matters more than the core telecom recovery

Telecom operations can produce steady cash flow, but investors rarely pay high multiples for stability alone. AI data centers are different because they monetize compute capacity backed by power, site control, and connectivity. SK Telecom is backing that shift with SK Hyper, with up to 750 billion won planned to be injected into the subsidiary by 2030 and a combined capacity of 5 gigawatts under development.

That is why the buildout matters. If successful, SK Telecom is not just selling connectivity; it is trying to become a provider of essential AI infrastructure.

The earnings call suggested the expansion is meant to be phased and customer-led, with efforts to secure clients before construction and finance projects through external investors and project financing to reduce pressure on the parent company. If that approach works, the business can look more durable than a standard telecom story.

The software push is optionality, not the core profit engine

SK Telecom has also unveiled A.X K2, a 688-billion-parameter foundation model, and is identifying use cases across manufacturing, defense, and bio. That can improve long-term margins and customer stickiness, but it still looks more like optionality than a current contribution to earnings.

The practical takeaway is simple: the next rerating debate hinges on AI data center revenue, SK Hyper's progress, and whether the company can fund expansion without overstretching the core business.

What the market is pricing into SK Telecom now

SK Telecom is being viewed as a hybrid: it still offers a Q2 DPS of 831 won while it tries to build a combined AI data center capacity of 5 gigawatts. After the recent rally, that balance is the point of contention. If the core business keeps generating cash and the dividend remains intact, investors have more time to wait on the AI buildout. If not, the story gets harder to defend.

What has to keep working

The bull case rests on a few linked conditions:

• AI data center revenue continues to scale.
• Capacity development stays tied to demand rather than pure speculation.
• Shareholder returns remain stable while investment progresses.

What could go wrong

The biggest risk is that the market is pricing the AI transition too early. A 5-gigawatt buildout would require substantial capital, and the company still has to secure land, power, water, and large customers while meeting the standards of global buyers. If AI spending gets ahead of revenue, the dividend cushion that makes the story attractive could come under pressure.

What investors should watch next

The next major checkpoint is the next quarterly report. Beyond headline profit and revenue, the more useful signals are:

• AI data center revenue growth
• Progress on customer commitments and construction phases
• Any change in the pace or funding structure of the SK Hyper buildout
• Whether the dividend remains stable as investment needs grow

This is still a watchlist setup, not a fully proven growth story. One sentiment note: speculation about Anthropic-related upside and a state-led AI model project may draw attention, but those remain narrative drivers, not current earnings drivers.