Second-quarter results for Korea's major information technology (IT) service corporations were mixed. Hyundai AutoEver beat market expectations on the back of strong enterprise resource planning (ERP) and cloud migration businesses, while Samsung SDS and LG CNS saw limited profit improvement due to factors such as increased upfront investment and project deferrals, respectively.
Progress in the artificial intelligence (AI) business, a future growth driver, showed a different trend from current results. Samsung SDS is crystallizing orders and operational outcomes centered on graphics processing unit (GPU) subscription services (GPUaaS) and AI data centers, and LG CNS plans to fully expand its external physical AI business starting next year after group-affiliate projects and proofs this year. In contrast, Hyundai AutoEver's role, business scale, and monetization timeline within Hyundai Motor Group's physical AI strategy remain undefined.
According to the industry on the 3rd, Hyundai AutoEver posted second-quarter revenue of 1.2507 trillion won and operating profit of 90.5 billion won. Both rose 20% and 11.3%, respectively, from a year earlier, beating market expectations. Results were led by the enterprise IT business. Revenue in this institutional sector was 1.0384 trillion won, up 27.9% year over year. The increase was driven by rising cloud demand and a higher share of high value-added projects such as building next-generation enterprise resource planning (ERP) for group affiliates. About 20 billion won of enterprise IT revenue carried over from the first quarter was also reflected.
Samsung SDS recorded second-quarter revenue of 3.7178 trillion won and operating profit of 231.8 billion won. Those increased 5.9% and 0.7%, respectively, from a year earlier. Cloud revenue rose 17% to 779.4 billion won, but operating profit growth was modest due to upfront investments to expand new external businesses in the public and financial sectors and intensifying competition. The operating margin of the IT service institutional sector also fell 0.6 percentage points from a year earlier.
LG CNS posted second-quarter revenue of 1.5208 trillion won, up 4.2%, but operating profit fell 9.2% to 127.9 billion won, missing market expectations. The decline was due to some affiliate projects' contract schedules being pushed to the second half and increased investment in future businesses such as new AI platforms and physical AI. Song Gwang-ryun, LG CNS chief financial officer (CFO), said at an earnings conference call held on Jul. 31, "Deferred projects are proceeding normally, and once contract signings and project execution in the second half get underway in earnest, they will contribute to earnings improvement."
However, industry watchers say it is difficult to gauge the mid- to long-term future based on such short-term results alone. In particular, the picture changes when looking at the report cards of AI businesses that each company is cultivating as a future growth engine. All three are supporting AI transitions for group affiliates, including AI infrastructure expansion and AX, and are positioning the physical AI business—which combines robots with manufacturing and logistics systems—as a next-generation growth pillar.
In this earnings release, Samsung SDS disclosed operation and order status for its GPU services and AI data center business. According to the company, it launched GPUaaS based on Nvidia's "B300" via the Samsung Cloud Platform (SCP) in March, and as demand has increased, remaining available capacity is limited. It expects the B300 to be fully operational starting in the third quarter. Lee June-hee, Samsung SDS CEO, said, "Having confirmed solid demand for the latest GPU models, we will continue to steadily expand supply."
Contracts also materialized in the AI data center business. Samsung SDS secured its first order in the first quarter in the corporate DBO business, which outsources design, construction, and operation of data centers, and said several business opportunities took concrete shape in the second quarter as well. It added that the scale of AI infrastructure, currently at 110 megawatts (MW), is planned to expand to 230 MW in 2029 and over 800 MW in 2031.
LG CNS is expanding its business scope around an enterprise AI platform, AI compute infrastructure, and physical AI. Following the first-quarter launch of the enterprise agentic AI platform "AgenticWorks," it unveiled the robot training and operations platform "PhysicalWorks" in the second quarter. In July, it also launched "XPUWorks," which provides AI compute resources such as GPUs and Neural Processing Unit (NPU) as a subscription.
In particular, the physical AI business plans to secure business cases this year through group-affiliate projects and on-site proofs and expand to external customers starting next year. LG CNS recently signed a contract worth 189.7 billion won with LG Electronics to supply GPU infrastructure and PhysicalWorks needed for humanoid robot training. It also completed the first proof of concept (PoC) conducted with Kurly. The company plans to focus on field validation and platform advancement this year, then from 2027 will fully expand businesses building robot transitions and robot data factories for manufacturing and logistics customers.
By contrast, while business opportunities are expected for Hyundai AutoEver from Hyundai Motor Group's investment in physical AI, its specific role, business scale, and monetization timeline have yet to emerge. Although the Robot Meta-Plant Application Center (RMAC), Hyundai Motor Group's robot training and validation facility, is scheduled to begin operations in Aug., the tasks Hyundai AutoEver will undertake and the related revenue scale have not been disclosed.
Brokerages also see the group's AI and robotics investments as potentially a growth opportunity for Hyundai AutoEver, but they say more confirmation of concrete business plans is needed. Shin Yun-cheol, a Kiwoom Securities researcher, said, "Contrary to once-aggressive market expectations, the specific scale, timing, and financing plan for new businesses to be carried out by Hyundai AutoEver have yet to be fleshed out," adding, "Despite successive meetings between Hyundai Motor Group and Nvidia top management, visibility into the new businesses remains foggy."