Samsung’s earnings report for Q4 2024 indicates that its AI chip business is not ready for big wins due to a combination of external challenges and internal shortcomings. While AI chips have been a growth driver for the semiconductor industry, Samsung has lagged behind competitors, particularly SK Hynix, in securing a strong foothold in this lucrative market.
@ SamFW
Samsung Electronics has posted KRW 75.8 trillion in revenue and KRW 6.5 trillion in operating profit in the fourth-quarter of 2024. Samsung’s full-year revenue was KRW 300.9 trillion, while operating profit was KRW 32.7 trillion.
Samsung’s capital expenditures in 2024 reached KRW 53.6 trillion, including KRW 46.3 trillion spent in the DS Division and KRW 4.8 trillion in SDC. In the fourth quarter, Capex was KRW 17.8 trillion, with KRW 16 trillion allocated to the DS Division and KRW 1 trillion to SDC.
Samsung’s AI plans
Samsung is doubling down on AI across its business divisions to drive growth and technological leadership. The company is leveraging AI to differentiate its smartphones, with the MX Business aiming to reinforce leadership in mobile AI through the Galaxy AI initiative. This strategy includes expanding AI-driven experiences across its flagship Galaxy S series, foldables, tablets, and wearables.
In the semiconductor sector, Samsung is optimizing its portfolio by accelerating the transition to advanced nodes, focusing on high-performance memory products like HBM, DDR5, and LPDDR5x to meet AI-driven demand.
Its System LSI Business is enhancing flagship SoCs and image sensors to support AI-powered mobile devices.
The Foundry Business is developing cutting-edge process technologies, including 2nm GAA and 4nm nodes, to cater to AI and HPC applications.
Samsung Display is strengthening its position in the premium segment by improving display technology, crucial for AI-integrated devices.
In the consumer electronics sector, Samsung is driving the adoption of AI screens under its “Home AI” vision, integrating AI into connected home experiences through SmartThings and enhancing security with Samsung Knox. This AI-focused strategy is expected to shape Samsung’s product roadmap and business performance in 2025.
Samsung’s AI challenges
One major setback has been Samsung’s inability to meet Nvidia’s strict requirements for high-bandwidth memory (HBM) chips. Nvidia, the dominant player in AI GPUs, has primarily relied on SK Hynix for HBM supply, leaving Samsung struggling to gain a meaningful share. Nvidia’s CEO even pointed out that Samsung needs to engineer a new design to meet its standards, signaling a fundamental gap in technical competitiveness, Reuters news report said.
Geopolitical factors have further complicated Samsung’s position. The U.S. government’s export restrictions on advanced semiconductor sales to China have hit Samsung particularly hard, as around 20 percent of its HBM sales were reliant on Chinese customers. This, combined with shifting client preferences toward more advanced AI chips, has slowed down Samsung’s progress in the sector.
The broader tech industry’s supply constraints have also played a role. The high demand and complexity of manufacturing Nvidia’s GPUs have resulted in delays for data center clients, indirectly affecting Samsung’s memory business. While the overall AI chip market is growing, Samsung has not been able to capitalize on this surge in the same way as its competitors.
Financially, the impact of these challenges is evident. Samsung’s fourth-quarter operating profit fell 29 percent compared to the previous quarter, with weak demand for smartphones and personal computers further limiting growth. The company’s decision to use Qualcomm’s processors for its latest Galaxy S25 lineup instead of its in-house Exynos chips also raises concerns about its semiconductor design capabilities.
Meanwhile, rivals like SK Hynix and Taiwan’s TSMC have reported record profits driven by the AI boom. In a striking shift, SK Hynix’s quarterly profit even surpassed Samsung’s for the first time, highlighting the competitive gap that has emerged.
Looking ahead, Samsung’s ability to regain lost ground in the AI chip market will depend on whether it can successfully supply 12-layer HBM3E chips to Nvidia in 2025. While the company expects an overall recovery in memory demand, it remains uncertain whether it can match its competitors in delivering the advanced chips required for AI applications.
Baburajan Kizhakedath
