In a resounding New Year’s declaration that has sent ripples through the global semiconductor market, Samsung Electronics co-CEO and chip chief Jun Young-hyun announced that the company’s next-generation HBM4 chips have received overwhelming praise from major clients. During his 2026 New Year address, Jun revealed that several key customers—widely believed to include AI titans like Nvidia—have explicitly stated that "Samsung is back." This feedback marks a significant psychological and technical turning point for the firm, which spent much of the previous two years scrambling to close the gap with its chief rival, SK Hynix.

The praise centers on the "differentiated competitiveness" of the 6th-generation High Bandwidth Memory (HBM4) solution. Unlike previous iterations that faced rigorous quality hurdles, the HBM4 prototypes have reportedly exceeded efficiency and thermal benchmarks in recent customer evaluations. Market analysts suggest that Samsung’s shift to 1c nm DRAM technology and its decision to optimize chips for specific AI accelerators have finally paid off, positioning the company to reclaim its title as a leader in high-performance memory for data centers.

Investors responded to the news with immediate enthusiasm, driving Samsung’s shares up by as much as 7.2% on the first trading day of 2026. This surge significantly outpaced the benchmark KOSPI index and reflected a renewed confidence that the company is no longer just a "follower" in the AI chip race. The rally comes at a critical time, as industry data from late 2025 showed Samsung trailing with a 35% market share compared to SK Hynix’s 53%. The positive HBM4 feedback suggests a potential shift in those rankings as volume production begins.

Beyond memory, Jun Young-hyun also highlighted a "great leap forward" for Samsung’s foundry business. He pointed to several massive supply deals signed in late 2025—including a landmark $16.5 billion agreement with Tesla—as proof that the company’s manufacturing arm is becoming a preferred partner for custom AI silicon. By integrating its HBM4 technology directly into its foundry offerings, Samsung aims to provide a "one-stop shop" for AI companies, a strategic advantage that rivals specializing only in memory cannot easily replicate.

However, the path forward remains fraught with external challenges. In a separate address, co-CEO TM Roh warned that 2026 would likely be a year of "pessimistic optimism," characterized by rising component costs and increasingly volatile global trade tariffs. To counter these risks, Samsung is reportedly planning a 50% surge in HBM capacity throughout 2026, converting existing lines at its Pyeongtaek P4 facility to meet a backlog of orders that are already rumored to be sold out through the first half of the year.

The competition is not standing still, as SK Hynix CEO Kwak Noh-jung acknowledged in his own New Year remarks that while the "AI surprise" is over, the era of "AI normalcy" has begun. SK Hynix is expected to start its own commercial production of HBM4 as early as February 2026. This setup creates a high-stakes "supercycle" where both South Korean giants will battle for dominance in the 16-high (16-die-high) HBM segment, which Nvidia has requested for its next generation of Blackwell-successor GPUs.

As 2026 unfolds, Samsung’s resurgence signals a healthier, more competitive AI ecosystem. If the "Samsung is back" sentiment translates into mass-market delivery, it could ease global supply constraints for AI hardware and accelerate the deployment of next-gen large language models. For now, the focus remains on the upcoming first-quarter quality qualifications, where the world will see if Samsung can turn its technical praise into the largest supply contract in the history of the semiconductor industry.