Tell me about Samsung and sk Hynix increasing ai spend

The surge in AI-related capital expenditures, projected to reach $725 billion by 2026 from major tech companies, has created a "memory chip super cycle" that disproportionately benefits South Korean manufacturers Samsung and SK Hynix . These firms are critical suppliers of High-Bandwidth Memory (HBM), a component identified as the primary supply chain constraint for building large-scale AI clusters [6, 22, 30]. This pivotal position has led to a dramatic financial windfall, with both companies' market capitalizations recently crossing the **$1 trillion threshold** [8, 20, 23]. Their financial power has grown to the point where their collective cash flow now reportedly exceeds that of all hyperscaler companies combined . Key AI developers like OpenAI have partnered directly with them for memory supply, and Nvidia has certified both, along with Micron, for its next-generation HBM4 chips [3, 4, 29].

SK Hynix has gained a significant competitive advantage by moving faster than Samsung to supply HBM to Nvidia, a strategic decision that has been widely credited for its recent success [9, 14, 27]. This early focus on AI-related memory chips has allowed SK Hynix's market capitalization to grow to **nearly the same size** as Samsung's over the past three years, a notable shift in the historical power balance . The results are evident in SK Hynix's recent earnings, which showed a five-fold increase in quarterly profit, driven almost entirely by demand for AI memory [18, 25]. This success has fueled a re-evaluation of the traditionally cyclical memory market, with some analysts suggesting a structural shift driven by sustained AI demand could justify higher, more stable valuations for these companies [15, 28].

Despite the overwhelming demand, the companies' investment strategies appear complex and are reported with some contradiction. Initially, memory manufacturers were hesitant to significantly ramp up HBM production due to uncertainty about the long-term sustainability of AI-driven demand . One source from March 2026 claims both Samsung and SK Hynix publicly stated they did not plan to significantly increase production capacity . However, more recent reports from April 2026 indicate that SK Hynix is, in fact, increasing its capital expenditures, influenced by strong revenue from AI leaders like OpenAI and Anthropic, as well as competitive pressure from Chinese rivals [10, 21]. This suggests a potential strategic divergence or a rapid evolution in planning as the scale of the AI infrastructure build-out becomes clearer.

The market is grappling with how to value this unprecedented growth. The combined profits for Samsung and SK Hynix are expected to be around **$400 billion** this year . Yet, despite a significant stock rally, the forward P/E ratios for both companies have been falling, indicating that extremely high earnings expectations are outpacing their stock price appreciation [7, 28]. While the AI boom has cemented their central role, some investor confidence has recently pulled back, reflecting lingering concerns about whether the current high level of demand for their chips can be sustained in the long term .