|
Feb 12, 2026
|
|
LONG
|
Aliaga states that while the "AI Scare Trade" hurts software, the "nuts and bolts" of AI are still in high demand. She explicitly notes, "Memory is really the bottleneck here, not just the raw intelligence... these agents require memory." As AI shifts from simple chatbots to "Agentic AI" (complex tasks), the computational load on memory chips increases exponentially. Pricing power for memory manufacturers (Samsung, SK Hynix, Micron) is surging because demand outstrips supply, and they have "real moats." LONG Memory manufacturers as the primary beneficiaries of the next phase of AI capex. Hyperscalers (Microsoft, Google, Meta) cutting capex spending due to lack of immediate ROI. |
Bloomberg Markets
China's Zhipu Jolts AI Race as 'Scare Trade' ...
|