# Memory Shortages Leave Hyperscalers Trailing Memory Suppliers > Source: > Published: 2026-06-21 19:35:44.326579+00:00 # Memory Shortages Leave Hyperscalers Trailing Memory Suppliers Jim Cramer wrote in his Sunday column for CNBC that the AI trade has boosted memory and semiconductor-equipment stocks while putting the major cloud providers at a disadvantage. Cramer wrote that **HBM** (high-bandwidth memory) is a key bottleneck, with **SK Hynix** holding roughly **60%** share and **Samsung** and **Micron** about **20%** apiece, per his column. He also noted that Apple acknowledged price increases tied to memory shifts and that **Microsoft** and **Meta Platforms** flagged higher component pricing on earnings calls, which contributed to elevated capex. Cramer observed that the stocks of the four hyperscalers, **Amazon**, **Alphabet**, **Microsoft**, and **Meta Platforms**, declined over the past month while the tech-heavy Nasdaq was up about **1%**, and he contrasted that with strong performance in memory and storage names such as **Sandisk**, **Western Digital**, and **Seagate** (CNBC). ### What happened Jim Cramer wrote in his Sunday column for CNBC that the AI-driven market rally has chiefly benefited memory and semiconductor-equipment stocks while major cloud providers have struggled. Cramer wrote, "They have run into a brick wall in this stock market. That brick wall is hardware." He identified **HBM** (high-bandwidth memory) as a crucial bottleneck and reported market-share figures of roughly **60%** for **SK Hynix**, and about **20%** apiece for **Samsung** and **Micron** (CNBC). Cramer also noted that **Apple** disclosed price increases tied to memory allocation shifts, and that **Microsoft** and **Meta Platforms** mentioned higher component pricing on their earnings calls as a contributor to large capital expenditures (CNBC). He contrasted the four hyperscalers' recent stock declines with the tech-heavy Nasdaq's roughly **1%** gain and highlighted strong rallies in memory and storage vendors including **Sandisk**, **Western Digital**, and **Seagate** (CNBC). ### Editorial analysis - technical context The column foregrounds a supply-side constraint in **HBM**, a DRAM variant engineered for large-scale AI workloads. Industry suppliers of HBM are highly concentrated, and that concentration creates pricing power and allocation pressure at the component level. Companies that depend on large GPU/accelerator fleets tend to consume disproportionate HBM capacity, which elevates procurement and capital-allocation complexity across the cloud and AI stack. ### Industry context Companies operating in comparable capital-intensive technology cycles often see hardware supply dynamics translate directly into relative market-performance differences between vendors and their suppliers. Observed market moves where semiconductor suppliers outperform platform owners are consistent with past cycles in which scarce, specialized components temporarily reallocated economic value toward upstream manufacturers. ### What to watch Indicators observers should follow include HBM capacity additions and fab announcements from major suppliers, public commentary from hyperscalers about component availability on future earnings calls, and pricing trends reported by memory and storage vendors. Changes in inventory disclosure, vendor allocation policies, or new entrants to the HBM supply chain would materially affect the competitive balance described in the column. ### Practical takeaway for practitioners For teams planning large-scale training or deployment projects, reported concentration in HBM supply suggests procurement timelines, cost assumptions, and capacity planning deserve closer scrutiny. Editorial analysis: Organizations budgeting for heavy GPU usage should factor in potential component-driven capex escalation and supply constraints as plausible inputs when modeling total cost of ownership for AI infrastructure. ## Scoring Rationale This is market commentary from Jim Cramer in his CNBC Investing Club Sunday column, analyzing observable HBM supply-chain dynamics and their effect on AI infrastructure costs and equity performance. The underlying dynamics -- concentrated HBM supply, elevated hyperscaler capex, storage stock outperformance -- are real and practitioner-relevant, but the card is sourced primarily from a single commentator's market-performance column rather than a primary news development. Placed in the Solid tier: genuine data-center infrastructure relevance for practitioners planning compute budgets, but categorically market opinion rather than a reported development. Practice interview problems based on real data 1,500+ SQL & Python problems across 15 industry datasets — the exact type of data you work with. [Try 250 free problems](/problems)