$NVDABullishLow

Nvidia and SK hynix to Partner as Jensen Huang Warns Memory Shortage Could ‘Last for Years’

Nvidia CEO Jensen Huang met SK hynix leadership in Seoul and, according to CNBC, said the global memory shortage could persist for several years. The report says Nvidia and SK hynix are deepening cooperation around high-bandwidth memory used in Nvidia’s Blackwell platforms, with SK hynix supplying an estimated 50%–70% of Nvidia’s HBM4 needs. Investors may view this as multi-year supply planning amid constrained capacity.

6/10
4/10
Low
Bullish
Ahead of any announced Nvidia–SK hynix cooperation details “tomorrow” (per article).
Supports a bullish AI-infrastructure narrative by emphasizing structural (multi-year) HBM constraint.

Signals multi-year HBM supply visibility for Nvidia’s next GPU cycles, potentially reducing near-term supply-cost uncertainty.

Article says Jensen Huang warned the global memory shortage could persist for several years and hinted an expansion of Nvidia–SK hynix cooperation.

Bias toward supportive sentiment for NVDA on any follow-through that confirms expanded HBM commitments.

Background

The article ties AI infrastructure buildout to a persistent high-bandwidth memory (HBM) bottleneck and describes a deepening Nvidia–SK hynix relationship.

Why it matters

Huang’s “several years” warning shifts expectations from short-cycle tightness to longer-duration capacity planning, which can influence procurement assumptions and investor positioning around AI hardware supply chains.

Market relevance

Treat as a catalyst for multi-year HBM allocation expectations; direct tradable impact depends on whether the hinted “tomorrow” announcement confirms binding volume/pricing terms.

Market effects

Reinforces that AI hardware bottlenecks are structural (HBM capacity/yield), not cyclical, which can reprice memory supply risk across the stack.

Highlights Korea-based HBM capacity planning and potential “AI factory” integration in Korea.

If shortages persist for years, global AI capex-to-hardware conversion may remain supply-constrained, affecting downstream server/accelerator lead times worldwide.

Alternative perspectives

The article is largely interpretive (“quickly interpreted” as expansion) and may not confirm binding contract terms; markets could already be pricing HBM tightness.

HBM supply expansion depends on yield/process improvements; if ramps slip, the “several years” narrative could increase volatility in customer allocation and pricing rather than smooth it.

Key entities

  • Jensen Huang

    Nvidia CEO quoted saying the global memory shortage could persist for several years and teasing a possible announcement.

  • SK hynix

    HBM supplier described as central to Nvidia’s Blackwell accelerators and next-gen platforms.

  • Chey Tae-won

    SK Group Chairman mentioned as part of the high-profile meeting context.

Related articles

$AVGOMed

Weekend Round-Up: Anthropic Eyes IPO, CrowdStrike's Stock Split, Broadcom's Mixed Results, SpaceX's AI De

Anthropic filed a confidential draft Form S-1 with the SEC, giving it the option to pursue an IPO, though share count and pricing weren’t set. CrowdStrike reported Q1 revenue of $1.39B vs $1.36B estimates and announced a 4-for-1 stock split. Broadcom’s Q2 revenue was $22.19B vs $22.27B estimates, with adjusted EPS of $2.44 vs $2.40. SpaceX signed a $920M monthly AI deal with Google.

$NVDAMedAI 8/10

Nvidia Acquires Kumo AI for $400 Million. What That Means for NVDA Stock.

Nvidia will acquire Kumo AI for $400 million, according to the article, to expand its AI model portfolio and potentially integrate Kumo’s models into its AI foundry software or use its researchers for enterprise foundation models. The company reported Q1 revenue up 85% YoY to $81.6B and EPS of $1.87. Nvidia raised Q2 revenue guidance to $91B (excluding China).

$GOOGLMedAI 9/10

Google Signs $30bn Cloud Computing Deal With Elon Musk’s SpaceX

Google signed a three-year cloud computing deal with SpaceX, committing about $30bn to secure AI infrastructure, according to regulatory filings. Google will pay SpaceX about $920m monthly from Oct 2026 to Jun 2029 for access to Nvidia-based computing, including ~110,000 GPUs and related hardware. The contract targets rising demand for Google’s AI services and can be terminated if chip access isn’t met.

$NVDAMed

Global Economy Briefing — June 6, 2026

US May payrolls rose 172K vs 85K consensus, with unemployment steady at 4.3% and wages up 3.4%, while consumer credit beat at $20.73B. Equities fell sharply as Treasury yields jumped; the 10-year moved above 4.5% and the 30-year above 5%, with markets pricing a 98% chance of a Fed hike. Brazil’s real positioning reportedly flipped as CFTC net longs fell to 47.0K from 71.7K.