New York, Feb 9, 2026, 16:07 EST — After-hours
- Micron slipped roughly 2.9% in after-hours trading, marking a second evening of jitters tied to HBM4 supply concerns for Nvidia’s upcoming AI processors.
- Samsung is said to be gearing up to mass-produce and ship HBM4 chips for Nvidia’s “Vera Rubin”—a move that intensifies competition with SK hynix and Micron.
- Eyes now turn to February’s HBM4 shipments and Nvidia’s GTC 2026 set for next month, as traders hunt for more clarity on the supplier mix.
Micron Technology dropped roughly 2.9% to $383.24 in late trading Monday, with investors reacting to new reports about which suppliers will provide high-bandwidth memory for Nvidia’s upcoming AI chips.
The action in the shares is notable: high-bandwidth memory, those stacked DRAM chips positioned near AI processors to crank up data speeds, now represents one of memory’s richest profit streams. Landing a supplier position on a top Nvidia platform can alter revenue outlooks in a hurry.
HBM sits among the most constrained segments in AI hardware. If investors sense Micron is losing share, or see signs it’s late in the ramp, the stock tends to move before the details are clear.
Samsung Electronics is set to kick off mass production of its sixth-generation high-bandwidth memory, dubbed HBM4, in the coming weeks, with shipments expected to reach Nvidia for use in its graphics chips shortly after the Lunar New Year break, according to industry sources cited by South Korea’s Korea JoongAng Daily. The paper reports Samsung cleared Nvidia’s quality checks and locked in orders for the “Vera Rubin” rollout. Korea Joongang Daily
Nvidia is gearing up to launch Vera Rubin at GTC 2026 in mid-March, TrendForce reported, citing Yonhap News and Hankyung. Early HBM4 shipments will mostly come from SK hynix, with Samsung projected to take a mid-20% slice and Micron close behind at about 20%. Samsung, the report notes, is touting data rates hitting 11.7 gigabits per second — that’s a “pin speed,” or how fast each connection moves data — and single-stack bandwidth reaching 3 terabytes per second. TrendForce
Industry chatter around timing and volume is swirling, BusinessKorea reported, adding that no official word has come out yet about early supply numbers. The outlet pointed to a line from SK hynix’s earnings call: the company is “proceeding with volume production as scheduled in accordance with customer requirements.” Businesskorea
Micron shares have slumped, and MarketWatch points to speculation the company might miss out on Nvidia’s initial Vera Rubin chip supply. Some are questioning if Micron’s HBM4 can deliver the pin speeds Nvidia needs. Mizuho’s Jordan Klein blamed the stock’s slide on “moronic press” reports doubting whether Micron’s HBM4 is up to Rubin’s speed standards. MarketWatch
No sign of panic across the chip sector. Nvidia gained roughly 2.4% late, while the VanEck Semiconductor ETF added 1.2%. The action zeroed in on Micron’s memory battle, not a sector-wide AI retreat.
Investors want more concrete answers—who’s shipping, what’s shipping, and on what timeline. Attention is turning to Samsung: will those rumored HBM4 shipments actually start in the third week of February? Nvidia’s GTC event follows in mid-March, and the street is looking for specifics on the Rubin platform and the official list of suppliers.
Banking on early “share” math carries its own risks. HBM qualification isn’t fixed—it shifts as yields fluctuate, packaging capacity changes, or customer testing uncovers issues. Supplier mix? That’s another moving target if a vendor slips up on volume or quality.
Micron’s next big move hinges on one thing: any shift—positive or negative—in the HBM4 qualification clock or early Rubin supply. Traders are eyeing February shipment data, plus whatever hints come out of Nvidia’s GTC 2026 in mid-March, in search of a clear signal.
