Industry News

Two Clocks in Storage: Spot Prices Ease While Allocation Locks to 2031

Samsung memory capacity for three years is mostly locked with guarantees sought to 2031, per IT之家. Why that and falling NAND spot prices are both true, and what to buy.

storage-allocation-two-clocks
The storage market is telling two stories at once, and buyers who only hear one of them will plan badly. On Wednesday we covered the easing: NAND spot prices slid about 20 percent, and by Thursday the relief had reached retail shelves in a 4TB flagship drive at 17.5 cents per gigabyte. Now the other clock has spoken. A report says most of Samsung's memory capacity for the next three years is already locked in, with some customers pushing for supply guarantees stretching to 2031 (source: IT之家, November 21, 2026). Client flash is getting cheaper at retail, and the capacity behind it is being sold out years in advance. Both facts are true, and both matter.

Why the two clocks do not contradict each other

The apparent paradox dissolves once the two markets are told apart. The spot and retail clock tracks the marginal price of commodity flash for buyers who purchase now: it eased this week, and consumer drives repriced within days. The allocation clock tracks who owns future production: it has been set years forward by the largest buyers in the market, who paid premiums and prepayments to secure capacity through the AI buildout. A falling spot price and a sold out future are not contradictions. They are the difference between the leftovers market and the reservations market.
The report's horizon is the detail that matters. When customers seek guarantees to 2031, they are pricing in the cost of having been burned: shortage memory converts into contract length, and contract length converts into who gets product when the next squeeze arrives. Long dated capacity deals are becoming the memory industry's normal, not an exceptional episode.

What the locked market means downstream

For downstream electronics makers, the strategic read in the reporting is direct: memory allocation is now a negotiation that happens years ahead, and BOM planning without locked supply is exposure. Any product whose bill of materials includes memory capacity on board competes for allocation against AI infrastructure buyers with deeper pockets and longer contracts. Latecomers do not get better prices. They get the leftovers the reservation holders did not want, at the moment the reservation holders release them.
The mechanics of the new normal are worth spelling out, because they differ from the shortage cycles buyers remember. In the classic squeeze, prices rose and everyone paid more; allocation stayed loosely tied to relationships and orders. In the long dated regime, capacity is reserved before it exists, with premiums and prepayments changing hands years ahead of delivery, which means the squeeze of the next cycle will not raise prices so much as simply exclude the unreserved. The lesson buyers are carrying forward from being burned is that supply security is bought in advance or not at all.
That reality splits the accessory trade into two strategic positions. Products that embed memory, drives with onboard storage and any device where capacity ships inside the product, carry allocation exposure and belong to buyers with supplier relationships deep enough to reserve. Products that provide slots for memory the customer supplies do not. The enclosure architecture, where the buyer brings the drive, is structurally immune to the allocation game, and that immunity is worth something in procurement terms that a price comparison alone does not capture.

The planning checklist for B2B buyers

The week's three storage stories compose into a practical checklist. First, sort the catalog by memory exposure: products that ship with flash inside sit on the allocation clock, products that host customer supplied drives sit on the retail clock, and the two should be planned on separate calendars. Second, for the embedded memory lines that must continue, start allocation conversations earlier than comfort suggests, because the reported horizon says the comfortable window closed two years ago. Third, for the slot architecture lines, lean into the retail easing and merchandise the drive plus enclosure pairing, because that pairing is where this cycle's price relief actually reaches the customer. Fourth, watch the two clocks separately: the spot market can ease while the allocation market tightens, and conflating them is how planning mistakes are made.

The enclosure read on this cycle

This is where the week's news composes into a coherent buyer strategy. Client flash at retail is easing, which makes bare drives cheaper and more abundant in every capacity tier. Allocation remains locked for products that build memory in. The rational assortment therefore leans toward the slot model: M.2 NVMe enclosures, storage hubs with drive bays, and dock ecosystems that accept the drives buyers choose for themselves. The buyer captures the spot relief on the drive, the manufacturer avoids the allocation exposure on the chassis, and neither side needs a 2031 contract to sleep well.
Mattzon's storage accessory line is built on exactly that architecture, with M.2 NVMe enclosure categories and SSD bay hubs designed so the customer supplies the drive, alongside OEM and ODM programs covering 89 percent of products on private tooling. Samples and small trial batches are accepted with no minimum order quantity, and production runs are scheduled in 3 to 25 days from our factory in Shenzhen, with CE, FCC and RoHS documentation for destination markets.
One clock says buy the drive now, and the other says think twice before buying anyone else's memory built in. Both hands agree on the enclosure.
Source: IT之家, November 21, 2026, ithome.com

Read next

More from the journal

Keep readers moving through related announcements, stories, and field notes.