A 30TB TLC enterprise SSD that cost around $3,000 in Q2 2025 was priced at roughly $17,500 by Q1 2026 — a 472% jump, according to Vdura. That single data point sums up the position UK IT leaders are now in. NAND flash contract prices rose 33-38% quarter-on-quarter in Q1 2026 and then accelerated further, while enterprise SSD contract prices surged around 80% QoQ over the same period. TrendForce and supplier ADATA now disagree on how bad Q3 2026 will get — but neither says it gets cheaper. This piece sets out the timing decision UK buyers actually face, with the risks of buying now weighed against the risks of waiting.
The numbers behind the NAND squeeze
NAND flash contract prices — the wholesale price manufacturers charge for raw flash wafers and components — rose 33-38% quarter-on-quarter in Q1 2026, according to TrendForce data cited across multiple trade sources. That was just the opening move. TrendForce's Q2 2026 figures showed NAND contract prices climbing 70-75% QoQ, a rate of increase that, notably, outpaced DRAM for the first time in the current pricing cycle.
For Q3 2026, the picture gets murkier. TrendForce itself forecasts a comparatively modest 10-15% QoQ rise. But ADATA chairman Chen Li-bai has stated that memory manufacturers have already announced NAND Flash price increases of 35-40% for the same quarter. That's a meaningful divergence between a market analyst's forecast and a manufacturer's stated position — and it matters, because it means the official forecast may be understating what buyers will actually be quoted.
Why supply is this tight
This isn't a temporary blip. Samsung, SK Hynix, Kioxia and Micron jointly scaled back NAND flash production in the second half of 2025, tightening supply just as demand from AI infrastructure investment accelerated. Business Research Insights estimates AI-driven applications alone now contribute to nearly 40% of SSD demand growth.
Compounding this, hyperscalers and cloud service providers are securing the bulk of available NAND supply through long-term agreements, according to TrendForce — which squeezes the allocation left over for enterprise and mid-market buyers outside those contracts. TrendForce also expects enterprise SSDs to be the largest single NAND application segment in 2026, driven by that same AI infrastructure build-out.
There is some relief on the way. Samsung has said it plans to ramp up V9 QLC NAND production from early 2026, starting with Phase 1 of a new plant (P4) at its Pyeongtaek campus, aimed at high-capacity demand. But TrendForce's broader assessment is that new fab capacity industry-wide is not expected to reach meaningful production volume before late 2027 or 2028 — meaning this is a multi-year supply constraint, not a quarter-long spike.
What it means for enterprise SSD price tags
The knock-on effect on finished enterprise SSDs has been sharper than the raw NAND figures alone suggest. TrendForce reported enterprise SSD contract prices surged approximately 80% QoQ in Q1 2026, with a further 48-53% QoQ rise projected for Q2 2026.
At the product level, Vdura's tracking shows a 30TB TLC enterprise SSD rising 472% in price between Q2 2025 and Q1 2026 — from around $3,000 to approximately $17,500. Separately, Vdura found that by Q1 2026, 30TB QLC SSD capacity cost 22.6 times as much as equivalent 30TB HDD capacity — a distinct comparison (QLC flash versus spinning disk) that underlines just how far the cost gap between flash and disk has widened, not just how much flash itself has risen.
Anyone modelling a storage refresh against last year's unit-cost assumptions is working from numbers that no longer hold. This is where an storage solution finder exercise earns its keep — matching workload requirements against tiering options before committing capex to all-flash where a hybrid or refurbished approach might do the job.
Buy now or delay: the two paths
The case for buying now is straightforward: every quarter of delay so far has meant a higher price on the next order, and neither TrendForce's 10-15% nor ADATA's 35-40% Q3 2026 forecast implies relief is imminent. Locking in capacity now means fixing a known cost against a market that has, so far, only moved in one direction.
The case for delay is less about price and more about flexibility — if a project can genuinely be pushed past 2026 without operational risk, waiting avoids committing capital into a market that could still correct. But the brief evidence here points against that being a low-risk strategy: hyperscalers are absorbing the bulk of available supply through long-term agreements, so buyers who wait risk not just higher prices but allocation shortfalls and extended lead times on top of the cost increase.
Use the IT finance calculator to model both scenarios against your actual capex cycle before deciding — the arithmetic changes considerably once financing structure and depreciation schedule are factored in alongside the unit price movement.
The market isn't waiting either
Whatever individual buyers decide, the broader enterprise storage market is not pausing for the price cycle to settle. IDC recorded the worldwide external OEM enterprise storage systems (ESS) market at $9.2 billion in vendor revenue in Q1 2026, a 22.7% year-over-year increase. All Flash Arrays (AFA) crossed the 50% revenue threshold for the first time in that quarter, generating $4.9 billion and representing 52.6% of total external OEM ESS revenue.
That combination — rising revenue and flash's growing share of it — suggests that competitors and peers are continuing to commit to flash purchases despite the price surge, rather than sitting out the cycle. Buyers who delay indefinitely aren't just risking price exposure; they're risking falling behind on infrastructure that rivals are already securing.
The UK buyer's calculus
Specific GBP pricing data for enterprise SSDs is scarce, but that doesn't insulate UK buyers — NAND and enterprise SSD contract pricing is set in a globally interconnected market, and the percentage increases reported by TrendForce and Vdura translate directly into higher landed procurement costs here, before any currency or freight factors are added. A UK enterprise budgeting against 2025 unit costs is likely to be working from a figure that is now materially out of date.
This is a case where extending the life of existing assets, rather than defaulting straight to new-build flash purchases, deserves serious consideration. Run existing arrays through a storage end-of-life checker before assuming a like-for-like flash refresh is the only option, and weigh refurbished storage for capacity that isn't performance-critical — it sidesteps the current contract-price spike entirely for workloads that don't need the newest NAND generation.
Where a refresh genuinely can't wait, model the array configuration through a RAID calculator to make sure you're not over-provisioning capacity at these prices, and explore the full range of storage solutions before committing to a single-vendor path. For infrastructure teams also navigating the Broadcom VMware exit, storage refresh timing is increasingly bound up with wider platform decisions — worth reviewing VMware alternatives alongside any hardware refresh business case, since both are capex decisions competing for the same 2026 budget.
Servnet's recommendation
Our reading of the evidence: buy now for anything mission-critical or already budgeted for 2026, because the divergence between TrendForce's 10-15% and ADATA's 35-40% Q3 2026 forecasts means the downside risk of waiting is asymmetric — you could be wrong by a lot, in the expensive direction. Delay is only defensible for genuinely deferrable capacity where refurbished or extended-life alternatives can bridge the gap until supply eases, which on TrendForce's timeline isn't expected before late 2027 or 2028.
Talk to a Servnet engineer before locking in a large order — allocation, not just price, is now part of the negotiation, and getting the sourcing strategy right matters as much as the price you agree.
- •Lock in capacity now for anything budgeted or mission-critical in 2026 — the Q3 forecast range (10-40%) only points upward
- •Treat hyperscaler long-term agreements as a warning: supply allocation is tightening as much as price
- •Extend life on existing arrays via an end-of-life review before committing to a full flash refresh
- •Consider refurbished storage for non-performance-critical capacity to sidestep the current contract-price spike
- •Model both buy-now and delay scenarios through a finance calculator before signing off capex
Sources
Every figure in this article traces to the sources below.
- •TrendForce — Q1 2026 NAND contract price rise of 33-38% QoQ
- •TrendForce — Q2 2026 NAND contract price rise of 70-75% QoQ, outpacing DRAM
- •TrendForce — Q3 2026 NAND price forecast of 10-15% QoQ
- •ADATA — Q3 2026 supplier-announced NAND increase of 35-40%
- •TrendForce — Enterprise SSD contract price surge of ~80% QoQ, Q1 2026
- •TrendForce — Enterprise SSD contract price forecast of 48-53% QoQ, Q2 2026
- •IDC — Worldwide external OEM ESS market revenue and growth, Q1 2026
- •Business Research Insights — AI-driven applications and SSD demand growth
- •SK Hynix — NAND production cuts by major suppliers, H2 2025
- •Vdura — 30TB TLC enterprise SSD price rise and 30TB QLC vs HDD cost ratio