Memory chip prices finally reached the thing you actually hold. Samsung raised the Galaxy Z Fold8 Ultra to $2,099.99 and the Z Flip8 to $1,199.99 at Unpacked yesterday, $100 more than their equivalents last year, and told CNBC the increase comes partly from rising memory costs. The new mid-tier Z Fold8 lands at $1,899.99. Preorders opened on the 22nd; general sale is 7 August.

I have been writing the supply side of this story for months and waiting for the moment it stopped being an industry problem and became a consumer one. This is that moment. DRAM contract pricing has climbed roughly 700 percent, and TSMC joined the surge on the logic side at the start of this month. Everyone knew the bill would arrive at retail eventually. It arrived on a stage in London.

For scale, look at where we were five months ago. In February the Galaxy S26 Ultra held at $1,299, and holding the line was itself the story. Samsung ate the input costs to protect a price point. Twenty-two weeks later it stopped eating them. Nothing about the phones explains the gap; the Fold8 Ultra is an iteration on last year’s design, not a reinvention. The gap is memory.

What makes this launch genuinely strange, and what I have not seen anyone say out loud, is that Samsung sits on both sides of the transaction. Samsung Electronics is one of the three companies named in the antitrust complaint over that 700 percent climb, alongside SK Hynix and Micron. Those are allegations and not findings, and I want to be careful about that. But the structural fact stands regardless of how the suit resolves: the division selling you a more expensive phone and the division selling the memory that made it more expensive report into the same building. When Samsung’s mobile arm cites memory costs, it is citing a transfer price from its own semiconductor arm, which is currently having the best year it has had in a decade.

That does not make the increase fake. Samsung Mobile has an internal cost line, and it genuinely went up. It does make the framing incomplete, because a vertically integrated company describing an input cost as an external force is describing a choice about where to book margin.

Ben Wood at CCS Insight read the number the other way, and I think he is right too: he was surprised the increase was not larger, which implies Samsung is still absorbing part of the rise rather than passing all of it through. Both things are true. Samsung is subsidizing the entry price to defend a competitive position, and it is also the supplier setting the input price it is subsidizing against.

The storage ladder is where I would look next, because that is where memory economics show up most nakedly. The Fold8 runs $1,900 at 256GB, $2,100 at 512GB, and $2,500 at 1TB. So 768 gigabytes of additional NAND costs you $600 at retail. The Ultra runs $2,100, $2,300, and $2,700 on the same steps. Storage tiers have always carried fat margin, but in a market where the underlying component genuinely is scarce, that ladder gets harder to read: some of it is scarcity and some of it is the oldest pricing trick in consumer hardware, and from outside you cannot separate them.

The part that worries me is what happens below the flagship tier. Samsung can absorb a hundred dollars on a two thousand dollar foldable. A $250 phone has no room to absorb anything, and memory is a much larger share of its bill of materials. The pressure that showed up as a $100 line item this week shows up further down the range as specification cuts, as configurations quietly disappearing, or as models not getting refreshed at all. The datacenter buildout is what is bidding memory away, and the buyers who lose a bidding war they never entered are the ones at the bottom of the market.

Watch what Apple does with the foldable iPhone that is expected shortly. Apple buys memory from the same three suppliers and faces the same input curve, without a memory division to offset it. If Apple holds its pricing, that tells you something about how much of Samsung’s increase was really unavoidable.

Sources