The handheld shortage is a packaging problem, not a chip problem

Ask almost anyone why handheld gaming machines have been hard to buy for the last two years and you will get the same answer: there is a chip shortage. It is a tidy explanation, it was broadly true in 2023, and it is now wrong in a way that matters if you are trying to work out when the situation improves.

Over six weeks I spoke to engineers and procurement staff at three manufacturers who between them build components for most of the handheld devices sold in Europe. All three described the same shift, independently and without prompting. The constraint is no longer wafer capacity at the front end of the process. It is advanced packaging, the step at the back end where separate pieces of silicon are bonded into a single component, and it has been the constraint since roughly the start of 2025.

None of the three would speak on the record, for the ordinary reason that their employers have supply agreements with the companies whose devices we are discussing. All three hold roles that put them in direct contact with the relevant schedules; two shared internal capacity planning documents, which we have read but are not publishing because they would identify the sources.

What actually changed

Modern handheld processors are rarely one piece of silicon. A typical part now combines a processor die, a graphics die and several stacks of memory, manufactured separately and then joined together on an interposer. That joining step is the advanced packaging, and for years it was cheap, fast and nobody outside the industry had heard of it.

Two things happened at once. Demand for the same packaging capacity from data-centre accelerators went up by an order of magnitude, and those customers pay considerably more per unit than a handheld games console does. Meanwhile the front end, the wafer fabrication everyone worried about in 2023, added substantial capacity that came online through 2024 and 2025. The queue did not disappear. It moved one station down the line, to a station with far fewer seats.

Why the old story persists

The explanation has not been updated publicly because no one in the chain has an incentive to update it. For a device maker, a chip shortage is a blameless act of God. A packaging shortage invites the follow-up question of why your order sits behind somebody else in the queue, and the answer to that is usually that you did not commit far enough ahead or did not pay enough. That is a commercial decision, and commercial decisions are answerable.

There is also simple inertia. The 2023 framing was repeated so widely that it became the default description, and once a phrase like chip shortage is doing the work in a headline there is no particular pressure to replace it. Two of the three people I spoke to said they had stopped correcting journalists who used it, because the correction required a twenty-minute explanation of back-end processes and nobody wanted the quote.

What the manufacturers say

The picture from the procurement side is more specific than the public messaging. One planner described allocation for the 2026 build season as effectively settled by the middle of 2025, with the remaining flexibility measured in single-digit percentages. A second put it more bluntly: the devices that ship on time next year are the ones whose makers signed capacity agreements two years ago, and price was not the deciding factor so much as willingness to commit before the product was finalised.

We put the substance of this to the three device makers whose products are most affected. Two declined to comment. The third said, in a written statement, that it does not discuss supplier arrangements, and pointed us to a public statement from March that refers to industry-wide component constraints. That statement does not distinguish between front-end and back-end capacity, which is precisely the ambiguity this piece is about.

The wafers have not been the problem for over a year. If you want to know when a device ships, ask who booked packaging in 2024. Capacity planner at a component manufacturer

What it means if you are buying one

Practically, three things follow. Availability will stay uneven rather than improving smoothly, because allocation is lumpy and set far in advance. Price cuts on current hardware are less likely than usual, since the constraint is not one a manufacturer can fix by discounting. And the devices most likely to be available are the ones from makers with long-standing supply relationships, which in this market does not correlate neatly with which device is best.

If you are waiting for a shortage to end before buying, the thing to watch is not wafer capacity announcements. It is packaging capacity coming online, which several suppliers have publicly committed to expanding through 2027. That is the number that will actually move the queue.

Method and disclosure

This piece is based on six interviews conducted between 4 August and 12 September 2026, three of them with people who agreed to speak only on background, and on two internal documents shared on the condition they would not be published or quoted directly. Nothing here was funded, suggested or reviewed by any manufacturer. Cold Open does not accept hardware samples, and the devices referenced in our reviews were bought at retail. If any part of this account is wrong we will correct it at the top of this page and date the correction.