In March 2026, PINE64 made an unusual hardware decision: instead of pushing the price of several devices sharply upward, PineStore chose to stop producing PineTab2, PineNote and PinePhone for the time being.1 PINE64 directly attributes the decision to the DRAM shortage and says PineStore founder TL Lim does not want to raise hardware prices significantly.1 Remaining inventory can be sold, but another production run has to wait for a better equation.
That is narrower than “PINE64 is leaving Linux hardware because of the AI bubble”, since the company's own post talks about a DRAM shortage, existing financial risk around a possible PinePhone 2, and a mobile-Linux market where cheaper and more powerful used Android phones can now become Linux targets.1
To be sure, the memory market itself shows the pressure clearly. TrendForce describes memory suppliers reallocating capacity toward server and AI demand during 2026, leaving conventional DRAM supply unusually tight and pushing contract prices sharply upward.23
PINE64 is a useful case for seeing what that market shock looks like once it reaches a relatively small hardware maker, where component cost and final retail price have much less room to drift apart.
An ordinary component can stop the whole product
RAM is not the reason anybody wants a PinePhone or PineNote; these devices are sold around openness, Linux, replaceable software and a community that can work directly on the platform, yet memory becoming too expensive is enough to make another manufacturing run unattractive.1
Electronics manufacturing is unromantic in that way, because the least distinctive line on a bill of materials can end up deciding whether the distinctive product exists at all.
PINE64 does not publish the exact DRAM share of each device cost or the retail increase that would have been necessary to preserve its margins, so inventing a reconstructed bill of materials would add certainty that the source does not provide. The confirmed point is simpler: PineStore considered the cost pressure large enough to pause production instead of passing a significant increase to buyers.1
That choice says something about the products' position in the market: an open Linux tablet or phone cannot absorb any price simply on the strength of being open, and at some point the comparison with mainstream or second-hand hardware becomes too difficult.
Large memory buyers play a different game
TrendForce describes a 2026 memory market in which suppliers are prioritising server applications while major cloud customers secure part of their supply through multi-year agreements.23
For the second quarter, the research firm forecast conventional DRAM contract prices rising 58–63% quarter over quarter, following an already severe first quarter.2 The forecast increase moderates to 13–18% in the third quarter, but the market remains undersupplied and capacity continues to favour server demand.3
Those are market figures rather than invoices paid by PINE64, but they describe the environment in which a small buyer has to negotiate while much larger customers are securing supply on different terms.
A large OEM can commit to substantial volume, spread higher costs across a wider range, negotiate long-term supply or accept lower margins on one strategic product. By contrast, a niche manufacturer has fewer levers, and the cost of a memory shock is harder to hide elsewhere in the catalogue. Orders are smaller, the range is narrower and the final selling price is often central to the product's appeal.
Small hardware makers are simply operating from the weaker side of that supply-chain asymmetry, with less volume available for negotiation and fewer product lines over which to spread the shock.
Low volume makes each cost decision visible
PINE64 also operates unusually because its devices often serve as open platforms whose value depends heavily on community software work.
That can change where development effort sits, but the company still has to buy SoCs, memory, displays, PCBs and batteries and then finance a production run, so community software work cannot make the silicon side of the bill disappear.
In the long term, low volume makes a component increase difficult to dilute, because there are fewer units and fewer product families across which to spread it. Adding tens of dollars to a niche phone can abruptly change its comparison with a second-hand Android device. PINE64 makes exactly that broader point around PinePhone 2: beyond the current chip shortage, the project is considered financially risky because the mobile Linux community can port software to cheaper and more powerful Android hardware.1
The shortage therefore does not create all of the fragility. It presses on a business model that already has very little room for a bad component cycle.
PINE64 did not stop making everything
Another important correction is that the whole product range has not been suspended: the March post says PineTime, PineBuds, Pinecil and PineVoice are not affected in the same way.1 By May, PINE64 was announcing PineVoice production and new PineTime Pro developer samples.
That makes “open-source hardware is no longer viable” a poor conclusion. Different devices have different memory requirements, bills of materials and production economics, so the same market shock reaches them in different ways.
The useful unit of analysis is the individual product rather than the brand as a whole.
Openness does not protect a product from component economics
The PINE64 case separates two kinds of openness that are easy to blend together.
The first is technical: accessible bootloaders, Linux support, modifiable firmware and a community that can maintain software around the hardware. PINE64 has built unusually open products on that axis.1
The second would be economic openness, meaning the ability to keep manufacturing those products under any market condition; no technical licence can provide that when a small company is buying components from the same industry as much larger customers and suppliers may prefer higher-margin server orders.
A phone can be fully flashable without making its RAM independent from the global memory market.
That is the more useful lesson from the pause. Open hardware does not live in a parallel universe made only of PCBs and good intentions; it still depends on fabs, contracts, minimum volumes and components whose industrial priorities can change far away from the community using the final device.
PINE64 has chosen not to hide that pressure behind a large retail increase. For several devices, its current answer is blunt: a new production run makes less sense than a product whose final price has stopped making sense.