A small hardware company that builds affordable Linux tablets and notebooks just told its community the production line is sitting idle. PINE64 posted in August 2026 that the shortage of two specific memory components has left nothing to build with, and that the earliest realistic restart is mid-2027. For anyone who buys their gear, or follows the open hardware scene, the news lands as a warning shot for the wider market.
The broader story is about how a component squeeze can cascade across an entire category of small vendors, not just one company’s bottom line. When memory supply gets tight at the foundry level, the smallest players feel it first and recover last.
What changed for PINE64’s product line
The company’s short public update was direct. Its manufacturing partners have no DRAM and eMMC inventory to allocate. The PineTab2 tablet and the PineNote e-ink laptop both have roughly twelve weeks of finished stock remaining in distribution channels, and once that runs out there are no new units planned. The PineTime smartwatch, Pinecil soldering iron, and PineVoice accessory are unaffected because they run on microcontroller silicon rather than the contested memory parts. The PineTime Pro, which has been in development for years, has working demo firmware across most of its sensors but cannot move into pilot manufacturing until memory pricing normalises.
I have kept a PINE64 device on my desk for testing since the original PineBook days, so the practical effect for me is that the test unit gets harder to replace, not that I lose access to Linux. The harder effect is for users who were planning to buy their first PINE64 in the next year. The window is closing. Existing owners also face a quieter risk. Repair shops that stock PineTab2 screens and batteries will start running out as soon as finished units dry up, because those parts share the same upstream supply chain as the main boards.
Why memory supply dried up so fast
PINE64 did not name the cause in its post beyond “DRAM and eMMC shortage,” but the industry context is plain. The same memory fabs that supply small hardware vendors have been allocating the bulk of their output to AI infrastructure customers running multi-year procurement contracts. A single hyperscale training cluster can consume more memory in a quarter than an entire year of consumer tablet production. When those orders land with the three memory makers that dominate the global market, smaller buyers get whatever is left.
SK Hynix executives said publicly that demand is expected to outstrip supply for years, not quarters. Micron and Samsung have signalled similar constraints. Component pricing has moved in step rather than smooth curves through the shortage, which means the relief, when it comes, will not arrive gently. It will arrive as batch pricing renegotiations that smaller buyers cannot always access. The other uncomfortable detail is that foundries cannot bring new capacity online quickly. A new memory fab takes 18 to 24 months from initial construction to qualified production output, which means the supply picture in 2027 depends on decisions made in 2025, not on anything the industry can do this year.
- Consumer DRAM allocations have shrunk because hyperscale buyers locked in multi-year deals first.
- New memory fabs take 18 to 24 months from initial construction to qualified output, so 2027 is the realistic floor.
- Spot market prices for DDR5 and eMMC are roughly double their 2024 levels as of mid-2026.
- Even where allocations exist, they often come with minimum order quantities that small vendors cannot meet.
The wider ripple effects on small hardware
PINE64 is the most visible casualty, but the same conditions are squeezing other small Linux hardware projects. The Orange Pi Neo, a Linux first handheld that was supposed to ship earlier this year, has been stuck in limbo because the bill of materials no longer works at consumer pricing. Two years of engineering effort with no launch date. Crowdfunded projects in the concept phase are in worse shape, because their backers expect a finished product at a price that may no longer be possible.
Most of those projects will not announce cancellation. A two person team that shelves a board they were designing simply stops updating their channel. The project becomes a forgotten repository link, and any potential successor has to start from scratch on community knowledge that is now five years old. That quiet attrition is harder to measure than PINE64’s official pause, but it is doing more long term damage to the open hardware world.
Trade-offs to keep in mind
There is no clean resolution here. The trade-offs are real and uncomfortable.
- If you buy a PineTab2 or PineNote now, you are paying retail for a device with limited future replacement parts. That is a real downside.
- If you wait for prices to come down, you may be waiting until 2028 for some categories, and stock may be gone by then anyway.
- If you switch to a non-Linux alternative to bridge the gap, you are locking yourself back into the walled garden that PINE64 existed to provide an exit from.
- If you do nothing, the small hardware world gets thinner by one or two vendors every quarter until the memory market rebalances.
None of these are good options. The honest answer is that this is a supply problem, not a planning problem, and nobody on the consumer side gets to make it go away faster.
What is worth doing this month
Now is the time to act if you were already planning to buy an open hardware Linux device. The window for buying it from a real distributor rather than the used market is now. Inventory is going to drain, and once it is gone the used market prices climb as scarcity sets in. Check PineStore directly rather than relying on Amazon or third party sellers, because the company has been transparent about where its remaining stock sits.
Users who already own a PineTab2 or PineNote should treat the device as the long term piece of gear it now is. Replacement screens, batteries, and charging boards will become harder to source as the support window narrows. Take care of what you have.
Backers of smaller projects like the Orange Pi Neo should watch for any signal that the team has shifted its timeline or its target price. Long silences usually mean the bill of materials will not work at the planned price, and that is the kind of news that arrives quietly rather than in a press release. Ask the team directly. A one sentence answer about whether launch is still realistic is worth more than six months of hope.
Anyone who cares about the open hardware world continuing to exist should consider buying something from a small vendor while you still can. The math is simple. Every sold device is a vote of confidence in the existence of that vendor. Every skipped purchase is one fewer vote in a market that is already losing vendors. This is not a feel-good suggestion. It is the only lever that actually moves the needle, and it only works while the vendors are still around to receive the orders.