Raspberry Pi just announced its third price increase in five months, and this one hurts.
The flagship Pi 5 16GB jumped from $120 to $220 -- an 83% increase that pushes what used to be affordable computing into premium territory. The reason is simple and frustrating: AI companies are hogging all the memory.
What just happened
On April 1st (and no, it was not a joke), the Raspberry Pi Foundation announced sweeping price increases across most Pi 4 and Pi 5 models.
The damage:
- Pi 5 16GB: $120 to $220 (+$100)
- Pi 5 8GB: $80 to $130 (+$50)
- Pi 5 4GB: $60 to $85 (+$25)
- Pi 4 8GB: $75 to $125 (+$50)
- Pi 4 4GB: $55 to $80 (+$25)
Over a dozen models saw increases ranging from $25 to $150. The company also launched a new Pi 4 3GB model at $83.75, positioned as the "affordable" option now that the 4GB version costs $80.
This is the third increase since November 2025. What started as a $35 computer is now pushing $225 for the flagship model.
To put that in perspective, here is a brief history of Pi pricing. The original Raspberry Pi Model B launched in 2012 at $35. For most of the next decade, that $35 price held across the mainline models. The Pi 4 launched at $35 for the 1GB model in 2019. Even the Pi 5 started at $60 for 4GB in 2023, which felt expensive at the time but reasonable for the hardware. The jump to $85 for that same configuration is a 42% increase. The $220 flagship would have been unthinkable three years ago.
Why your Pi costs more than your streaming stick
The culprit is LPDDR4 memory. That stands for Low-Power Double Data Rate 4, and it is the same type of memory chip that powers your Pi, your phone, and increasingly, AI data centers.
According to Raspberry Pi CEO Eben Upton, the cost of LPDDR4 DRAM has risen seven times over the past twelve months. Memory manufacturers like Samsung, SK Hynix, and Micron are prioritizing lucrative AI infrastructure contracts over consumer hardware.
Makes sense from their perspective. Google and Microsoft pay premium prices for guaranteed supply. They sign multi-year contracts worth billions. Raspberry Pi competes for the leftovers.
Here is how the memory market works in simple terms. There are only a handful of companies in the world that manufacture DRAM chips. Those companies have limited production capacity. When demand from one sector explodes, as it has with AI, the manufacturers allocate their output to whoever pays the most. AI training clusters need enormous amounts of memory. A single large language model training run might use thousands of GPUs, each paired with high-bandwidth memory. The sheer volume of AI-related orders has absorbed so much production capacity that everyone else in the market feels the squeeze.
The result: a supply squeeze that is hitting every corner of affordable computing. Gaming handheld maker Ayaneo already canceled their Next 3 device. Other manufacturers are pushing indefinite delays rather than eat the costs. Smartphone makers are reportedly holding off on higher-memory models because the component costs no longer make sense for mid-range pricing.
A quick timeline of the price hikes
The increases did not happen all at once. They escalated:
- November 2025: First round of increases, mostly $5 to $15 bumps across the lineup. Felt manageable. Most people shrugged.
- January 2026: Second round. The Pi 5 8GB went from $70 to $80. The higher-memory models jumped more aggressively. Forums started getting noisy.
- April 2026: The big one. The Pi 5 16GB nearly doubled. The Pi 4 8GB jumped $50. The tone shifted from "this is annoying" to "this changes the hobby."
Three rounds in five months. Each one described as a response to worsening supply conditions. Each one bigger than the last.
What this means for your projects
If you have been planning a home lab build, media center, or IoT project, the math just changed.
For basic projects: The Pi 4 3GB at $83.75 can still handle most single-board computer tasks. File servers, Pi-hole, simple automation, VPN servers, and lightweight Docker containers. You probably do not need 8GB anyway. Most home lab tasks barely touch 2GB of RAM. The 3GB model is genuinely fine for the majority of builds covered on this site.
For demanding builds: That AI model you wanted to run locally? The computer vision project? High-memory builds now cost what a decent laptop did two years ago. A Pi 5 with 16GB at $220 plus an SSD, case, and power supply puts you well over $300. At that point, a used mini PC from eBay with an Intel i5, 16GB of RAM, and a 256GB SSD might be a better deal for pure performance.
For existing setups: If you already have Pis running, do not mess with them. Replacing a working Pi 4 8GB will cost you $125 instead of the $75 you paid originally. Protect what you have. Maybe add a UPS. Definitely do not tear down a working Pi to "upgrade" it right now.
For classroom and education use: This is where the price hikes hit hardest. The Pi was born as an educational tool. Schools and coding clubs that used to buy ten Pis for $350 now need $850 for the same order. The Foundation knows this is a problem, which is partly why they released the 3GB model at a lower price. But the optics are rough. The "affordable computer for everyone" message gets harder to sell at $84.
The alternatives are not great either
The memory crisis is not just hitting Raspberry Pi. It is affecting the entire single-board computer market.
Orange Pi, Rock Pi, and other alternatives are seeing similar pressure. Some have better specs for the money. The Orange Pi 5, for example, still offers strong performance per dollar. But most alternatives lack the Pi's ecosystem, community support, and software compatibility. Raspberry Pi OS, the documentation, the forums, the sheer volume of tutorials and project guides -- that is a real advantage that cheap alternatives do not replicate.
Used Pis are holding value better than most electronics. Check eBay. Working Pi 4 4GB units are selling close to new retail prices from six months ago. In some cases, used boards are selling for more than what they originally cost new. That is unusual for electronics and tells you something about how people are feeling about supply.
For certain use cases, the alternatives make sense even without the price crisis. An ESP32 microcontroller costs $5 to $10 and handles sensor projects, simple automation, and IoT tasks without needing a full Linux operating system. An old laptop or a $100 used mini PC from Dell or Lenovo can run Docker, media servers, and home lab software with more power and storage than any Pi. The Pi's advantage was always its price-to-capability ratio plus its community. When the price leg weakens, the calculation shifts.
What Raspberry Pi is not saying
Upton frames these as temporary measures, promising to "reverse increases when circumstances abate." But no timeline comes with that promise.
That absence matters. It suggests even manufacturers close to memory supply chains cannot forecast relief.
The company is also being careful not to blame AI directly, referring instead to "competition for fab capacity." Diplomatic, but everyone knows where that competition is coming from. OpenAI, Google DeepMind, Meta, Anthropic, and dozens of other AI companies are collectively spending hundreds of billions on compute infrastructure. That spending drives demand for GPUs, memory, power supplies, networking equipment, and datacenter capacity. Consumer electronics like the Pi are collateral damage.
There is also a question about the Foundation's pricing philosophy. Historically, Pi models held their launch prices for the entire product lifecycle. A Pi 4 at $55 stayed $55 until it was discontinued. This new pattern of rolling price increases is a departure, and it changes how people plan projects. You used to be able to bookmark a build guide and come back in six months knowing the parts list would cost the same. Not anymore.
My honest take
This sucks, but it is probably not temporary.
AI infrastructure spending is measured in hundreds of billions. The Raspberry Pi Foundation's entire revenue is around $200 million. When hyperscalers need memory, they get it first. That dynamic is not going to reverse just because hobbyists are unhappy.
The original Pi launched during a very different time, when mobile chips were getting cheaper and memory was abundant. That era is over. Memory demand from AI is expected to grow for at least the next several years as training runs get larger and inference infrastructure scales out.
We are moving toward a world where "affordable computing" starts at $100, not $35. The Pi will survive, but the audience shifts from curious hobbyists to serious makers and small businesses who can justify the cost for specific projects.
I do want to say this clearly: the Raspberry Pi Foundation is not the villain here. They are a relatively small buyer in a market dominated by companies with essentially unlimited budgets. Upton has been more transparent about pricing pressure than most hardware CEOs. The situation is genuinely difficult, and there is no obvious solution on the Pi side.
What to do next
If you need a Pi now: Buy the 3GB model or find what you need used. Do not wait for prices to drop. If history is a guide, the next announcement is more likely to be another increase than a decrease.
If you are planning projects: Consider whether you actually need a full Pi. ESP32 boards handle many IoT tasks for under $10. Older laptops or mini PCs might be better for media centers and home lab servers. A $60 used Dell OptiPlex with an i5 processor will outperform any Pi for most server workloads.
If you are building a home lab: This is a good time to evaluate whether multiple Pis make sense, or if one decent mini PC would be more cost-effective. A single $150 mini PC can run a dozen Docker containers that would have required three or four Pis to match.
If you already own Pis: Take care of them. Update the software. Use quality power supplies. Back up your SD cards. These boards are now worth more than what you paid for them, and replacing a dead one costs real money.
If you are just browsing: Wait. These prices will likely get worse before they get better. And when they do get better, it will be gradual, not a sudden return to $35.
The Pi revolution democratized computing for a generation. What comes next probably will not be as cheap, but it might be more focused on people who actually build things rather than those just curious about the idea.
That is not necessarily bad. It is just different.