The Numbers
Raspberry Pi announced its second price increase in three months on February 2. Models with 2GB RAM are up $10, 4GB models up $15, 8GB up $30, and 16GB units up $60. The flagship Pi 5 16GB now costs $205 - up 70% from its original MSRP of $120.
The Pi 500+ keyboard PC hits $260. Only 1GB models (Pi 4 at $35, Pi 5 at $45) hold their original pricing.
What's Driving This
Global LPDDR4 prices doubled in Q4 2025, according to Raspberry Pi founder Eben Upton. The cause: memory manufacturers reallocating fab capacity from commodity DRAM to high-bandwidth memory (HBM) for AI accelerators.
Samsung committed 60,000 wafers per month to HBM4 production by September 2025. OpenAI's "Stargate" project with Samsung and SK Hynix reportedly consumes 40% of global DRAM capacity - roughly 900,000 wafers monthly. TrendForce now projects DRAM contract prices will rise 90-95% this quarter, up from earlier estimates of 60%.
Nvidia's datacenter systems add pressure on LPDDR supply - the Vera Rubin NVL72 alone requires 54TB of LPDDR5X per rack.
What This Means For Edge Deployments
Enterprise teams running Kubernetes edge nodes on Pi clusters face immediate budget impacts. A 10-node K3s cluster using Pi 5 8GB units now costs $1,250 versus $800 six months ago - a 56% increase.
The pricing shift narrows Pi's advantage over x86 alternatives. A fully-kitted Pi 5 8GB (with case, power, storage) now approaches mini PC territory at $160-180, though power consumption still favors Pi for battery-powered edge applications.
Alternatives Worth Evaluating
Orange Pi 5 variants remain cheaper but lack Pi's ecosystem maturity and arm64 Docker support consistency. For machine learning inference at the edge, used Intel NUCs or Lenovo ThinkCentre Tiny systems offer better performance per dollar if power budgets allow.
Upton calls the shortage temporary, expecting reversal as HBM production matures and regular DRAM capacity returns. History suggests 6-12 months before fab rebalancing takes effect.
The real question: at these prices, does Raspberry Pi still make sense for production edge deployments, or is this the push enterprise teams needed to re-evaluate their hardware strategy?