Ask HN: Does anyone care about OpenPOWER?

Hacker News Discussion ↗

OpenPOWER and IBM’s POWER architecture are seen by many as technically solid and uniquely attractive for fully auditable, blob-free systems, but their high cost and limited hardware options keep them a niche choice. Participants contrast this with the rapid rise of ARM and RISC‑V, which benefit from cheap development boards, broader ecosystems, and fewer licensing constraints. There is lingering enthusiasm for open hardware workstations like Raptor’s Talos series, yet doubts remain about long-term viability, software support, and vendor commitment in the face of mainstream x86 and ARM dominance.

Cost and Hardware Availability

  • Raptor’s Talos/Blackbird systems are widely seen as very expensive (multi‑thousand dollars, entry configs cited ~$4–6k+), especially relative to commodity AMD64 desktops.
  • Some argue the pricing is normal for server‑grade gear (ECC, dual sockets, IPMI, quality PCB, US manufacture). Others counter that these are “desktop‑like” boards with last‑gen POWER9 and weaker price/performance than x86 servers.
  • A recurring complaint: there are no cheap POWER SBCs or hobbyist boards. People compare this unfavorably to sub‑$100 ARM and RISC‑V boards.
  • Used datacenter POWER gear exists but comes with typical downsides: noise, power, form factor.

Openness and Security as the Main Selling Point

  • OpenPOWER interest is driven mostly by fully auditable, blob‑free systems and owner‑controlled firmware.
  • Raptor is highlighted as unique in pushing for deblobbed NIC firmware, open BMC, and secure boot chains.
  • Some value this enough to justify the cost and treat it as business‑grade or IP‑protection hardware rather than consumer gear.

IBM, POWER10, and Roadmap Problems

  • IBM’s POWER10 requires non‑redistributable firmware blobs, including on the memory controller, making it unusable for Raptor‑style fully open machines.
  • This effectively stranded Raptor on POWER9 and killed short‑term hopes for a Talos III based on POWER10; some users are waiting for a future Power11 or alternative platform.
  • Debate exists over whether Raptor “torched” its roadmap or was forced into a dead end by IBM’s choices.

Ecosystem and Software Experience

  • Opinions diverge: some report good distro support (e.g., mainstream Linux with plenty of prebuilt packages).
  • Others cite real pain: lack of Firefox JIT, no native Electron apps, fragile niche distros, and general “fighting the ecosystem,” pushing them back to x86_64.
  • POWER is viewed as unfriendly to hobbyists: expensive entry, sparse documentation/help (e.g., for hypervisor mode), and weak community support.

Competition: RISC‑V, ARM, x86, and Others

  • Many see no compelling advantage of OpenPOWER over RISC‑V at the ISA level; performance is viewed as mostly an implementation/funding issue.
  • RISC‑V’s success is attributed to a very small base ISA, cheap licensing (none), easy extensibility, and rapid growth in embedded and low‑end boards.
  • ARM and x86 dominate desktops, laptops, servers, gaming, and supercomputing; POWER largely survives in big‑iron enterprise and some niche workstations.

Future Prospects and Niche Enthusiasm

  • Enthusiasts highlight bi‑endian support, open hardware projects (LibreSOC, POWER notebook efforts), and features like transactional memory and CAPI, though real‑world use is unclear.
  • Skepticism is strong about long‑running crowdfunded POWER projects and the likelihood they will deliver competitive, affordable hardware.
  • Some foresee a “post‑PC” world where desktops become a niche like standalone cameras; in that scenario, OpenPOWER’s already‑niche economics might age better than today’s mass‑market PC ecosystem.