OpenWRT turns 20; wants to launch their "first upstream supported" design

Original Article ↗ Hacker News Discussion ↗

OpenWrt’s 20th anniversary coincides with plans for its first officially branded router, built with Banana Pi and aimed at being fully open, GPL-compliant hardware sold at roughly cost for under $100. Commenters weigh the benefits of a “reference” device with first-class upstream support, robust recovery features, and predictable long-term availability against challenges like certification costs, performance trade-offs, and competition from cheap or more powerful alternatives. Many express willingness to pay a premium to support OpenWrt directly and escape proprietary, short‑lived vendor firmware, while others question hardware choices, Wi‑Fi capabilities, and whether separate professional-grade routers and access points remain a better value.

Project and Device Overview

  • OpenWrt is planning its first “upstream-supported” hardware, built with BananaPi and aimed at being a showcase for full GPL compliance and first-class OpenWrt support.
  • Target price is under $100; proceeds per unit go as donations to the project, not salaries.

Funding, Governance, and Commercial Involvement

  • Some worry about commercialization corrupting the project (CyanogenMod comparison), but several comments note OpenWrt doesn’t pay developers and only licenses its trademark in exchange for per-device donations.
  • Debate on infrastructure costs: one side claims it could be run very cheaply on free cloud tiers, others with CI/CD experience say this severely underestimates resource needs.
  • There’s unease about paying some developers directly from shared funds; proposals include using money for infra, meetups, and project-wide activities instead.

Hardware Design and Specifications

  • Based on a MediaTek SoC (MT7981B) plus a Wi-Fi 6 NIC (MT7976C), with 2× Ethernet (2.5G + 1G), USB-C UART, read‑only recovery bootloader, NVMe (M.2 2242; thread notes the “2042” spec is likely a typo), RTC, and some industrial interfaces (e.g., Modbus).
  • Some praise: unbrickable design, USB‑C UART, strong focus on openness and upstreaming.
  • Critiques: only two Ethernet ports, no integrated switch, asymmetric 2.5G/1G, NVMe seen by some as overkill, extra dev‑oriented features (JTAG, Modbus, RTC) viewed by skeptics as unnecessary cost.

Price, Scale, and Market Position

  • Under-$100 goal seen as challenging due to small volumes and radio certification costs, but maybe feasible at cost, especially versus similar Filogic-based devices.
  • Comparisons to sub-$100 routers on Amazon lead to clarification that this design aims for openness, compliance, and support rather than bare-minimum BOM.

Comparisons to Other Hardware

  • Turris Omnia cited as the de facto “OpenWrt router” today: very capable but expensive, overkill for many; this new board is seen as a more mainstream “Corolla” alternative.
  • GL.iNet, NanoPi, FriendlyElec, and others discussed: many ship OpenWrt-based or “franken” firmware, often with limited upstream contributions and shorter support windows.

Use Cases, Alternatives, and Concerns

  • Enthusiasts want to buy it to support OpenWrt and get a well-supported, open router; some would pay more than $100 if it ensured long production runs and stable support.
  • Others prefer x86 + pfSense/OPNsense/Mikrotik/Unifi/Ruckus for performance, features, or Wi‑Fi quality; they see OpenWrt as less compelling beyond cost and openness.
  • Some worry about long-term Wi‑Fi driver support and note Linux often lags one Wi‑Fi generation behind for AP use.
  • A few hope it becomes a good target for other OSes (e.g., OpenBSD) and serves as a stable reference platform for years.