What happened to 3Com?

Once a dominant maker of Ethernet cards, hubs and early network servers, 3Com is held up here as a case study in how successful hardware companies can fade. Commenters trace its arc from high‑quality NICs, server OSes and bold bets like Palm and US Robotics, through missteps such as failed software partnerships, weak high‑end routing, and a chaotic exit from enterprise networking before its assets ultimately landed at HP/HPE. Along the way, they contrast 3Com’s engineering strengths and memorable products with the rise of cheap “good enough” integrated networking on motherboards and the strategic choices that left little room to adapt.

Corporate trajectory

  • 3Com evolved from pioneering Ethernet NICs and servers to a broader networking vendor (hubs, switches, routers, IPS, ATM, Huawei JV).
  • Eventually exited enterprise networking abruptly, alarming large customers who had just bought major deployments.
  • Pieces ended up inside HP/HPE/Aruba; some commenters note Comware-based switches are still sold and solid.
  • Several see 3Com as another DEC‑style story: acquired and mismanaged by later HP/Compaq-era leadership.

Hardware products & reputation

  • Early NICs (3c509, 3c595, 3c905) are widely remembered as “they just worked,” especially before Plug and Play and under early Linux/BSD.
  • 3Com gear was often the premium alternative when cheap NE2000/Realtek cards misbehaved.
  • Some enterprise products had serious flaws: buggy SuperStack switches, strange resets under load, and at least one “managed” switch whose config wouldn’t persist.

Enterprise networking & competition

  • Cisco dominated high-end routing; multiple commenters argue 3Com never had competitive core/edge routers (full BGP table, carrier-class).
  • 3Com was stronger in LAN switching and midrange enterprise, but software quality on routers/switches drew criticism.

Software, OS, and Microsoft partnership

  • Earlier, 3Com built its own network OSes (3+Share on DOS, then 3+Open on OS/2 with LAN Manager).
  • A major fixed-fee licensing deal around OS/2 LAN Manager failed; 3Com exited LAN software and handed IP to Microsoft, which then helped displace Novell.
  • Some debate how much OS/2/LAN Manager code really fed into early Windows NT networking.

Palm, US Robotics, and mobile

  • 3Com owned Palm/US Robotics, then spun Palm off in 2000 to “focus on networking,” just before the dot-com bust.
  • Several argue Palm/Handspring/Treo could theoretically have evolved into an iPhone‑class platform, but most think the gap in software, UX, and ecosystem was too large.

NIC integration & market shifts

  • Commenters link 3Com’s decline partly to Ethernet, sound, and basic graphics being integrated onto motherboards, wiping out much of the add‑in card market.
  • Cheap NICs (e.g., Realtek 8139) were technically inferior but “good enough” once CPUs were fast, eroding the premium market 3Com had dominated.

Support, warranties, and longevity

  • Lifetime warranties on NICs and switches were feasible because failure rates were extremely low; some people still get firmware and replacements decades later.
  • Others note that “lifetime” guarantees can be quietly redefined or dropped post‑acquisition, making them less meaningful.

Anecdotes & design details

  • Multiple fond reminiscences: first drivers written for 3c509, early Ethernet lab work, rock‑solid 3Server file servers, and friendly human support lines.
  • Hardware Easter eggs include a “test mode” on 3c509 cards activated with a pencil mark over a PCB pad to recover from bad boot ROM configs.
  • Stories about 3Com founders/early engineers (e.g., meticulous performance analysis, Ethernet history) highlight the company’s engineering‑driven origins, contrasted with its later strategic missteps.