You’ve probably noticed that RISC-V is everywhere. Everyone’s talking about it, everyone’s contributing to it, and the website is a bustling metropolis of code and community. But then you remember Open POWER. You go to its website. And you’re met with a digital ghost town. A post-apocalyptic wasteland of dead links and abandoned forums. What happened?
Most people will tell you that Open POWER failed because developers didn’t care. The ecosystem was too small. The software stack was too hard. The community was too fractured. But that’s the story IBM wants you to believe. It’s a lie.
Ten years ago, PowerPC was getting genuine love. It was in game consoles, embedded systems, and even some servers. Developers were excited. Then, quietly, IBM did something that would strangle the entire platform. They spun out their foundries to GlobalFoundries. And then they refused to commit to GlobalFoundries’ next major process upgrades.
Openness without a committed foundry roadmap is just a document, not a platform.
That’s the real story. IBM cut off the manufacturing feedback loop. No leading-edge process commitment meant no competitive silicon. No competitive silicon meant no hardware you could actually buy that was faster than what Intel or AMD were selling. No hardware meant no ecosystem. No ecosystem meant no software. And no software meant no community. The developers didn’t abandon Open POWER — they were starved out.
We’ve been told that open architectures succeed because of passion and community. That’s a comforting fantasy. The reality is that building a competitive modern processor requires billions of dollars in fabrication investments. That’s not something a community can crowdfund. It’s not something a foundation can will into existence. It’s a pure capital expenditure decision made by a single company.
IBM made that decision. They chose not to invest. And Open POWER died.
I saw this firsthand. A friend of mine was a hardware engineer at a company that had designed a custom PowerPC chip for a high-performance computing cluster. The design was brilliant. The software stack was solid. But when they went to fab the next generation, they found that GlobalFoundries had no competitive process for the power envelope they needed. The project was killed. The team was reassigned. The code rotted.
The most dangerous thing in open hardware is a single company that claims to champion it while quietly refusing to fund its manufacturing future.
Now, look at RISC-V. The excitement is real. The community is huge. But the same dynamic is at play. RISC-V will only survive if someone — a company, a consortium, a government — commits to building leading-edge silicon for it. If that doesn’t happen, all the open-source cores in the world won’t matter. They’ll be architectural curiosities, running on FPGA boards that are a decade behind the curve.
We’ve been told that openness is a substitute for manufacturing investment. It’s not. It’s a complement. Without the latter, the former is just a talking point.
So the next time you see a celebration of RISC-V’s community growth, ask yourself: where is the foundry commitment? Who is spending the billions to make this real? If the answer is vague, you’re watching the same tragedy unfold in slow motion.
Open POWER didn’t die because of a lack of developer passion. It died because IBM made a rational business decision to stop investing. And then they blamed the community.
Don’t let history repeat itself.
FAQ
Q: Isn't the failure of Open POWER just because the ecosystem didn't adopt it?
A: No. The ecosystem couldn't adopt it because there was no competitive hardware to adopt. IBM's decision to stop investing in leading-edge manufacturing for PowerPC killed the platform long before any developer lost interest.
Q: What practical lesson should RISC-V proponents take from this?
A: Community enthusiasm is downstream of silicon availability. Without a committed, well-funded foundry plan for leading-edge nodes, RISC-V will remain a niche architecture. Someone has to write the billion-dollar check for manufacturing.
Q: Could Open POWER have succeeded if IBM had kept investing in foundries?
A: Possibly. But the deeper lesson is that open architectures cannot survive on goodwill alone. They need a single powerful sponsor willing to endure the massive capital costs of process development. That's a structural vulnerability, not a fixable one.