The company building the thing that might centralize all power just published a warning about centralized power—while missing its own growth targets.
The Summary
- OpenAI released five principles for superintelligence development, warning that without decentralization, ASI could concentrate power in dangerous ways
- The company fell short of internal revenue and user targets heading into AI earnings season, per WSJ reporting
- The irony: OpenAI—the most centralized player in AI—is now positioning decentralization as essential while struggling to hit its own centralized growth projections
The Signal
OpenAI just published governance principles that read like a crypto manifesto written by someone who doesn't actually believe in crypto. The five-point framework warns that superintelligence could "concentrate power" without proper decentralization mechanisms. It's a remarkable document from a company that is itself the poster child for AI power concentration.
The timing makes it stranger. WSJ reports OpenAI missed internal revenue and user growth targets just as the AI industry heads into earnings season. The gap between projections and reality matters because OpenAI's valuation—and the entire AI investment thesis—rests on exponential growth assumptions.
"The company building the most centralized AI infrastructure just discovered the word 'decentralization' exists."
Here's what OpenAI is actually saying: superintelligence needs distributed control, transparent governance, and mechanisms to prevent power from pooling in a few hands. Every decentralized AI crypto project has been screaming this for two years. Bittensor, Fetch.ai, Ocean Protocol—they've been building the infrastructure OpenAI is now admitting might be necessary.
But words are cheap when you're sitting on a centralized model worth tens of billions. The real test is whether OpenAI will actually decentralize compute, model weights, or governance. Nothing in the principles suggests they will. It reads more like preemptive reputation management—acknowledge the risk, propose vague solutions, continue building the central tower.
The revenue miss adds context. When growth slows, companies get creative about narrative. "We're concerned about power concentration" sounds better than "ChatGPT Plus subscriptions aren't scaling like we projected." Both can be true. But the sequence matters.
Key tensions emerging:
- Centralized AI labs acknowledging decentralization matters while maintaining full control
- Crypto AI projects gaining legitimacy as even OpenAI admits distributed models might be necessary
- Growth assumptions behind AI valuations starting to show cracks in real user adoption data
The Implication
Watch what OpenAI does, not what it publishes. If this warning leads to actual decentralized governance mechanisms, open model weights, or distributed compute networks, that's signal. If it stays at the level of principles documents while the company maintains total control, it's reputation theater.
For decentralized AI builders, this is validation. OpenAI just admitted the thing you're building might be essential. The question is whether the market figures that out before or after the centralized model shows more cracks. The revenue miss suggests the clock is ticking faster than the big labs want to admit.