We mined liquidity while the code slept.
That line comes from the 2020 DeFi Summer, when I watched farmers dump millions into unaudited pools, chasing APY like it was free money. The same pattern is now repeating in AI infrastructure. A new non-profit called Current AI just raised $400 million from Google and the French government. The promise? A free, open, decentralized AI layer — a World Wide Web for artificial intelligence. Tokens are not involved. There is no token. No airdrop. No DAO. Yet, as a battle-tested trader who has survived the Parity hack, Terra's collapse, and the 2024 ETF arbitrage grind, I smell a familiar scent: euphoria masking technical flaws.
Context: The Open Infrastructure Mirage Current AI is not a model builder. It is not a cloud provider. It is an attempt to build a non-profit foundation that aggregates open-source AI components — datasets, models, compute — into a single interoperable layer. The backers include Google (through its cloud division) and the French government (pushing for European AI sovereignty). The $400 million is seed capital, not a revenue-based valuation. The goal is to lower the barrier for any developer to train or deploy AI without being locked into a proprietary API.
This sounds noble. It sounds like Linux for AI. But let me translate that into the language of on-chain risk: Current AI is a smart contract with no formal verification, governed by a multi-sig where two of the keys are held by entities with conflicting incentives. I have been auditing smart contracts for eight years. I know what happens when governance is opaque.
Core: The Governance Trap – A Lesson from DeFi’s Past Based on my audit experience, the real challenge is not technical — it is coordination and incentive alignment. Current AI wants to be a public good. Public goods, by definition, suffer from the tragedy of the commons. In blockchain, we solve this with token incentives, slashing conditions, and on-chain voting. Current AI has none of that. It has a foundation board, likely dominated by Google and French appointees.
I saw this exact dynamic play out in the 2020 Uniswap V2 liquidity mining experiments. We were all chasing yield, but the real yield came from understanding the depth of liquidity, not the APY. In AI infrastructure, the real value is not the open code — it is the trust in governance. If Google can influence which models get prioritized compute, or if France can demand content filtering to align with EU values, then the openness is an illusion. The board becomes the central point of failure.
Let me be precise: The $400 million is enough to build a decent cloud orchestration layer, but not enough to compete with AWS, Azure, or Google Cloud on raw compute. Current AI will rely on aggregating third-party resources — Google Cloud credits, European supercomputers, donated GPU cycles. This creates a multi-tenant risk: if one provider (say Google) decides to pull support or change terms, the entire infrastructure could stall. We have seen this in DeFi — lending protocols that depended on a single oracle collapsed when the oracle was manipulated. Current AI’s oracle is its funding and political will.
Contrarian Angle: The Open Infrastructure Might Increase Centralization Here is the counter-intuitive insight most announcements will miss. A “free” infrastructure layer with a centralized governance board can actually accelerate the very centralization it claims to fight. Why? Because small developers will build on Current AI assuming it is neutral, but when a conflict arises (e.g., a model generating hate speech, or a competitor wanting to use the same compute), the board will make decisions that favor its backers. This is not malice; it is the nature of concentrated power.
I have seen this pattern in blockchain. Ethereum was supposed to be a world computer, but after the DAO hack, the core team made a unilateral decision to hard fork. That was governance centralization. In the same way, Current AI’s “free” infrastructure could become a walled garden where the gates are invisible — open to all, but with subtle preferential treatment for Google’s products or French government priorities.
Moreover, the $400 million figure is strategically vague. How much is cash? How much is Google Cloud credits? How much is French tax incentives? In my 2024 ETF arbitrage strategy, I learned that the actual liquidity premium is hidden in the spread between on-chain price and ETF price. Similarly, the actual power of Current AI lies in the difference between its stated open nature and the real governance levers.
Data Point: Compare with HuggingFace, which has over 500,000 models and a commercial business. HuggingFace’s platform is open but the company is for-profit. Current AI is non-profit but backed by for-profit entities. Which one is more likely to stay open? Based on my experience with blockchain foundations, the non-profit ones often become the most political because they have no profit motive to override internal politics.
Takeaway: The Battle Is Over Governance, Not Code As a copy trading community founder, I have seen traders lose money chasing the next big narrative without understanding the underlying trust layer. Current AI is no different. The technology is secondary. What matters is who writes the rules, who can change them, and who pays the cost when something breaks.
Liquidity is just trust, digitized and leveraged. In AI, trust is governance, digitized and enforced. Until Current AI publishes its full governance charter, board composition, and conflict-of-interest policy, treat this as a speculative signal, not a technical breakthrough. The code may be open, but the real game is being played in boardrooms — and code alone cannot audit a budget.
We rode the wave until it broke our boards. Let's not let this wave break the open AI dream before it even starts.
The question remains: Will Current AI become a Linux for AI, or just another walled garden with open windows?
