OKX Europe chief predicts an AI-native “super currency” bigger than any fiat-backed token
The next trillion-dollar currency in crypto may not be a stablecoin at all, according to Erald Ghoos, chief executive of OKX Europe. In comments reported by CoinDesk on 29 August 2026, Ghoos argued that the industry should be preparing for an AI-native monetary instrument purpose-built for autonomous agents, one he suggested could dwarf every currency in existence today.
“There will be an AI currency coming,” Ghoos told the publication. “This is not going to be fiat, for sure.” He described the eventual instrument as “a stablecoin or some other crypto token” and predicted it could become “by far, by far, the largest currency that this world has ever seen.”
The remarks land at a moment when the stablecoin market is enjoying unprecedented institutional acceptance, with USDT and USDC entrenched as the dominant settlement assets across exchanges, payment corridors and increasingly mainstream finance. Ghoos’s claim inverts that orthodoxy. In his framing, the largest onchain currency of the next decade may be defined not by dollar pegs or issuer branding but by function: fast settlement, predictable value and compatibility with autonomous software.
Crucially, Ghoos did not claim to know what the instrument will look like. He predicted a single “super currency” for AI agents but stressed that its final form remains an open question. “It could be a stablecoin, or it could be something else,” he said. “Let’s see what works.”
Why machines, not humans, may drive the next monetary layer
The argument underpinning Ghoos’s forecast is structural rather than speculative in the ordinary sense. AI agents, not only human users, are positioned to become major transactors on blockchains. Where human commerce is bounded by wallets, apps and working hours, autonomous software can negotiate, pay and settle continuously, at machine speed, across counterparties that never sleep.
That creates a demand profile that today’s stablecoin model only partially serves. Machine-to-machine commerce requires an asset that settles instantly, holds predictable value between the initiation and completion of a transaction, and can be held, escrowed and programmed by software without human intervention. Dollar-backed tokens satisfy the value-stability requirement by design. Whether they are optimal as native money for agents, rather than as a human-facing proxy for dollars, is precisely the debate the CoinDesk piece reflects.
The scale of the opportunity is what gives the argument force. AI systems could eventually handle trillions of dollars in payments, trading and services. Whoever supplies the settlement asset for that flow would not simply be running another token. They would sit on foundational infrastructure for crypto, fintech and AI simultaneously, a position arguably more strategically significant than any single exchange, custodian or payment network occupies today.
This explains why the framing matters beyond one executive’s opinion. The stablecoin wars of the past decade were fought over human demand: trading collateral, remittances, dollar access in emerging markets. A machine-driven monetary layer would be contested on entirely different terms, including latency, programmatic verifiability, fee predictability at high frequency, and resilience under conditions where no human is watching the transaction at all.
The stablecoin incumbents face a category question, not a competitor
For the incumbent issuers, the immediate commercial threat is limited. Nothing in Ghoos’s comments suggests dollar-backed tokens are about to be displaced from their core human-facing use cases. The question posed is subtler: whether machine commerce will grow into a distinct category with its own native asset, expanding the market beyond USDT- and USDC-style products into something that has not yet been clearly named.
That is not how stablecoin competition has been understood until now. The prevailing assumption, reinforced by legislative progress on stablecoin regulation in major jurisdictions and by the entry of banks and payment giants into tokenised deposits and fiat-backed tokens, has been that the market consolidates around regulated dollar instruments. Ghoos’s comments suggest a second trajectory running in parallel: an instrument that may borrow stablecoin mechanics, such as price stability and redemption assurance, without being a stablecoin in the regulatory sense at all.
The distinction carries real regulatory weight. If the AI super currency is “not going to be fiat”, as Ghoos put it, it falls outside the dollar-pegged frameworks that legislators have spent years constructing. Regulators would face an asset designed for high-velocity machine settlement whose issuer, reserve structure and governance model may not fit existing categories. For firms building in the space, that ambiguity is both the opportunity and the risk: a new asset class with no settled rulebook.
It also reframes the competitive landscape for exchanges and infrastructure providers. OKX, through its European leadership, is signalling early positioning in a market that does not yet exist. The bet is that whichever asset becomes the settlement layer for autonomous agents, the venues and wallets that support it early will capture a disproportionate share of the resulting volume.
Readers tracking the broader evolution of onchain money can follow developments in our stablecoin coverage, where the machine-economy debate is increasingly shaping issuer strategy.
Function over branding: what the next monetary layer may look like
The most consequential idea in the CoinDesk piece is that the next major onchain monetary layer may be defined less by branding than by function. The dollar won the stablecoin era partly through network effects and the institutional credibility of its issuers. An AI-native currency would succeed or fail on engineering criteria.
Fast settlement is the first requirement. Agents operating at machine speed cannot tolerate the friction of assets whose finality is uncertain or whose transfer costs are unpredictable. Predictable value is the second: an agent holding balances between tasks needs assurance that the asset will not drift materially in the interval, which is why some form of stability mechanism, whether peg, algorithm or reserve, seems likely. Compatibility with autonomous software is the third and least solved: the asset must be operable by code alone, with permissions, limits and recovery paths designed for counterparties that have no customer support hotline.
Ghoos’s own hedge, that the winner “could be a stablecoin, or it could be something else”, is the honest read of the situation. The industry has a poor record of predicting the form factor of winning monetary instruments. The stablecoin itself began as a trading convenience and became systemic infrastructure. An AI currency could follow the same path in compressed time, or it could emerge from a design space nobody is currently watching.
What seems harder to dispute is the direction of travel. Autonomous agents are already executing onchain transactions, and their share of activity is set to grow. Money follows usage, and usage is migrating toward machines.
Closing analysis
Ghoos’s prediction should be treated as a thesis, not a forecast with a date attached. No asset currently in circulation plausibly qualifies as the “largest currency the world has ever seen”, and the distance between a compelling conference argument and a functioning monetary layer is measured in years of engineering, liquidity formation and regulatory negotiation.
Yet the thesis deserves attention precisely because it comes from the leadership of a major exchange, the layer of the stack that sees transaction demand before almost anyone else. If machine-to-machine payments scale into the trillions, the settlement asset for that flow becomes the most strategically important piece of infrastructure in digital finance. Whether it is a stablecoin, a token with an as-yet-uninvented structure or something between the two, the contest has effectively been declared open. The market beyond USDT and USDC may not yet have a name. According to the CEO of OKX Europe, it may not need one for long.
