The Move to AI in Cryptocurrencies: Hype, Infrastructure, and a Two-Year Countdown

If Conses Hong Kong 2026 had an unofficial theme, it wasn’t bitcoin or regulation. It was artificial intelligence – and the research to find out what it really means for digital currencies.

AI has appeared in almost every context: main stage keynotes, side event sessions, venture capital meetings, and even the post-conference atmosphere. But the speeches were not unified. They ranged from government officials in Hong Kong supporting the machine economy to venture capitalists declaring that the cycle of hype around artificial intelligence in cryptocurrencies is over.

Sponsored

Sponsored

The company’s artificial intelligence agents have already been implemented

At the Gite side event, Sophia Jin, Hong Kong CTO at… ByteIn addition – the company’s technology arm of ByteDance – several major cryptocurrency exchanges are already using the company’s AI agent products. He demonstrated three production use cases: intelligent customer service that integrates deep research and business scenario matching; multi-agent research systems with parallel data collection; Automate AML workflows with human oversight at decision points.

The most notable detail was the safety engineering. BytePlus puts guardrails outside the agent orchestration layer – a kill switch that can immediately shut down agents if they violate specified limits. Jin predicts that within two years, every exchange employee will have an enterprise-level AI assistant, while onboarding new users will become much easier with AI personalized learning.

Two years until the AI ​​thinks about you

Ben Goertzel, CEO of decentralized AI marketplace SingularityNET, offered more Hours Excitement for the conference. It gave humans about two years before AI surpassed them in strategic thinking.

“The human brain is better at taking the imaginative leap to understand the unknown,” Goertzel said in the consensus. But this will not last. “We’re going to enjoy it for a few more years.”

While his Quantum project can predict Bitcoin’s short-term fluctuations with high accuracy, Goertzel noted that long-term strategic thinking remains uniquely subjective — until now. He described the current bear cycle as a “stress test” for the infrastructure that will eventually host AGI.

foot Bitget CEO Gracie Chena more realistic point of view. In a panel discussion on proxy trading, I compared existing AI trading bots to apprentices – faster and cheaper, but requiring supervision. She noted that historical data-driven models did not encounter events such as 10/10 liquidations, making human intervention necessary in unfamiliar circumstances. But in three to five years, he predicts, AI could replace many human roles.

Sponsored

Sponsored

Saad Ng, CEO of proxy trading startup PiP World, responded that humans may not be the right foundation. “As humans, we are very emotional. We cannot compete with AI solutions,” he said, noting that 90% of day traders lose money.

Build a payment layer for agents

If the main stage provides visibility, the side events try to build the plumbing.

At the Stablecoin Odyssey event in Soho House, the session “Building Blockchain Payments for the Agent Economy” will focus on the infrastructure that AI agents really need. Nellie Tan, Head of Payments at Monad, presented: X402 protocol from Coinbase — an inherently on-chain payment standard over HTTP — and argued that proxy payments generate transactions at “data speeds,” requiring a transfer rate of thousands to millions per second.

Eddie, CEO of payment broker AEON, described this change as a transition to the interface. When consumers interact via AI agents rather than apps, every business interaction goes through a single point – and the last mile is always a payment. His company processes what he described as 80% of cryptocurrency payments through partnerships with OKX, Bybit and others.

Sponsored

Sponsored

The question of which blockchain AI agents could choose remains open. Matt Tokay, CMO of OP_CAT Layer, noted that no one knows yet whether agents will choose chains based on training data, experience, speed or security. The answer depends on the transaction – large asset transfers prioritize security, while consumer purchases prioritize speed.

Are cryptocurrencies the currency of artificial intelligence – or just another hype cycle?

The most notable endorsements come from outside the industry. Hong Kong’s Finance Minister Paul Chan Moo-po used his appearance to represent AI agents as an economic force that cryptocurrencies are uniquely positioned to serve.

“With AI agents able to make and execute decisions autonomously, we may begin to see the first forms of what some call the machine economy, where AI agents can hold and transfer digital assets, pay for services, and transact with each other throughout the chain,” said Chan.

Binance CEO Richard Ting pushed the matter further. “If you think about agent AI, like booking hotels and flights and any purchases you make, and how you think those purchases will happen — it’s going to be through cryptocurrencies and stablecoins,” he said. “So cryptocurrencies are the currency of artificial intelligence, if you think about it.”

Sponsored

Sponsored

But venture capitalists have focused on the broader narrative of “AI + cryptocurrencies.” Canonical Crypto’s Anand Iyer described the moment as a low-key event. “We went through a period of foam. Now it’s about knowing where the real strength lies,” he said. Both Ayer and Kelvin Kuo of Spartan Group criticized the overinvestment in GPU markets and attempts to build decentralized alternatives to OpenAI or Anthropic – projects that require much more capital than cryptocurrencies can provide.

Instead, both see the potential in tailored solutions that start with a specific problem. Proprietary data, regulatory boards or market advantages are now more important than technical innovation. Kuo’s advice to the founders was simple: “Twelve months ago, just having a ChatGPT shell was enough. This is no longer true.”

What is formed?

Conversations among industry participants indicate that a framework is taking shape: stablecoins that act as a barrier of value for agent transactions, prediction markets that process price information, artificial intelligence systems that execute trades and operations, and physical robots that extend the loop into the real world. It is not a single project or protocol – it is a thesis about where cryptocurrencies and AI intersect productively, without relying on the speculative cycles that have driven previous revolutions.

A parallel thread runs through decentralized AI. Current systems are centralized and opaque. The idea of ​​transparent and verifiable AI networks governed by communities aligns with the founding principles of cryptocurrencies – and Goertzel, among others, pointed to the growth of such projects at the event as evidence that convergence is underway.

The cycle of pure speculation cannot return. But at the Consensus Hong Kong conference, the argument that AI gives cryptocurrencies a raison d’être was presented at the same time by a government platform, an exchange and a venture capital meeting. This is a different kind of consent.