)
Over the last three years, the artificial intelligence (AI) narrative has been dominant. Naturally, discourse around this topic is followed closely by the nuclear energy required to power data centres and its sustainability. Nevertheless, AI’s momentum has pushed out the environmental narrative, as governments utilise public-private partnerships (PPPs) to drive AI integration across all sectors.
This is plain to see with the Trump administration. Case in point, shortly after President Trump’s inauguration, the Stargate Project – in conjunction with OpenAI – was announced at the White House, worth around $500 billion USD. Likewise, when President Trump headed the tech/defence delegation into Saudi Arabia, $600 billion USD worth of commitments were made. Within months, the Oval Office had splashed a trillion dollars on AI projects.
The AI narrative is powerful because it is seen as the ultimate tool towards automation and amplification of force. This is also why it is seen at the forefront of a geopolitical race against China.
But where is the blockchain narrative now? And can it fit into the fast-evolving AI ecosystem?
At a glance, these two narratives clash. AI is all about centralized data centre infrastructure that makes it possible to cost-effectively deliver computations, be they in text, image or video output. In contrast, blockchain is all about running processes on a decentralized ledger to create a trustless economy.
Yet, digging in deeper, it is easier to see why blockchain plus AI makes sense.
AI: Ushering in the Age of Fakery and Slop
Following the launch of Google’s Veo 3 text-to-video generator, it is becoming even clearer that the online space will suffer severe erosion of trustworthiness. When AI generates videos in a style resembling handheld shaky cams, even authentic footage can be labelled as ‘AI’.
Accordingly, it becomes easier to dismiss anything, as it is reasonable to assume that anything not directly viewed in physical reality is a mere manipulation of pixels. It is then easy to see why trust would become high in demand. And there is no better trust-facilitator than a trustless blockchain network. Specifically, blockchain can be used to:
Create immutable records as a reference point of traceability and auditability. This directly translates to content provenance.
Create smart contracts to automate data validation bounties and compute marketplaces for AI workloads. For example, Story Protocol (IP) tokenises and automates royalties and licensing permissions.
Create a ledger resistant to tampering and to single points of failure.
Incentivise participation in an automated, decentralized manner.
How would these blockchain features manifest exactly in a cohesive ecosystem? First, let’s consider the implication of AI models constantly churning out content for users. Over time, this means that the available public dataset pool for their training becomes polluted.
In a 2023 paper titled “Nepotistically Trained Generative-AI Models Collapse”, researchers found that even small amounts of regurgitated AI content causes rapid degradation of quality outputs. Even more worryingly, that “this distortion extends beyond the text prompts used in retraining, and that once affected, the models struggle to fully heal even after retraining on only real images.”
One way to circumvent this rising problem is to monetise private data. But that is only the first step.
ASI: The All-In-One Tokenised AI Ecosystem
The first step in ensuring future AI model training remains coherent is to incentivize the selling of private data not sourced through web-scraping. Ocean Protocol does exactly that, with its marketplace driven by OCEAN token:
On the automated market maker (AMM) called Ocean Market, in the vein of Uniswap, data sellers mint tokens to sell off-chain datasets.
Through asks and bids via smart contracts, customers buy these tokens to gain access to datasets.
And because the process is tokenised (encrypted on Ethereum), the sellers’ personal information is safeguarded, which further incentivizes people’s participation. Complementary to Ocean is Fetch.AI (FET) and SingularityNET (AGIX) protocols.
While Fetch.AI monetizes AI agents to perform tasks on behalf of users and dApps, SingularityNET is a general-purpose network to monetize AI services.
In late March 2024, the developers of all three protocols agreed to combine FET, AGIX and OCEAN into a single Artificial Superintelligence Alliance (ASI) token to streamline the fusion of AI and blockchain.
CUDOS protocol (CUDOS) also joined in as a cloud computing platform for Web3 deployment. For optimal exposure, users should visit ASI calculator to understand the conversion rates of all four tokens – OCEAN, AGIX, CUDOS, FET – as well as migrate existing tokens to Artificial Superintelligence Alliance.
On major centralized exchanges, ASI is available as FET, given its 1:1 FET/ASI conversion ratio. Year-to-date, FET is down 47%, which may present a solid opportunity for exposure in the scenario of a renewed altcoin season – an outlook shared by several leading investment newsletters.
Is It Possible to Profit from Tokenised AI?
If AI models are properly trained, the sky is the limit. The aforementioned Fetch.AI is all about autonomous economic agents (AEAs) scouring Web3 to make optimal profit-making and profit-taking decisions… among other things.
This would not be possible in TradFi, consisting of isolated, non-blockchain networks with heavy involvement of human intermediaries. But across dApps, everything boils down to interaction with smart contracts. Consequently, this leads to an arms race between different AI agents.
The general exposure to such an arms race comes from Virtuals Protocol (VIRTUAL), available on two major blockchain networks – Ethereum via Base and Solana. The VIRTUAL token democratises the ownership of deployed AI agents, wherein each AI agent is tokenised and paired with VIRTUAL in liquidity pools.
Whatever service the agent provides, such as LUNA social media influencer, VIRTUAL token pays for it, with churned revenue pooling back into the agent’s treasury. Of course, the more evolved and adopted Web3 gets, the greater the potential for AI agents to be deployed, interact with dApps and generate revenue.
In addition to ASI/FET, this positions VIRTUAL as one of the key AI tokens to consider for exposure. Year-to-date, VIRTUAL price is down significantly. But, if altcoin season is to take off by the end of 2025, timely market entry into these AI tokens could present bigger potential gains than the heavily capitalised Bitcoin.