Tiger Research: What AI services do cryptocurrency companies offer?

This report is written by Tiger Research. Cryptocurrency companies are generally facing "fear of missing out" (FOMO). From exchanges to security firms, they are all racing to launch AI-driven services. We will explore why they chose to take action at this time.

Key Points Summary

• Cryptocurrency companies in areas such as exchanges, security, payments, and research are synchronously launching AI services.
• Unlike previous cycles, companies like Coinbase and Binance, which have proven to be profitable, are leading the trend. AI has shifted from a theoretical concept to a practical necessity.
• The motivations for adoption vary across different industries: exchanges aim to prevent user churn; security companies aim to fill audit blind spots; payment infrastructure targets the emerging gig economy.
• Having a feature and actually using it are two different things. The "FOMO" and competitive pressure in the AI field are accelerating its application, far exceeding actual demand.
• Real demand and competitive anxiety are both at play. Distinguishing between value-creating adoption and mere labeling adoption is a key issue.

1. Cryptocurrency Companies are Offering AI Services

AI is the most attention-grabbing field in today's global market. General tools like ChatGPT and Claude have integrated into daily life, while platforms like OpenClaw have lowered the barrier to building intelligent agents.

Although the cryptocurrency industry missed this wave, it is now integrating AI across various verticals.

What AI services are these companies offering? Why are they entering this market?

2. How Cryptocurrency Companies Adopt AI Technology

2.1 Research

Cryptocurrency research has structural issues: on-chain data, social sentiment, and key indicators are scattered across various platforms, making verification difficult. General AI often returns inaccurate answers to cryptocurrency queries.

Projects like Surf address this issue by providing AI research tools specifically for cryptocurrency that can integrate scattered data sources. Among all cryptocurrency AI application scenarios, research has the lowest entry barrier for ordinary users, requiring no programming or trading expertise.

2.2 Trading

Exchanges are leading the application of AI in trading.

Methods vary. Some approaches directly expose proprietary trading data to users; others allow users to issue natural language commands to AI agents, which complete the entire process from analysis to execution.

Exchanges have been offering APIs for many years. The difference today is the addition of a layer: interfaces like MCP and AI Skills enable non-developers to access exchange functionalities through AI agents. Tools that were once limited to developers can now be accessed through natural language.

This aligns with the broader trend of community transformation. Non-developer users are increasingly building automated trading strategies through AI agents without writing any code. They simply describe the strategy, and the agent constructs and runs the algorithm.

For exchanges, this is both an opportunity and a challenge. As the number of AI users grows, user loyalty to a single exchange may decrease, as traders can execute trades anywhere. The reason exchanges are adopting AI is simple: to quickly attract users and keep them active on the platform.

Trading involves real asset management, requiring higher judgment and responsibility than research. However, as the entry barrier lowers, this field is also opening up to ordinary users.

2.3 Security/Audit

Traditional smart contract audits rely on manual line-by-line code reviews, which are slow, costly, and lack uniform standards among different auditors. Now, AI has been integrated into the workflow: AI first scans the code, followed by targeted deep reviews by human auditors. This improves speed and coverage without replacing auditors.

CertiK is a typical example. The company has previously faced criticism for audit projects that were later exploited maliciously. However, these incidents occurred outside the audit scope. Audits check the code at a specific point in time and do not include continuous monitoring.

CertiK uses AI to address this shortcoming. It adds real-time post-audit monitoring capabilities and publishes monitoring results through a public dashboard. Since the expanded monitoring scope is driven by AI rather than manual operation, both CertiK and the projects it audits benefit from this.

In the security field, the application of AI does not disrupt existing services but expands the scope of human work: improving accuracy during audits and filling post-audit blind spots. For blockchain security companies, AI is not a new business area but a tool to address existing security vulnerabilities.

2.4 Payment Infrastructure

AI agents need payment channels to participate in economic activities: for example, paying API fees, purchasing data, and buying services from other agents. For agents, the most natural payment method is an on-chain wallet paired with stablecoins.

Two models are emerging. The first is a universal protocol that embeds payments into HTTP requests, allowing agents to automatically settle on-chain when accessing paid APIs. The second is a payment plugin for specific agents, where agents can only execute payments within pre-set permissions and limits.

Payment infrastructure is the area most closely linked to stablecoins. However, since the payment entities are AI agents rather than humans, a fully operational model has yet to emerge.

Circle, the issuer of USDC, is also receiving attention. The company has released a proposal aimed at connecting its Gateway payment infrastructure with the x402 protocol and invited developers and researchers to review and contribute.

This is not a mature market, but the market has begun to digest this development trend. One of the key drivers of Circle's stock price increase is its AI agent payment model. The implementation speed of payment infrastructure will be slower than in the other areas mentioned, but it has become one of the most prominent macro themes in the current market.

3. Why Cryptocurrency Companies are Entering the AI Field Now

When ChatGPT was launched in November 2022, both AI and cryptocurrency were not yet mature. Although AI models were impressive, they could not reliably perform tasks. The cryptocurrency industry was severely impacted by the collapse of FTX and a widespread trust crisis.

Since then, AI has made rapid advancements. In the past year, the functionality and practicality of all mainstream models have significantly improved. In contrast, the cryptocurrency sector has merely "utilized" AI during the same period: filled with AI-labeled "Meme coins," underperforming AI agents, and marketing-driven hype. Decentralized AI infrastructure projects continue to emerge, but when objectively compared to equally capable native AI services, their quality is clearly lacking.

Today, the gap is widening further. In the AI industry, infrastructures like MCP (which enables agents to directly call external tools) and OpenClaw (which supports no-code agent building) have made the era of agents a reality. Meanwhile, cryptocurrency companies are just beginning to take action.

What sets this apart is who the actors are. No longer are they emerging startups claiming to be AI-driven, but established companies with mature profit models: Coinbase, Binance, and Bitget. These companies are launching AI services not for marketing purposes; they are driven not by immediate profits but by the fear of falling behind: FOMO.

The actions of Coinbase CEO Brian Armstrong exemplify this urgency. He issued a directive to all engineers to launch AI coding tools within a week and fired employees who did not comply.

However, it is also crucial to maintain a clear mind. Take trading automation as an example: agents can view prices and propose strategies, but how many users will truly trust agents to manage their funds for real-time trading? Moreover, has the x402 protocol really been applied in the real world?

Ultimately, the adoption of AI in the cryptocurrency field is not about chasing trends. With the arrival of the AI era, companies are actively taking action to avoid losing market position. Having a feature and truly utilizing that feature remain two different issues. But who is taking action is critical.

Imagine the AI industry as a swimming pool being filled with water. Those who jumped in before were merely pretending to swim. Now, those jumping in are former national team surfers. No one knows how high the water level will rise or whether this swimming pool will turn into an ocean. But cryptocurrency will not be drowned in the flood.