A Tale of Two Themes: AI & Tokenization

Introduction

As the digital asset ecosystem matures, investors are seeking more nuanced and targeted ways to gain exposure to specific trends—particularly themes that drive innovation across blockchain networks.

In this report, we explore how Kaiko’s Thematic Indices offer a unique means of tracking emerging trends in digital assets, namely AI and tokenization.

Inside the Index: What Makes a Theme?

Unlike our other index products, which offer broad market exposure or access to well-established trends, the Thematic Indices focus on assets that serve as foundational infrastructure in emerging applications.

The nascent nature of these markets makes it challenging to define assets by theme, as there can be overlap and noise among projects. Kaiko’s Thematic Indices are designed to meet demand while offering a curated selection of assets that truly align with key narratives.

Both the Kaiko AI and Tokenization Indices use our standard ranking approach. Eligible assets are ranked according to liquidity and market capitalization, with each criterion weighted at 50%. The weighting for individual assets is capped at 30%.

Catching the AI Hype Train

Artificial Intelligence has been a major theme across all markets in recent years. It has even permeated everyday life, as more and more people interact daily with large language models like ChatGPT or Claude. However, defining what constitutes an AI-related crypto project has become increasingly difficult, as token projects seek to ride the tailwinds of Nvidia-driven euphoria.

The Kaiko AI Index tracks a basket of the seven leading digital assets at the intersection of AI and blockchain technology. Constituents are chosen based on our internal vetting and taxonomy process, which is discussed in more detail below. As of the latest rebalance at the beginning of April, the assets included are: NEAR, ICP, FET, RENDER, GRT, AIOZ, and AKT.

These assets were deemed suitable for the Kaiko AI Index based on market consensus and our own research. Essentially these are assets that integrate AI with blockchain technology in a meaningful way.

We place a particular focus on innovations in decentralized AI, distributed computation, and AI-enhanced blockchain applications. An asset must satisfy one or more of the following criteria for inclusion in this thematic index.

• Lead in integrating AI with blockchain technology.
• Focus on enabling decentralized AI or distributed computation.
• Be widely adopted with significant market activity in AI-driven blockchain applications.

Tokenize This, That, and the Other Thing

While a less popular narrative in the broader economy, tokenization has become a dominant topic of discussion within crypto-native circles in recent years. This may be due to the promise of tokenization and its potential to serve as a “Trojan horse,” enabling the widespread adoption of blockchain technology in everyday finance.

In the most basic sense, tokenization is the process of creating a digital version of a real-world asset. This can apply to anything from real estate projects to treasury bonds, with the latter currently being the most popular use case. At present, billions of dollars’ worth of U.S. Treasury bills have been tokenized across more than a dozen blockchains.

The current construction of the Kaiko Tokenization Index focuses on assets that enable the tokenization process. This includes assets that finance the infrastructure and technology facilitating these transformations. An asset is eligible for inclusion in the index if it meets one or more of the following criteria:

• It is a market leader in the tokenization of real-world assets.
• It has a primary focus on enabling the tokenization of real-world assets.
• It serves as a means of payment for infrastructure and technology targeting the tokenization of real-world assets.

Macro Headwinds Weigh on Thematic Trends

The Kaiko Tokenization Index posted returns of nearly 28% over the past six months, while the AI Index shed 31% of its value during this period. The downturn in the AI Index mirrors the broader market decline, as sentiment around the emerging technology has worsened.

As such both indices are down dramatically since January, with AI down 45% and Tokenization down 30%. This coincided with the reality of Trump’s second term setting in as volatility ticked higher and risk was repriced across markets.

(Deep)Seeking Answers for AI’s Decline

The decline in AI-related digital assets reflects overall market sentiment and mirrors the performance of the world’s largest AI-related stock, Nvidia.

Jensen Huang’s Nvidia is down over 20% in the past six months, as trade tensions and political uncertainty have created additional risks for investors in the leading AI chip producer. Uncertainty driven by Trump-era tariffs persists, along with his stated desire to keep the U.S. out of any major global conflicts.

Another major risk factor for AI-related assets was the release of Deepseek in January, which spurred fears that the China-based firm would radically change the outlook for all AI companies with its low-cost, efficient product.

Furthermore, since the DeepSeek-driven volatility, the returns for the AI index have failed to reach previous highs. In fact, since the DeepSeek-driven sell-off, the AI index hasn’t had a single daily return above 10%. Its losses have also exceeded 10% more times since January 27 than in the previous eight months.

Clearer Outlook for Tokenization

Despite the broad decline in markets the risk profiles for both thematic indices are quite different. The Kaiko Tokenization index offers better risk-adjusted returns at present. Its Sharpe ratio is around 0.9 currently, while the AI index has a negative Sharpe of 0.37. These ratios are calculated on a rolling six-month basis, hence the tokenization index mandates a positive Sharpe in spite of recent declines.

Conclusion

While AI has dominated headlines in recent years, the major shift in markets since the beginning of Trump’s second term has massively affected prices.

One reason for tokenizations relative outperformance compared to AI is related to it’s use case and its potential for change. While both themes can have radical proponents, tokenization is by its nature less of a widespread narrative and as such major shifts in market sentiment don’t have as pronounced an affect on prices.

Tokenization is about market structure and improving on current processes, its selling point is somewhat less radical than AI’s which proponents often claim will change how everything in day-to-day life works. The more radical a theme, the more prone to drastic changes in sentiment.