InfoFi means information finance: using financial incentives to produce, assess or distribute information. In crypto conversations on X, the label also became associated with systems that measured authors' attention and connected participation to potential rewards. The broader idea includes markets designed to answer questions; the social-media version focuses on who attracts attention and how projects reward that contribution. Vitalik Buterin's explanation and Kaito's account of InfoFi illustrate those different uses.

The distinction matters when you read an author. A reward system can give someone a reason to explain a project well, and a reason to keep that project in your feed. You still need the author's evidence to assess a particular claim.

The examples below were checked on 6 October 2026. Kaito's old Yaps program and incentivized leaderboards have been discontinued. Its current documentation describes Kaito Studio, a marketplace for selected brand–creator collaborations. Historical guides to earning Yaps describe a model that has since changed. Kaito's closure confirmation, current Studio documentation.

What does InfoFi try to measure?

Start by separating three things:

  • Information: the claim or explanation in a post. For example, an author might describe how a protocol handles withdrawals.
  • Attention: the response to that material, such as engagement or a platform's estimate of the author's influence. A mindshare measure describes a share of attention within the conversation the service measures; its meaning depends on the service's method and coverage.
  • Reward: what a program makes available under its own rules. A score, eligibility decision and actual payment are separate stages.

These categories help you read a leaderboard. An account may be effective at getting a subject discussed while leaving an important technical claim unexplained. Conversely, a useful explanation can reach relatively few people. The score and the explanation answer different questions.

The broader term has a different example in Buterin's November 2024 essay: design a prediction market around a question, then let readers use the resulting probabilities as one information source. Participants have financial incentives; readers can examine the output without betting themselves. He also cautions against trusting the chart alone. That is a different mechanism from rewarding someone for bringing attention to a brand. From prediction markets to info finance.

A worked example: Kaito's February 2025 allocation

Kaito's early Yaps model makes the connection between attention and rewards concrete. Read it as a historical example.

On 17 February 2025, Kaito announced that its Yaps snapshot had been taken at 04:00 UTC. A snapshot fixes the inputs at a particular time; later posting does not change what was captured then.

Now open the Initial Community and Ecosystem Claim section of Kaito's tokenomics. It describes an initial allocation of 10% of the token supply across groups including early Yappers, Genesis NFT holders and ecosystem participants. Kaito said its assessment went beyond total Yaps. It included factors such as alignment with Kaito, long-term loyalty, participation and onchain reputation. The same document describes Yaps through activity, engagement and insight signals.

How the historical Kaito allocation used Yaps

February 2025 example

  1. Step 1

    Contributions and engagement

    Inputs measured by Yaps.

  2. Step 2

    Yaps snapshot

    Captured on 17 February 2025 at 04:00 UTC.

  3. Step 3

    Allocation assessment

    Yaps considered alongside other criteria.

    Condition Additional criteria included ecosystem alignment, participation and onchain reputation.

An attention score alone did not specify an exact allocation or establish whether a post was correct.

Conceptual reading of the published criteria; no individual score, allocation or payment is reconstructed.

The reading conclusion: If you were reading a participant's enthusiastic Kaito post, that possible incentive would be relevant context; the post's technical assertions would still need their own sources.

What changed on X in January 2026?

On 15 January 2026, Nikita Bier announced that X would no longer allow apps rewarding users for posting on X and had revoked those apps' API access. He attributed the action to AI-generated low-quality content and reply spam. The current X Developer Policy, under Pay to engage, prohibits using the API to provide monetary or virtual compensation for actions such as posts, likes and replies.

That day, Kaito founder Yu Hu announced the retirement of Yaps and incentivized leaderboards. His 16 January follow-up described Yaps as already closed. The change had a defined scope: it did not mean the whole Kaito business had shut down.

Kaito's current product overview still describes research, capital-formation and creator products. For the creator side, Studio's documentation describes matching brands with suitable creators and measuring campaign performance.

The Studio terms, updated 6 March 2026, make the newer arrangement tangible: eligible creators can submit price quotes for campaigns, brands choose whom to work with, and accepted creators enter separate agreements. Submitting a quote or content does not automatically entitle someone to a reward. That is a reason to check the current campaign and relationship instead of applying old Yaps instructions to Studio.

What does this change when you choose authors to read?

Give a ranking a limited job. It may help you discover an author or a conversation. To decide whether an explanation is useful, open its supporting documentation, data or argument. A model's assessment of influence cannot settle a claim about how a protocol works.

Identify what the author may receive. A campaign payment, a possible token allocation and a referral commission create different incentives. Look for the stated relationship near the relevant post. Participation alone does not prove that every opinion was purchased, and an unclear relationship should remain an open question.

Read the evidence separately from the enthusiasm. Several authors promoting the same project may be responding to the same program. Follow their sources: five posts that all repeat one announcement still provide one underlying announcement. Look for what each author adds, such as a test, a specific limitation or a well-supported disagreement.

For a completed example of assessing an author's actual X statement, continue with our crypto KOL guide. It follows a Base announcement from Jesse Pollak to the documentation behind the milestone, then assigns the account a specific purpose in a reading list.