1. Future Miner Revenue Issues

Problem Background

To maintain the secure operation of the PoW system, miners need sufficient revenue.

Currently, over three million Kaspa can be mined per day (October 2025), but 93% of the coins have already been mined.

In 2029, only about 300,000 can be mined per day.

In 2032, only about 30,000 can be mined per day.

Kaspa's design is to subsidize miners through transaction fees in the future to maintain secure system operation. However, current transaction fees are too low, ranging from only a few thousand to tens of thousands of Kaspa per day.

Unless the price rises significantly or transaction fees increase substantially...

October 2025: Core Team Discussion

The core development team discussed the issue of transaction fees in the official group https://t.me/kasparnd/4796.

Core Topic: Discussion of a proposal to dynamically set the network's minimum transaction fee rate (min fee threshold) through a miner voting mechanism to address the blockchain network's security budget issue.

Summary of Main Points:

  1. Dynamic Fee Rate Proposal

    • Proposed by Michael Sutton: Let miners set the minimum transaction fee rate through additional data fields in blocks, with all nodes in the network (including miners) counting every 10 minutes and taking the 25th percentile of the voting values as the current minimum fee rate standard.
    • Purpose: To avoid fixing fee rates in code, allowing miners to adjust dynamically according to market supply and demand, while assuming miners are cooperative and not greedy, needing to reach consensus through mechanisms intelligently.

  2. Three-Stage Analysis of System Security Budget (elaborated by Yonatan Sompolinsky)

    • Stage One (Severe Shortage of Demand): Real transaction demand is far below network security budget requirements, and no mechanism (such as EIP-1559, smart fee estimation) can solve the security budget problem. Believes Kaspa may currently be in this stage.
    • Stage Two (Sufficient Demand but Imperfect Mechanism): Real demand is sufficient to cover the security budget, but existing mechanisms cannot incentivize users to reveal their true willingness to pay (not "incentive-compatible"). This stage requires dynamic fee discovery mechanisms (such as proposals or EIP-1559) to optimize.
    • Stage Three (Demand Exceeds Supply): Network is healthy, no complex mechanisms needed, regular bidding (such as "highest bidder wins") can meet demand.

  3. Basis for 25th Percentile

    • To prevent individual miners from accepting low-fee transactions for profit, triggering a "race to the bottom" vicious cycle, the P2P network needs to follow the voting results of the majority of miners (such as 75%).
    • If the threshold is set at 90%, only 11% of miners can disrupt coordination; 50% is safe but easily manipulated. Choosing the 25th percentile (i.e., the standard supported by 75% of miners) strikes a balance between preventing malicious price suppression and resisting manipulation.

  4. Significance of Discussion

    • Even if currently in Stage One, exploring Stage Two mechanisms in advance helps with market price discovery, testing users' true willingness to pay, and preparing for future transitions.

Following the previous discussion, here are three additional distinct solutions proposed by Yonatan.

Solution Three: Implement Useful Proof of Work

  • Core Content: A completely independent solution is to introduce useful proof of work in future planning.
  • Path and Conditions:
    • This is a long-term solution.
    • Need to first observe and learn from the pouw team's practical results and successful experiences, then adapt them to the DAG (Directed Acyclic Graph) structure.
    • Such major changes require absolute majority consensus from the community to implement.

Solution Four: Create New Token Standard (KRC-21)

  • Core Content: Another distinct approach is to invent a new token standard—KRC-21—for real DeFi tokens (not meme coins) on L2 or built-in L2. Tokens designed with this standard will generate meaningful inflation (equivalent to taxation) and pay it to miners to support L1 security.
  • Implementation and Vision:
    • This standard will not be enforced through the consensus protocol, but believes it can form a strong social consensus.
    • To better support Kaspa, hope such tokens can mandatorily require exchange with Kas first through public AMMs or CLOBs.
  • Background Inspiration: This idea originates from Ethereum community discussions on how L2 can support L1 security and achieve interest alignment.

Solution Five (Most Strategic Core Work): Promoting Kaspa Development Stage Upgrades

  • Core Content: The most important and most strategic work is helping Kaspa advance from its current stage (from Stage One to Stage Two, then to Stage Three).
  • Current Status and Outlook:
    • This aligns with Kaspa's established development trajectory.
    • Many Kaspa community members (including the speaker) are working towards this.
    • Expect to share more details in the coming weeks.

2025-10-12: FreshAir08's New Proposal

Background

FreshAir08, a core team member of Kaspa, mentioned today (2025-10-12) his recent research with Yonatan:

(1). Article Translation

Translation:

Over the past few days, let me briefly explain what @hashdag (Yonatan) and I have been researching:

A 2017 paper by Lavi, Sattath, and Zohar proposed an alternative fee mechanism for blockchains. At first glance, its core idea seems counterintuitive—miners charge fees from all transactions in their blocks equal to the fee of the lowest-paying transaction included in that block (even for those transactions that agreed to pay higher fees). The idea is to incentivize miners to optimize "transaction count × minimum value transaction," thereby naturally excluding those transactions that pay fees far below other transactions. It can (roughly) be seen as a smart and adaptive implementation of "minimum fee." The ideal effect of this mechanism is that it incentivizes users to pay higher fees to miners in a way that is independent of block size (i.e., even when the mempool is smaller than a single block's capacity).

First, it should be noted that, to my knowledge, this solution has never been implemented in the real world.

Secondly, the above holds true for blockchains. One of the challenges lies in how to apply this to DAGs. Before proceeding, I want to clarify a common misconception about Kaspa:

---------------Interlude: Inclusive Blockchain---------------

Kaspa's high throughput is not primarily derived from transactions in parallel blocks—even transactions not selected for the chain are discarded, Kaspa still has greater capacity than traditional blockchains simply because it allows higher block production rates for a given block size (...for given security parameters and minimum hardware requirements, etc.). However, as the number of parallel blocks increases, this simple effect diminishes—if you discard transactions from parallel blocks, throughput is wasted. This throughput can be recovered through a solution called "inclusive protocol." In short, this specifies an order for parallel blocks, where all transactions that do not conflict with previous transactions are recorded in the ledger as usual. Theoretically, we can now let every block achieve its maximum throughput.

But why don't all parallel blocks contain the same transactions, thereby wasting our theoretical gains? Well, if they all choose the same transactions, sometimes they'll be very successful, but most of the time they'll get nothing. Then, miners who choose transactions that other miners haven't chosen will, on average, do better! Even if those transactions pay lower fees. Long story short, the finding is that for them, the "correct" ("equilibrium," with about five asterisks for caveats) approach is to randomly select transactions, but weighted by the transaction fees paid. A lot of simplification has been done here, specifically to note. Coincidentally, this also tends to bring good throughput to the system. This is currently the case with Kaspa, but I emphasize again, this is not the only source of Kaspa's high throughput.

---------------------End of Interlude---------------------

So, basically, although theoretically we could take a step back and abandon "inclusiveness" (sacrificing some valuable throughput), we hope to combine this new fee mechanism with insights from the inclusiveness protocol currently used by Kaspa.

After adopting the new fee mechanism, the equilibrium state will change, and as miners adopt different strategies, the original guarantees about user behavior also need to be re-examined—so once we've established the miners' equilibrium, we also need to explore the users' equilibrium.

We hope that after completing all analyses, we can discover that either we can maintain similar user behavior guarantees as in the original paper, or we will attempt to find a variant of the original solution to achieve this.

By the way, if people with relevant background also want to research this problem, we also very much welcome it. Like everything else, long-term decentralization of research is also a goal.

Two final points:

  • a) Any change to fee mechanisms, no matter how good, cannot replace the importance of increasing demand. This effort should proceed in parallel with building new utility features for Kaspa (such as vProgs).

  • b) This is a medium-term solution (but I believe this is serious research that should have been done long ago) and requires a hard fork. Some modification suggestions that can be implemented immediately have been proposed within the community. These suggestions may lack theoretical basis in the long run, but are usually harmless—for example, miner-adjustable minimum fees. Such solutions often fall into the prisoner's dilemma, and as I said, I believe fee policy alone cannot change the big picture currently—but if someone wants to implement this, I personally won't oppose it because it's harmless. In my opinion, this sounds like a good entry-level project for getting into Rust version of Kaspa development.

(2). Comparison with Kaspa's Current Fee Mechanism

Compared to Kaspa's existing mechanism, the core advantage of this paper's mechanism is: it can intelligently and adaptively establish and maintain a "market minimum fee" standard in a decentralized system, thereby potentially increasing miners' overall revenue and improving the network's security budget.

  1. Creating a "Fee Floor" to Resist "Tragedy of the Commons"

    • Current Mechanism Problem: When block space is sufficient, miners have strong incentives to include any transaction with fees greater than zero, because it's better than nothing. This easily leads to fee competition approaching zero, forming a "prisoner's dilemma" that weakens network security.
    • Paper Mechanism Solution: This mechanism changes miners' incentives. Miners no longer simply pursue filling blocks, but pursue maximizing (block capacity × minimum fee in block). This incentivizes miners to ignore those transactions with excessively low bids, thereby naturally forming a "fee floor" in the market.

  2. Incentivizing Users to Pay Higher Fees

    • Current Mechanism: If users only want to be included in blocks without acceleration needs, they have strong incentives to pay extremely low fees.
    • Paper Mechanism: If the fee paid by users is below the current market's "floor price," their transactions will be ignored. If users want to be included, they must pay fees at least equivalent to the current "floor price." This directly pulls users into the system of maintaining network fee levels.

  3. May Be More Suitable for Future High Throughput, Low Congestion State

    • If Kaspa successfully achieves its high throughput goals, the network may be in a non-congested state most of the time. Under the current mechanism, this easily leads to transaction fees approaching zero. The paper mechanism precisely provides a solution for this "successful scenario," maintaining a meaningful fee market even when block space is not scarce, providing continuous support for network security (security budget).

Although the paper mechanism has theoretical advantages, Kaspa's current mechanism also has strong vitality due to its simplicity and efficiency.

  1. Simple and Effective, Easy to Implement

    • The current mechanism is very intuitive, does not require changing user behavior patterns (users still bid according to the "highest bidder wins" intuition), and is easier to implement in complex DAG structures.

  2. High Resource Allocation Efficiency During High Demand

    • When transaction demand is vigorous and the mempool is full, the current mechanism can effectively allocate block space to those transactions with the highest willingness to pay, which is itself an efficient market behavior.

  3. Good Synergy with "Inclusiveness Protocol"

    • As pointed out in the article, the current fee-weighted random selection strategy happens to incentivize miners to choose different transactions, thereby perfectly utilizing the DAG's inclusiveness protocol and recovering potential throughput losses caused by parallel blocks. This is an "unintentional but very beautiful" fit. Changing the fee mechanism might break this delicate balance.

(3). Related Papers

  1. Potentially More Suitable for Future High Throughput, Low Congestion State
    • If Kaspa successfully achieves its high throughput goals, the network may be in a non-congested state most of the time. Under the current mechanism, this easily leads to transaction fees trending toward zero. The paper mechanism precisely provides a solution for this "successful scenario," maintaining a meaningful fee market even when block space is not scarce, providing continuous support for network security (security budget).

Although the paper's mechanism has theoretical advantages, Kaspa's current mechanism also has strong vitality due to its simplicity and efficiency.

  1. Simple and Effective, Easy to Implement

    • The current mechanism is very intuitive, doesn't require changing user behavior patterns (users still bid according to the "highest bidder wins" intuition), and is easier to implement in complex DAG structures.

  2. High Resource Allocation Efficiency During High Demand

    • When transaction demand is strong and the mempool is full, the current mechanism can effectively allocate block space to those transactions with the highest willingness to pay, which is itself an efficient market behavior.

  3. Good Synergy with "Inclusive Protocol"

    • As the article points out, the current fee-weighted random selection strategy恰好 encourages miners to choose different transactions, thereby perfectly utilizing the DAG's inclusive protocol and recovering potential throughput losses caused by parallel blocks. This is a "serendipitous fit" but very wonderful. Changing the fee mechanism might break this subtle balance.

(3). Related Papers

Last Update: