Go back button

Upvest secures €900k in funding from the IBB ProFit program to solve the transaction fee problem on Ethereum

Fee estimation on Ethereum is cost-inefficient:

Transaction fees are a complex topic that platforms have to deal with when building applications on Ethereum. On the decentralized network, the execution speed of a transaction is dependent on the network traffic at a given time, as well as on the transaction fee, which is sent along with the transaction. Thus, setting a fixed fee for every transaction will likely result in an over- or underfunding of the transactions. Overfunded transactions add up in terms of cost, especially when the amount of transactions sent is very high. Underfunded transactions, on the other hand, come with the risk of being dropped, so never included in mined blocks. This is a problem since the transaction is stuck in the mempool, and it requires signing another transaction to replace the underfunded one. 


As of today, existing approaches to fee recommendations are static and do not come with an enterprise standard (SLA). Upvest has built a fee recommendation engine to cover this missing piece of software in the Ethereum ecosystem.


Fee estimation neural network:

Upvest developed a fee recommendation API based on a continuously evolving machine learning model. We are constantly collecting real-time network signals from numerous independent data sources, including for example the number of unconfirmed transactions or the number of active miners.

Our currently deployed model has been trained with Ethereum network signals gathered over a period of one month. The Initial benchmarking results against the current fee recommendation solutions promises crucial performance improvements by 18%.

“Fee estimation using Neural Networks (FeNN) Solution”


I am managing transactions on Ethereum. Why should I use it?

As explained above, the Upvest fee recommendation engine is designed for startups and companies that are executing a large number of transactions on the network. These companies are dependent on a meaningful fee recommendation system to optimize their operating costs. Moreover, the transaction execution speed is optimized while locking up capital is avoided entirely. It is relevant to highlight that these benefits gain importance for an increasing transaction count and/or transaction volume.


Upvest offers the service in the form of an easy to use API, which does not require any blockchain specific knowledge. You get the execution speed that you pick, no need to think about fees. The product is fully integrated into the Upvest product suite and comes with an enterprise-grade SLA.


How does it tie in with Upvest’s product suite:

Our clients are using the fee recommendation engine in the context of primary security token issuance, and for transactions in the secondary market between investors. Since the technology is use-case agnostic, it can easily be used by exchanges as well.


What is next?

To complete the Upvest transaction toolbox, we are currently working on a transaction confirmation risk calculator to account correctly for incoming Ethereum transactions. By recommending the optimal waiting time for block confirmations, our customers are able to account for these transactions as early as possible, while mitigating their risk.

Upvest is also a keen supporter of the open source community, and once this project has reached maturity, large parts of the solution and codebase will be shared with the community.


How can I participate and get access:

We are currently optimizing the fee recommendation engine, and also looking into building a tool for delivering insights on transaction confirmation levels. If you’re an engineer and this sounds interesting to you, please reach out to us to contribute. If you are a corporation that would like to try the fee recommendation tool, apply here or write to us directly to get access to our private beta. We are very excited to hear your feedback.


Learn more about the project in our technical rundown of what we built and our website about the fee estimation engine.


More about Upvest: