On October 1, the Ethereum Foundation, in collaboration with the Open Anonymity Project, introduced zkAPI, a system designed for making payments for AI models and other chargeable APIs without linking requests to the user's payment identity. This solution is already operational on the main network of the second-largest cryptocurrency by market capitalization.
Today the EF and OA announce zkAPI—a new means for private AI
— Ethereum Foundation (@ethereumfndn) October 1, 2026
Picture this: deposit ether into a vault, sign a zk proof, get a fixed amount of private inference. Advances like these help us move past middlemen's requirements that we link our data and identities.
Read more below https://t.co/LBy0Xz6cyi pic.twitter.com/7tOHRmXM4r
Users can make a one-time deposit of funds, such as ETH or USDC, into a smart contract vault. This deposit is then represented as a private note, which can be utilized through zero-knowledge proofs.
After validating the proof, the server issues a short-term API key with a spending limit. This key allows requests to be sent directly to the AI provider.
Once the key expires, the server deducts the actual cost of usage from the balance.
This setup separates the payment and content aspects. The zkAPI server only sees the payment's occurrence and amount, but not the specifics of the requests. Conversely, the AI provider can see the prompts but does not know the identity of the payer. The system identifies repeated use of the same balance through a special identifier called a nullifier.
The client is compatible with OpenAI and Ollama interfaces, enabling existing applications to connect without modifications. Developers have highlighted various scenarios for use, including blockchain RPC, image and video generation, VPNs, and payments between AI agents.
However, zkAPI does not guarantee complete anonymity. The provider can still see the content of requests and the IP address, and sessions can be linked based on time, personal information, or writing style.
To enhance protection at the network level, developers recommend using a VPN or Tor.
zkAPI is a practical implementation of the ZK API Usage Credits concept, introduced in February by Davide Krapis, head of the dAI direction at the Ethereum Foundation, and Ethereum co-founder Vitalik Buterin.
Authors Withdraw EIP-8363 from Hegota
Meanwhile, co-authors of EIP-8363 have decided to withdraw the proposal from consideration in Hegota. One of the co-authors, Jerome de Tychey, president of Ethereum France, explained that the update was an "inappropriate platform for decisions regarding emission policy."
📋Withdrawing EIP-8363 from consideration for Hegotá.
— Jerome de Tychey 🦇🔊 (@jdetychey) October 1, 2026
EIP-8363 rapidly became one of the most commented-on EIPs in the history of the Ethereum-Magicians forum, with 200+ comments in a few weeks. As we progressed through the Hegotà CFI (Consideration For Inclusion) process,… https://t.co/VjoEW11OYH pic.twitter.com/Lp7spyZk0p
EIP-8363, or Tapered Issuance Burn, proposed burning an increasing share of validator consensus rewards as staking grew. With 60.25 million staked ETH, which constitutes about half of the supply, 100% of rewards would be burned. The transition was suggested to span 18 months.
The authors believe that a large proportion of locked coins poses threats to the security, neutrality, and resilience of the network against takeover, as well as to ETH's role as money.
However, industry participants and protocol developers pointed out that reviewing a separate hard fork is not a suitable format for emission-related decisions. The co-authors agreed and moved the issue to a different process, with the Lido team offering assistance. EIP-8363 remains open and retains its draft status.
Among the key discussion topics, de Tychey mentioned security, industry impact, burning curve parameters, validator composition, and implications for solo stakers.
It is noteworthy that at the end of September, Buterin described Ethereum's evolution into a "cryptographic world computer" — a system that integrates blockchain, privacy, verification, and decentralized off-chain components.
Follow ForkLog on social media
Telegram (main channel) Facebook X