Weekly Project Updates: Ethereum Launches zkAPI, Base Rolls Out Cobalt Upgrade, Balancer and Blast to Shut Down, etc

1. Tether-backed Utexo Plans to Issue USDT on Bitcoin Network This Month link

Utexo, backed by Tether, plans to issue USDT on the Bitcoin network this month. Viktor Ihnatiuk, co-founder of Utexo, stated that the firm has obtained a commercial license to issue USDT on Bitcoin and intends to provide APIs, SDKs and cloud infrastructure for exchanges, wallets and payment service providers. Built on the RGB protocol and Bitcoin’s UTXO model, Utexo keeps most transaction data off-chain. It focuses on three types of use cases: private USDT transfers, direct swaps between native BTC and USDT, and native BTC-collateralized lending without wrapping BTC onto other blockchains.

USDT was first launched on Bitcoin via the Omni protocol in 2014; Ethereum and Tron later became its primary circulation networks. Utexo also plans to extend support to the Lightning Network after the Bitcoin-based USDT goes live.

2. Ethereum Foundation and Open Anonymity Launch zkAPI to Hide API Payment Identities via Zero-Knowledge Proofs link

The Ethereum Foundation and the Open Anonymity Project have launched zkAPI, which uses zero-knowledge proofs to decouple API usage from payment identity and is now live on the Ethereum mainnet. Users first deposit limits such as ETH and USDC into an on-chain Vault. Afterwards, zero-knowledge proofs are used to authorize API access for a single request or an entire session. Service providers can verify payment validity without learning the specific source of deposits or payer identity, and the payment layer cannot view request content. zkAPI supports generating short-lived, rate-limited API Keys for compatible interfaces including OpenAI and Ollama, with settlement based on actual usage. Funds are held in smart contract custody, and users can withdraw on-chain by exiting on their own. The solution is currently mainly for AI inference and can be extended to RPC, image generation, VPN and machine-to-machine services. It cannot hide identity-related risks stemming from IP addresses, request timestamps or prompt content itself.