In a recent X Spaces discussion, leaders from Fetch.ai and Secret Network shared insights into their collaboration to create verifiable, private AI agents. The speakers included Fetch.ai founder and CEO Humayun Sheikh, Developer Relations lead Nikolay Dimitrov, Secret Network Foundation Executive Director Lisa Loud, and Secret Labs CEO Alex Zaidelson.
Their shared goal is to bring integrity and privacy to autonomous AI systems, especially as use cases begin to include financial services and healthcare.
“Privacy plays a very important role,” said Sheikh. “Especially when there’s banking or medical data involved.”
Agent-first vision
Fetch.ai focuses on building agentic systems, where autonomous agents carry out tasks for users across multiple blockchains. According to Dimitrov, their lightweight framework is accessible and already supports millions of deployed agents.
“The agent-first approach makes sense,” said Loud. “We’re not measuring development in months anymore. We’re seeing meaningful outcomes in days.”
Secret Network complements this with infrastructure for confidentiality and verifiability. Their technology includes confidential virtual machines that can produce attestation proofs, showing that agents are running exactly the code they’re supposed to.
“Even changing the version number in the Docker file alters the attestation result,” said Zaidelson. “That’s how precise this is.”
Early use cases
A shared project currently underway uses a breast cancer detection model secured in a confidential VM. Medical professionals can trust that the results are accurate and untampered, with all data processed privately.
“It’s a serious domain,” said Zaidelson. “We’re building tools to support high-stakes decisions.”
Beyond healthcare, Dimitrov pointed to financial and industrial applications. Fetch.ai agents already have wallet addresses and can execute transactions. With the addition of secure infrastructure, these agents could serve enterprise users seeking both automation and control.
On trust and data
The speakers repeatedly returned to the importance of data privacy. Loud highlighted the risks of model leakage through prompt engineering and emphasized the need for opt-in confidentiality.
“You can extract personal information from models using clever prompts,” she said. “This is not theoretical. It’s happening.”
Dimitrov noted that many users assume AI models are private by default, when in fact their data may be stored and used indefinitely. He argued that cryptographic attestation creates a more trustworthy foundation for deploying AI in sensitive areas.
Cosmos collaboration
The integration between Fetch.ai and Secret Network took just over a month. Both teams cited the Cosmos ecosystem as an enabler of fast, modular development.
“Cosmos is for tinkerers,” said Zaidelson. “You can take components and assemble something new without starting from scratch.”
Looking ahead
The teams plan to introduce templates for developers to build verifiable agents and deploy them in secure environments. As Dimitrov put it, “We may see a Cambrian explosion of utility, informed by real demand rather than guesswork.”
For now, the focus is on combining speed of development with cryptographic assurance. The partnership is early but already demonstrating practical applications.
“We’re not just watching this space,” said Zaidelson. “We’re part of shaping where it goes.”
Watch the Spaces here (on video)
