Bryce Ferguson, CEO and Co-Founder of Turnkey – Interview Series

Bryce Ferguson, CEO and Co-Founder of Turnkey, is a product-focused operator with deep experience in crypto infrastructure and fintech, having played a key role at Coinbase where he helped scale Coinbase Custody into the world’s largest crypto custodian and launched staking and governance products, before leading crypto expansion at Trade Republic and earlier driving product strategy at Stitch Fix and consulting at Bain & Company; he also served as a scout for Sequoia Capital, reflecting a strong connection to early-stage innovation as he now builds Turnkey.

Turnkey is a crypto infrastructure company focused on secure, programmable wallet and key management systems, offering developers APIs to create embedded wallets, automate transactions, and manage private keys without exposing sensitive data, all within hardware-isolated environments designed for verifiable security and control; built by a team with deep experience in institutional crypto custody, the platform enables companies to generate wallets at scale, sign transactions programmatically, and integrate crypto functionality directly into applications without relying on traditional seed phrases, while maintaining auditability and granular policy controls.

You have had a front-row seat to the evolution of institutional crypto infrastructure, from helping grow Coinbase Custody into the largest crypto custodian in the world to leading the launch of crypto trading at Trade Republic. What lessons from those experiences ultimately led you to found Turnkey, and what problem in crypto infrastructure did you believe still needed to be solved?

Working at Coinbase Custody was an amazing experience. When I joined, it was just 180 people, and felt like the center of gravity for the crypto industry. I got to learn a ton about how crypto was evolving and all of the foundational infrastructure being put in place.

I worked with Jack Kearney (now my co-founder) on building Coinbase Custody. He was the first engineer and I was the first product manager. We essentially built the first institutional crypto custodian from the ground up, scaling it from 0 to $100 billion in assets. Three years in, Coinbase Custody was the dominant player. I was hungry for another 0 to 1 experience, and got introduced to the founders of Trade Republic.

At Coinbase, I was building infrastructure. At Trade Republic, I was using a lot of these infrastructure providers, and got to see what the rough edges were, what was working, and what wasn’t. Jack also moved on from Coinbase and became CTO of Polychain. Similar to me, he was using all of these crypto custody providers and banging his head against the wall. In 2022, we reconnected and started thinking about how we should pull infrastructure out of custody and go as low level as possible, with really flexible tools around how you can create and manage wallets as part of whatever application you’re building.

You hear that now, and folks think “of course, API-powered wallets seem obvious.” But in 2022, you couldn’t do that. All developers had access to were these clunky, point-and-click web interfaces that people were dealing with that they’d try to layer products on top of.

Turnkey focuses on programmable wallet infrastructure and private key management for developers. Why has key custody remained one of the most persistent bottlenecks in crypto innovation?

Every transaction in crypto starts and ends with a private key. Creating, storing, managing, securing, and using private keys are extremely hard problems to solve.

Traditional key management products struggle to keep up with new chains, ecosystems, and use cases because they have overfit their offering to a single narrow use case. As a result, builders are hobbled by their custodians or key management providers. Although these may be simple abstractions for developers, it isn’t possible to keep up with the explosion of ecosystems and use cases without crumbling under technical debt and maintenance burden.

Many teams are now experimenting with AI agents that can execute onchain transactions automatically. From your perspective, what are the biggest security challenges when machines, rather than humans, control wallets and execute trades?

When wallets act on behalf of humans, security is about access control, protecting keys and authenticating users. But when wallets begin acting on behalf of autonomous agents, the core security question changes. It’s no longer just about who has access, but what decisions are being made and why.

AI agents introduce a faster, more unpredictable threat model. Their decisions are shaped not by passwords or signatures, but by data, some of it unverified, adversarial, or incomplete. They can be manipulated subtly, not by breaking in, but by shaping the inputs they rely on. A malicious price feed, a corrupted prompt, a forged onchain signal all become vectors of influence.

This creates new design tensions. How do you build a wallet that moves fast enough for agents to operate autonomously, but cautious enough to resist invisible manipulation? How do you let agents execute without friction, while still surfacing enough context for meaningful oversight?

The role of wallet security in this new paradigm isn’t to eliminate risk entirely, but to make risk observable, bounded, and recoverable.

Turnkey often refers to “agentic wallets.” How do these wallets enforce governance rules, transaction limits, or authentication policies while still allowing autonomous agents to operate at machine speed?

There’s a lot to unpack here. Let’s start with the three critical security properties for Turnkey agentic wallet infrastructure:

1. Separation of control: The end user or business owns the wallet and sets the rules. The agent operates within those rules.

2. Zero key exposure: The agent never touches private key material. It receives the ability to sign transactions, not access keys.

3. Cryptographic enforcement: Policies are evaluated in the secure enclave, which means there’s no way to bypass them from application code.

With this security architecture in place, Turnkey’s policy engine enables programmable controls over signing behavior. Organizations can enforce transaction constraints, approval flows, and scoped permissions for agents, preventing unauthorized interactions even in complex multi-agent environments.

Turnkey enables autonomous agents to operate at speed within the policies defined by the end user or business ‘owner’. This includes sub-100ms signing latency, which is up to 100x faster than MPC-based alternatives, enabling agents to react to market movements and onchain events in real time.

We are beginning to see projects where AI agents execute trades directly from natural language prompts. Do you believe this model will become mainstream in crypto trading, or are there fundamental risks that still need to be solved first?

Yes, I think this will become more mainstream. One of our customers, Spectral Labs, builds tools for creating and deploying autonomous AI agents that operate fully onchain. Its platform turns natural language into executable blockchain actions, making automation and DeFi access simple and decentralized. Having worked with them closely, I believe we’ll continue to see this model adopted more widely in crypto trading.

Infrastructure often shapes entire ecosystems. How do you think agent-driven finance will change the architecture of exchanges, wallets, and on-chain protocols over the next five years?

With the rise of agentic finance, wallet infrastructure providers must ensure their policy engines and security controls give agents robust programmable spending policies, transaction limits, counterparty restrictions, approval flows, and real-time revocation mechanisms. Most importantly, all of this must be enforced cryptographically and executed with low latency.

Turnkey provides APIs that allow developers to create wallets, sign transactions, and automate onchain actions programmatically. What kinds of new applications or products does this infrastructure enable that were not previously feasible?

A common piece of feedback we get from Turnkey clients is “I couldn’t build this anywhere else.” One example of how our low-level infra powers novel use cases is the work we’re doing with World, co-founded by Sam Altman and Alex Blania.

World App is using Turnkey’s infrastructure for a programmable recovery system. World App encrypts each user’s recovery bundle locally on-device. Turnkey facilitates access by managing the bundle’s encryption key within secure enclaves, where the key is protected and can only be used in response to user-authenticated actions. The result is a flexible, secure recovery flow that keeps users in control and provides a better tool for account recovery. We’re now starting to see this use case across other businesses and industries as well.

Security in crypto has traditionally centered around hardware security modules and cold storage. With AI agents executing transactions continuously, do we need a new security model for automated financial systems?

Here’s how I think of it: just like employees need scoped credit cards to transact, agents need scoped wallets to transact. This means that wallet infrastructure serves as the control plane for agent transactions, which requires a security model that’s made up of a few key areas:

1. Verifiable security: Today, a lot of crypto infrastructure is still a black box. Many developers still rely on trust that wallet transactions are being executed as intended. Verifiable systems remove that assumption by offering evidence, allowing anyone to confirm that the code running is approved and untampered. Having wallets that are verifiably secure is particularly important when agents are performing tens of thousands of transactions daily. Humans need to ensure that the transactions are being executed as they intended.

2. Audit trails: On a related note, every transaction must produce cryptographic evidence that is secure and tamper resistant, which can be used for compliance, debugging, or incident response.

3. Transaction policies: Agentic wallets should have programmable transaction policies, designed with the goal of ensuring that the agent is only allowed to do the minimum set of actions for its required task. For instance, if your agent sweeps funds to a treasury, it doesn’t need permission to call arbitrary contracts.

4. Human-in-the-loop approvals: For sensitive operations, both the agent and a human should be required to approve a transaction before it executes. This provides defense-in-depth even if the agent behaves unexpectedly. Humans should also always have the ability to revoke permissions in real time.

Looking ahead, what does the long-term vision for agentic finance look like? Could we eventually see autonomous trading systems, DAO-managed agents, or even AI-native financial institutions operating entirely on-chain?

We’re already starting to see the initial stages of autonomous trading systems and AI-native financial institutions. In the long term, I see a world where every day, billions of autonomous agents are safely engaging in economic activity, following programmable policies while transacting with other agents at extraordinary speeds. Agents will operate with scoped wallets, enforced by policy, forming the foundation of a financial system that is autonomous, but still controlled, auditable, and secure by design.

Thank you for the great interview, readers who wish to learn more should visit Turnkey.