Every cycle in Web3 produces a wave of applications that promise to bring the next hundred million users onchain. Most of them stall for the same reason: the rails underneath are not ready. Transactions are too slow or too expensive, assets are stuck on one chain, or the tooling makes simple things hard.

That pattern shapes how we invest. Before we ask whether an application can win, we ask whether the infrastructure it needs exists. Often the more durable opportunity sits one layer down.

The pattern we keep seeing

When a new category takes off, the teams that capture the most lasting value are rarely the first consumer products. They are the teams that made those products possible: the bridges, the compute layers, the payment rails and the developer tools that every application quietly depends on.

Applications compete for attention. Infrastructure competes for trust. Once developers build on a piece of infrastructure and it works, switching is costly, and that creates a kind of staying power applications rarely have.

What makes infrastructure investable

Not every protocol is a good infrastructure bet. We look for a few specific signals:

• A clear bottleneck. The team is solving a problem builders already complain about, not one they might have someday.
• Developer pull. Teams integrate because it saves them real time or cost, not because of incentives alone.
• A path to sustainable fees. Usage turns into revenue in a way that does not depend on constant subsidies.
• Composability. The product becomes more valuable as more teams build on top of it.

The best infrastructure is invisible to the end user and indispensable to the builder.

Where we are looking now

Our current focus includes interoperability between chains, confidential and verifiable compute, and payment rails that make stable value easy to move. Each of these removes friction that shows up in almost every product conversation we have with founders.

We are also paying close attention to the infrastructure AI agents will need to operate onchain: identity, payments between machines, and ways to prove that work was actually done.

What this means for founders

If you are building infrastructure, expect a longer road to obvious traction, and expect us to be comfortable with that. We care more about the quality of your early integrations than the size of your launch. If you are building applications, we still want to hear from you, and one of our first questions will be which rails you depend on and how confident you are in them.

At the earliest stage there is very little data. There may be no product, no users and no revenue. What there is, always, is a team. That is where almost all of our work begins, and it is the part of a pitch we spend the most time on.

Over time we have found that three qualities matter more than anything else.

1. Market insight

We look for founders who understand a specific problem better than almost anyone else. Often that insight comes from having lived it: running a trading desk, building on a chain that frustrated them, or working inside a community that lacked the right tools.

A broad vision is useful. A sharp, specific understanding of why the current solutions fail is far more valuable.

2. Shipping speed

Crypto markets move quickly, and the teams that win are usually the ones that learn fastest. We pay attention to how often a team ships, how quickly they respond to feedback, and how honestly they talk about what did not work.

We invest in the rate of learning, not the current state of the product.

3. Staying power

Every founder in this space will live through at least one difficult market. We want to back teams who plan for that from day one: careful with runway, clear about their priorities, and building something they would still want to build if prices went sideways for two years.

What we do not over-weight

• A polished deck. Clarity matters, design does not.
• Vanity metrics. Incentivised users and follower counts tell us little on their own.
• Early token price. We care about the product and the community, not short-term market moves.

How to reach us

If this sounds like your team, we would like to hear from you. Send a short note on what you are building, why now, and why you are the right people to build it. A clear page of writing is often more useful to us than thirty slides.

AI agents and crypto have become one of the most discussed combinations in technology. Some of that attention is deserved. Some of it is a familiar pattern: a real idea wrapped in a lot of noise. Our job is to tell the two apart.

Why agents and crypto fit

An autonomous agent that needs to pay for data, compute or services runs into a wall in traditional finance. It cannot open a bank account or pass a credit check. Onchain, it can hold value, sign transactions and interact with open protocols without asking permission.

Crypto also offers something agents badly need: verifiability. When an agent claims it did a piece of work, onchain systems can make that claim checkable.

Where the value is real

• Execution. Agents that automate complex onchain tasks, such as managing positions or routing liquidity, and save users real time.
• Verifiable compute. Infrastructure that proves which model ran, on what input, and that the output was not tampered with.
• Machine-to-machine payments. Rails that let agents pay each other for services in small amounts, reliably and cheaply.

An agent that cannot hold value or prove its work is just a chatbot with a wallet address.

Where the story runs ahead

Many projects today are thin wrappers: an existing model, a social account and a token. There is nothing wrong with experiments, but a token alone does not create an economy. We look for usage that would still exist if the token disappeared.

What we look for

We are most interested in teams building the plumbing that serious agent products will need, and in applications where an agent does something a human genuinely cannot do as well or as cheaply. Strong technical founders, honest benchmarks and real users matter far more here than a compelling narrative.

Most coverage of crypto focuses on prices, launches and speculation. Meanwhile, one of its most useful applications has been growing steadily with far less noise: moving money.

The quiet part

Businesses paying suppliers across borders, teams paying contributors in many countries, and people sending money home are all discovering the same thing. Stable value on open rails can settle faster and more cheaply than many traditional routes, and it works at any hour.

These users rarely care about the technology. They care that the payment arrived, that it cost less, and that they could track it.

Why now

Several things have changed at once. Stablecoin liquidity is deeper. Wallets are simpler and safer. On and off ramps are easier to access in more markets, and rules in several regions are becoming clearer for the businesses that want to use these rails.

The best payments products do not feel like crypto at all.

Streaming and programmable payments

Beyond simple transfers, crypto makes new payment shapes possible. Salaries can stream by the second instead of arriving monthly. Payments can release automatically when conditions are met. Revenue can split between many parties the moment it arrives.

What we look for

• A focused use case. One corridor, one customer type, done extremely well.
• Real volume from real users. Repeat usage matters more than headline totals.
• Compliance built in. Teams that treat regulation as part of the product, not an afterthought.
• Great experience. Payments that are simple enough for someone who has never used crypto.

Payments may never be the loudest category in Web3. We think it may be one of the most important.