AI Isn't Waiting for CLARITY. Neither Will Financial Institutions.

The CLARITY Act stalled in the U.S. Senate on Tuesday. Two days later, the SEC moved anyway. Yesterday, the SEC introduced an Innovation Exemption creating a temporary pathway for certain tokenized U.S. stocks to trade onchain.

I think that sequence may ultimately be more interesting than the vote itself.

Congress didn't provide comprehensive regulatory clarity. Technology didn't stop. Regulators didn't stop. Entrepreneurs certainly aren't going to stop.

So why would financial institutions?

I've spent much of this year writing about what happens when AI, programmable financial infrastructure and entrepreneurship collide. My argument has never really been about crypto. It's about how these forces change who can create financial products, how quickly new businesses can be built and eventually even who, or what, the customer is.

CLARITY failing doesn't change that thesis. If anything, the last few days made me more convinced of it.

CLARITY was attempting to establish a more durable federal market structure for digital assets. That didn't happen this week. Then the SEC demonstrated what happens next.

Its Innovation Exemption creates temporary, conditional relief for certain tokenized U.S. stocks. The SEC isn't pretending this is the final answer. Chairman Paul Atkins described it as a bridge toward more durable rulemaking.

That’s the entrepreneurial playbook: don’t wait for perfect certainty. Create enough room to experiment, invest enough to learn, and let real world evidence inform what comes next.

When I wrote CLARITY Matters, I argued that regulatory clarity changes what institutions are willing to invest in, build, partner on or acquire.

The other side of that argument is equally important: Waiting for certainty is itself a capital allocation decision.

Entrepreneurs don't eliminate uncertainty before deploying capital. They put enough capital behind an opportunity to learn something, preserve optionality and then decide whether the evidence justifies investing more.

Financial institutions need the ability to do the same.

In Infrastructure, Not Apps, I argued that the bigger opportunity around Ethereum isn't another generation of speculative applications. It's financial infrastructure.

Stablecoins are becoming payment rails. Assets are moving onchain. Money is becoming programmable.

Yesterday's SEC announcement makes that considerably more tangible. We are now talking about a regulatory pathway for tokenized versions of U.S. public stocks, carrying the same underlying shareholder rights, to trade using onchain infrastructure.

AI is accelerating everything at the same time. We're not even four years removed from the launch of ChatGPT, yet AI is already changing how software gets built, how knowledge work gets done and how companies operate.

In The Best AI Strategy Isn't About AI, I argued that corporations spend too much time asking how AI can improve their existing businesses and not enough time asking what completely new businesses become possible because AI exists.

That leaves financial institutions with two AI jobs.
1. AI transformation: make the institution you already have dramatically better.
2. AI growth innovation: create new business models, revenue streams and ventures that don't exist today.

Transform what you have. Build what's next. Most enterprise AI investment today is focused on the first. There is enormous value there. I just don't think it's enough.

Corporate venture building isn't new. Highline Beta began working with RBC on its venture building approach back in 2017. RBC subsequently built RBC Ventures and ultimately integrated its venture capabilities into RBCx.

What has changed is the stakes, speed and economics.

AI is making it dramatically cheaper and faster for small teams to build. So I've been thinking about a simple question: What if financial institutions treated even 5% of their innovation and transformation capital as actual venture capital?

Not another portfolio of pilots. Capital explicitly allocated to discovering, building and owning new businesses. Y Combinator invests $500,000 into every startup accepted into its program. Imagine that startup is an AI native fintech.

Those founders aren't being given $500,000 to produce a strategy deck or proof of concept.They're trying to build a company. Ship product. Get it into customers' hands. Learn quickly. Demonstrate traction. Create something worth more than the capital that went into it.

So what should a financial institution expect from $500,000 invested into an Agentic Finance venture? The expectations for what $500,000 can accomplish should be dramatically higher than they were five years ago.

A financial institution isn't a startup. Nor should it try to become one. Banks are optimized for scale, security, regulatory compliance, risk management and trust.

Put a five person AI native venture through the same procurement, technology, compensation and governance systems as the core bank and you'll eliminate most of its speed advantage.

The answer isn't making the bank behave like a startup. It's creating a different operating model for building ventures around the bank.

Give a small entrepreneurial team enough independence to move quickly while giving it access to the institution's customers, data, distribution, capital, regulatory knowledge and trust.

Those are advantages most fintech founders would kill for.

My Highline Beta co-founder Ben Yoskovitz has been exploring this from another angle in Did AI Kill the Lean Startup? and They Tried to Kill Product Managers. Now Everyone Needs to Be One. His argument is that AI is accelerating the Build-Measure- Learn cycle while making product judgment more important. When building gets cheaper, knowing what to build becomes more valuable.

That's exactly why the venture model matters.

Don't try to make the bank operate like a startup. Give the startup the unfair advantages of the bank without making it operate like one.

Agentic Finance is an obvious place to start. AI agents will need identity, permissions, accounts, payments and eventually access to capital. Ethereum and stablecoin infrastructure are emerging as likely rails for at least some of that activity.

Nobody knows exactly what that financial system will look like five years from now. That's the point. Entrepreneurs don't wait until the future is obvious before they start building it.

Regulatory uncertainty isn't a reason to wait. It's a reason to change how you build.

Build working products. Get them into customers' hands. Create ventures when the opportunity belongs outside the core. Kill the ones that don't work. Put more capital behind the ones that do.

CLARITY will evolve. Regulation will evolve. The technology certainly will. The financial institutions that succeed won't have predicted all of it correctly. They'll have built the capability to keep moving as the evidence changes.

Because the future of finance won't wait for permission to be built.

Highline Beta launched the Highline Beta Shop this past week, built for corporate builders, innovators and entrepreneurs who would rather build the thing than talk about building it.

If you're a CIO in BETA subscriber and want to pick something up, contact me and I'll send you a subscriber discount code.