Investing In: Upgrading the Internet
One of the areas the Fly VC team and I have been thinking about most actively is what is needed to upgrade the internet for an agentic future. Today, the internet is a global necessity that has transformed the world in countless ways, but it’s rooted in a human-first framing that may not reflect the future. It seems increasingly clear that the era of AI-powered agents is quickly approaching, and they could become the primary driver of economic activity—at least if the accelerationists have their way. In that light, are the systems we have today fit for the future we envision?
As I speak with founders on the front line, I’m increasingly of the view that the answer is no, and the opportunity to build the infrastructure that will define the intelligence economy is now. How this all will play out is very unclear, and I suspect the future will be messy, imperfect, and almost biological in its manifestation. I do, however, think there are some near-term challenges that, if not addressed, will stunt or even prevent the agent economy from flourishing.
Opportunities today
The clearest example, in my mind, is the reinvention of the financial rails required for agents to economically participate in our world today. The promise of agents and AI writ large is to free us from the drudgery of the day-to-day administrative work that takes up so much of our time. Menial tasks like purchasing groceries or buying plane tickets are a commonly cited promise of the future, and because these are automations of existing behaviors, it seems likely that our existing infrastructure can be bent to enable these new capabilities with little resistance. It’s no wonder that such tasks make up a good portion of the portfolio from industrialized company builders today. Interesting but ultimately uninspiring.
For example, spend accounts with rules-based purchasing are relatively commonplace within large companies that want to provision their employees with the goods they need while preventing wasteful expenditure. It seems relatively straightforward to extend these controls to personal agents or “digital employees.” Payman AI, for example, has built sandboxed financial accounts for AI agents that sit alongside corporate financial systems in a fully compliant way.
But as purchasing behavior matures beyond the mundane, new controls will be required to provide the assurance and stability that the market will demand. Custodians and escrow systems are the obvious examples, particularly those that have been natively built for the blockchain such as Palisade. Enormous security innovation has come about in order to ensure discipline and security over digital assets . Multi-party computation and mulit-sig approval technology will likely extend into agentic systems, ensuring that human managers still maintain control over significant funds. Though these systems are rigid and slow by design, this may prove fatal to an agentic future. At a minimum, our KYC/AML architectures would need to be completely revamped because they can no longer rely on the traditional idea of an identity. Actually intelligent agents open up a host of new security challenges, which our portfolio company Lakera is actively addressing.
Quick anecdotal side note. Most of my conversations with financial leaders tend to gravitate towards payment revocability rather than robust approval. For many, the immutability of blockchain solutions is a big bug, not a feature. This suggests that either more centralized solutions will take hold or stopgap products like agent-focused cyber insurance will be required.
Opportunities tomorrow
Where things get particularly interesting is when agents are allowed to execute in ways we never have. For example, agent-to-agent microtransactions—where average order size would plummet and total purchasing volume would explode. A financial Jevons paradox, if you will.
What new behaviors emerge when settlement time and cost approach zero? The combination of near-frictionless transactions and 24/7 autonomous execution would unlock entirely new markets that would be preposterous today. My personal white whale is autonomous vehicle traffic orchestration. Imagine a world where a Wayve-powered car actively participates in dynamic, real-time bidding of passing and yielding—a Disney Fast Pass for the road where time is literally money. That may sound like a capitalist dystopia to some, but to me and others, it’s the future.
The most confounding question is whether agents will cohabitate our internet or build their own entirely? The web is a human invention, built for our understanding of biologically based sentience. We’re a far cry from achieving eight billion internet users, but it’s very possible that we’ll soon(ish) see 80 billion agents. The internet, as it’s built today, will be crushed by the weight and volume of what’s coming. Does this mean we reinforce what we have or build an entirely separate agentnet? Is this agent-first web fully headless? Will we soon need human on-ramps?
I have no idea, but I’m looking to speak with those who do.