Matt Ober is a managing partner at Social Leverage. Matt was most recently the chief data scientist at Third Point, where he built the data analytics and technology platform used to enhance the firm’s investment capabilities in equity, structured credit, venture capital and cryptocurrency.

AI agents are beginning to do more than analyze financial information. They can now trade stocks, make purchases and monitor markets for opportunities on a user’s behalf.

Subscribe to the FR

That creates an opening for a new layer of fintech infrastructure. Prediction-market data, trade routing, risk analytics and compliance tools will all be needed. But fintech investors should not expect the companies building them to become overnight successes.

When agentic commerce meets agentic trading

Robinhood is going agentic, with two recent announcements showing how AI agents could change both trading and commerce.

Agentic trading is now available on Robinhood. Our portfolio company Alpaca.Markets has brought agentic trading to everyone and is now offering something similar to its clients.

Tell Robinhood to analyze your portfolio and diversify into more energy stocks and fewer tech stocks based on XYZ. Let the agents trade with their own pool of capital across equities and, in the future, crypto, options and prediction markets.

The other big release, which I haven’t seen anyone else do, is agentic commerce on Robinhood’s Gold credit card. Users can set up virtual cards, give an agent a set amount of money and let it agentically monitor for buying opportunities.

Sneakerheads, for example, could say, “Buy me the Air Max 1 Amsterdam shoe if a size 9 goes below $280 on StockX, GOAT or any other reseller site.” Or maybe you have heard your partner say, “I really want to buy this pair of jeans or shirt, but I am going to wait for it to go on sale.” Now your agent will handle that.

If you previously spent time hunting for deals, thinking of gifts you might buy at a lower price, hunting for airplane tickets or hotels if the price was right, or finding vacations within your budget, NOW your AI agent can do it. It will be interesting to watch how people deploy these capabilities.

Prediction markets need their own picks and shovels

Prediction markets are growing, but the tools and data supporting them still need to be built.

I was catching up on Robinhood’s prediction-market growth, its partnership with Kalshi and its own prediction-market infrastructure. I was reading an article and it made me realize just how big the opportunity could become.

What if prediction markets become bigger than options or futures markets? Not sports prediction markets, but markets covering financial, economic and other nonsports events.

When I think about the opportunities they are both on the retail and institutional sides. They range from trade-order routing to analytics, social copy trading, leverage, compliance and data.

A lot needs to be built. Many companies are starting to build, and some tools exist, but no winners have emerged yet because it’s too early. We likely will see more exchanges as well as more consolidation.

Where I think the biggest new companies could be

Normalized data infrastructure, or the “Bloomberg/FactSet for prediction markets”
Enterprise APIs for funds, AI models and corporations
Execution and routing technology that aggregates liquidity across venues
Risk and portfolio analytics that integrate prediction exposures with traditional assets, allowing for scenario analysis and backtesting
Hedging tools leveraging prediction markets
OTC market makers for prediction markets that lack liquidity
Vertical AI agents that monitor thousands of markets, explain probability changes and recommend trades or hedges
Compliance tools that monitor for insider trading
Retail apps that allow for agentic trading while surfacing underpriced markets and large volume swings

It’s early, but there is volume and interest. Good financial infrastructure takes years to build. The companies building it just need to stay alive long enough for the markets to really grow.

Fintech takes time

Building companies takes time. Fintech is a long road.

In a world of AI, there are both expectations and examples involving startups hitting $100 million in revenue in their first 12 months. Should this be the expectation, or is it rare? I am in the camp that it is not reality.

Yes, for some AI companies or unique businesses, it’s possible, but for very, very few. Some companies need to go for it, grow at all costs and be the category winner, or they won’t survive. But for most others, building a good, successful business takes time, pivots and survival.

That is particularly true in fintech. The infrastructure takes a long time to build. The regulations, licenses and laws tend to drag out timelines and are not easy to acquire or meet. The partnerships, especially with banks, are a long process. Building trust, like in many industries, is something that takes time, lots of it.

Enjoying? Sign up here

If I look at Social Leverage portfolio companies where we invest in the seed stage, Robinhood took nine years to go public, while eToro took more than 12 years. Alpaca.Markets is from our 2018-vintage Fund III, Street Context took 12 years to get acquired by BlueMatrix and SecFi is from our 2018 Fund III.

There are lots of other examples across our portfolio, but also across the fintech ecosystem.

Sometimes it’s about surviving, staying alive and building for the long haul. It’s hard building a startup; fintech is no different.