Jon Ma, CEO of Artemis, on building 24/7 AI agents for trading & investing

Over the last year, we’ve interviewed founders & CEOs about private company secondaries (Ben Haber, Noel Moldvai), tokenized equities (Han Qin, Xavier Ekkel), prediction markets (Kurush Dubash), and crypto exchanges expanding into multi-asset financial platforms (Arjun Sethi).

To understand what happens to investing as all of those assets increasingly sit in the same portfolio, we reached out to Jon Ma, co-founder & CEO of Artemis, which started in 2022 as a crypto fundamentals & data platform and is now building AI investing agents across public companies, private companies, and tokens.

Key points from our conversation via Sacra AI:

For more, check out this other research from our platform:

Artemis started in 2022 as a crypto data and fundamentals product for investors. What did you believe crypto investors were missing that products like Dune, Messari, Token Terminal, and others weren’t providing?

We met around 2020, when Sacra and Public Comps were getting started. At the time, making private and public company financials comparable was still an unsolved problem.

In 2022, I was at Whale Rock, and we were trying to look at tokens as a new asset class. It was unclear how to evaluate Avalanche, Solana, Ethereum, or Bitcoin. There was no revenue, no Rule of 40, no financials that made these assets comparable. It was irritating: how do I decide which tokens to invest in? People were pitching me all sorts of NFTs and tokens they wanted me to buy, but I wanted to have an independent view. There was no Sacra or Public Comps for crypto.

Dune wasn't really solving that. Dune allowed you to create a bunch of charts. Messari had a lot of good research—we've hired a bunch of wonderful people from Messari at Artemis—but they weren't as good at the data side. DeFi Pulse and DeFiLlama had a lot of metrics specific to DeFi.

I'm a normie software guy who loves Sacra and Public Comps. I wanted revenue and revenue multiples. I couldn't get that for crypto, so I wanted to build it to inject rationality into the space.

The other problem was that things weren't comparable. You had Ethereum, Polygon, Avalanche, Terra Luna, and people saying, "This thing is going to the moon." But why? How many users are using it? How many developers are using it? Even those simple questions couldn't be answered by the early tools circa 2021. I wanted to make the space comparable in a way that made sense to me. It was essentially Public Comps for crypto.

How much of the challenge was creating new kinds of data versus normalizing existing onchain metrics into comparable measures?

With public or private software companies, ARR is standardized: MRR times 12, and you can pull that from Stripe. The interesting thing about blockchains is that because they're all architected differently, finding an apples-to-apples definition of something like a user is very difficult.

For example, on Solana, a program or application can pay the fee on behalf of the user. So who is the user? Is it the program? Is it the application? Or is it the person getting free gas to interact with the application? Ethereum and EVM chains have a different architecture.

A lot of our problem was figuring out how to create the equivalent of a GAAP standard for crypto. How do you normalize some measure of demand or user activity across these different blockchains? A lot of it was normalization and using our best judgment.

What is Artemis today? Is it primarily a financial data company, an AI investment analyst, a research platform, or the beginnings of a new investing network?

In 2022, we were squarely building Capital IQ or Bloomberg for crypto—an Excel plugin and all the tools that someone coming from traditional finance would want to evaluate tokens.

In 2026, one of the unfortunate realizations is that there are only a handful of tokens that really make sense to buy and hold long term. We've pivoted from focusing on a crypto data platform to building investing agents for every investor, whether that's a retail investor, an independent investor, a hedge fund analyst, or an institutional analyst.

I want to build the investing agent of my dreams: something that can pull data on public companies, private companies, and tokens in one place so I can build dashboards and charts and develop conviction in an investment thesis. We've pivoted from crypto data toward a thesis-driven investment research platform.

You’ve said many of the liquid token funds Artemis originally served increasingly invest in Coinbase, Robinhood, fintech, AI stocks, and other public equities alongside tokens. Did expanding beyond crypto come primarily from customer pull, or from a broader conclusion that “crypto investing” is disappearing as a distinct category?

A lot of the larger funds that were staple crypto token investors—the Multicoins and Paradigms of the world—have shifted toward investing in equities like Robinhood, or private companies like Securitize.

What I've observed is that people are investing in tokens, equities, private companies, AI, robotics, and other frontier technologies. If you're starting a long-only crypto token fund in 2026, it's probably not going to be very easy to raise from LPs. I personally wouldn't recommend it.

Everyone invests in everything now. If you're a leading hedge fund, you can invest in prediction markets, tokens, private companies, and public companies. Investing only in tokens doesn't really work as a strategy anymore.

What lessons do you take from Messari about what worked and what didn’t in building a standalone crypto research and data company?

It's still possible to run data businesses in crypto. Companies like Dune are necessary. They can serve hedge funds, institutional funds, and governments that need to track stablecoin activity, understand tokenization, or see how much volume people are trading on Hyperliquid. Those are great businesses, and they'll continue to exist.

What is gone is the dream that everyone in the world is going to be a token investor and that you're going to have a huge swath of long-only, crypto-only token investment funds. That dream didn't play out. Crypto is now a sub-asset class for a lot of asset managers and hedge funds.

Prediction markets are interesting both as a new asset class and as a new dataset. When does an investor use a Polymarket or Kalshi probability as an input into another investment decision, versus actually expressing the view by trading the event contract itself?

If you and I started a hedge fund today, we'd probably want an analyst focused on prediction markets. If there's a biotech company we're looking at, what are the odds that the FDA is going to approve its drug? There's probably a robust prediction market that gives you a useful signal, and that could affect whether you buy the stock.

Prediction markets are also interesting because they let you invest directly in very tight theses. Hedge funds are beginning to use these signals to hedge bets around earnings. You can take a very directional bet on whether Tesla will beat expectations for vehicles shipped. Maybe you have a strong view on that and can make money on the prediction market, whereas even if Tesla beats consensus significantly, the stock itself might not go up.

That's interesting because you can invest directly in a specific thesis about a company metric. The limitation today is liquidity. These markets are still quite small.

What happens when investment strategies can be composed across prediction markets, equities, crypto, perpetuals, and eventually tokenized real-world assets?

One of our data scientists loves investing in tokens, private companies, and public companies. He's using Sacra to underwrite assets on Augment Markets. He's using Artemis onchain data to underwrite the small number of tokens he finds compelling. He's also using the public filings, financials, and consensus estimates that we're aggregating at Artemis to make decisions about which token, private company, or public company to invest in.

That's what is interesting in 2026. Our thesis at Artemis is that everyone is becoming a managing director and portfolio manager. You can invest in prediction markets, private companies, public companies, and tokens.

The question becomes: how do you filter through the noise, build conviction in one of these assets, and double down? It's really hard to invest in 2026. There are so many good companies and assets to invest in.

Ideas are plentiful. Conviction is not. You can read a post on X or Substack, read research from Sacra, or read a sell-side research note, but how do you actually develop conviction about what to invest in?

Take fintech and payments. You could invest in Stripe in the secondary market, Adyen in the public market, Ramp, crypto card startups like Rain or Kast, or a token like Ether.fi. What is your bet? Should you just invest in Stripe because maybe Stripe wins agentic payments and commerce and some of these smaller startups get subsumed?

You need the data and the tools in one place to underwrite those choices.

What new infrastructure is required for that world? Is the hard problem finding the signal, understanding the correlations, sizing the positions, routing execution across venues, managing collateral, or monitoring the combined risk after the trades are on?

I think there will be sets of agents and interfaces on top of brokerages that allow a person to express a worldview and a thesis, use AI to build models, set a price target, and then execute the investment.

Once the investment thesis plays out, the agent can automatically trade or sell on your behalf, connected to a Robinhood MCP, Coinbase MCP, Schwab, or another brokerage. You can focus on the thesis and not have to worry about the trade execution. I think that execution should ultimately be automated with MCPs and agents once the thesis has been expressed.

A large portion of prediction-market volume today is still sports and short-duration events. Do you think that is simply the distribution wedge that bootstraps liquidity, or could sports remain the economic center of the category while financial and informational markets are more intellectually interesting but smaller?

I'm probably in the camp that financial prediction markets aren't going to be that exciting for the average investor. They'd probably prefer to bet on the World Cup or the NBA Finals than whether Tesla is going to beat vehicle expectations by 10%.

I'm more interested in seeing how much hedge funds adopt those financial markets—how the Citadels of the world start using them on Kalshi and Polymarket. That could create real volume and become interesting, but I don't think it's there quite yet.

Robinhood, Coinbase, and Kraken started from very different places but are increasingly converging across stocks, crypto, derivatives, prediction markets, tokenized assets, and other financial products. How do you think about their different strategies today?

Brokerages have very little differentiation in 2026. Schwab, Interactive Brokers, Fidelity, Robinhood, Coinbase, Kraken, Gemini—they're increasingly offering crypto, stocks, options, derivatives, and other products.

The unsolved question is: what should I actually buy? What should I invest in?

My view is that these brokerages will eventually be forced to choose who they serve. Robinhood clearly serves traders. That's why it has a social feed showing the trades people have made. It could acquire something like TradingView or another product that's very trading-oriented.

Schwab, Interactive Brokers, and Fidelity serve a different customer. Those customers probably aren't buying some random token on Robinhood. They're buying and holding Tesla or SpaceX and reading good research that gives them—or their agents—conviction to buy and hold something.

I think platforms will bifurcate into trading platforms and long-term investment platforms, and their M&A strategies and the products they package on top will become very different.

Robinhood recently opened trading to third-party AI agents through MCP. If the investor increasingly delegates research and decision-making to an agent, does the broker become more like a commodity execution and custody layer, or do liquidity, account relationships, regulatory infrastructure, and embedded financial products make the broker even more powerful?

Robinhood Social has been interesting because Robinhood lets users share their trades. But I think the bigger opportunity is for agents and interfaces to sit on top of these brokerages.

A person should be able to express a worldview and thesis, use AI to build models and set a price target, and then have the system execute the investment. Once the thesis plays out, it can automatically trade or sell on your behalf.

The broker becomes the execution layer underneath that. You shouldn't have to worry about the execution if your thesis can be reflected through an agent.

At the same time, I think brokerages will differentiate by the kind of investor they serve and what they add on top. Trading platforms and long-term investment platforms will package very different products around that execution layer.

How much value do you think AI investing agents create for long-term investors versus active traders?

I struggle with Artemis building investing agents for traders because I'm not a trader. I'm a terrible trader.

Fundamental investing—buying and holding and looking at fundamentals, including a lot of what Sacra and Artemis provide—is where we want to build. I also think it's a bigger market.

There are a lot more millennials, boomers, and older investors who don't want to trade. When I talk to people over 30, they'll often say, "I don't trade, I invest." The market for people who want good research, insights, and theses to help them buy and hold is really big, and I think it's only going to grow.

You and I are effectively portfolio managers with teams of analysts pitching us ideas about what to do and what to buy and sell. We ultimately decide: that makes sense, that doesn't make sense, or, "All right, agent, once this threshold is hit, make the investment on my behalf."

Composer took a different approach: use AI to turn an investment idea into transparent rules, backtest it, and then automate the strategy. Robinhood is opening the door to agents with much more discretion. Which model do you think wins—AI helping humans execute on deterministic strategies, or agents continuously deciding what to buy and sell?

The idea that an agent will just go out there and make you money is very seductive. You see that a lot on X. But at some point, it's going to become fairly crowded, and the Jane Streets and Citadels of the world will figure out how to build really good trading agents that capture a lot of that alpha.

I think the space will shift toward having a worldview and conviction. Maybe you believe biotech is going to change significantly. Maybe you think AGI is going to happen and therefore you should invest in a certain set of assets.

I'm more in the camp that people will open the Artemis terminal one day and decide, "I believe in this idea. I believe in this idea. I don't believe in this idea. Go execute on my behalf."

The individual acts like a portfolio manager. That's different from plugging an agent into Robinhood, closing your eyes, coming back later, and finding that it made a 30% IRR for you. I can see that happening, but I don't think it's the end-all, be-all. I think that kind of automated trading alpha will be fairly commoditized.

Artemis has talked about building an investing network where retail investors, institutions, and even public and private companies can publish and pitch investment theses. What is the fundamental object in that network: a post, a thesis, a financial model, a portfolio, a track record, or something else?

Investing agents aren't a new or novel idea, and I think that market will be fairly competitive.

Our thesis is to develop a network on top where intelligent people can share investment theses, research, models, and buy-side-quality research, and aggregate all of that in one place. It's somewhat like Seeking Alpha, which I really like, but with the supply side perhaps a tick higher in quality—someone from a bank, a hedge fund, or an independent analyst.

That's what we're hoping to build: an investment thesis platform on top of everything else we're building.

Robinhood Social is interesting because trades and performance can be verified rather than simply claimed. Is verified track record the missing primitive that differentiates a real investing network from finance Twitter, Reddit, or traditional social investing products?

We've seen versions of social investing before. Commonstock was cool. It was basically a Facebook feed for investing, it really blew up during the pandemic, and Yahoo Finance acquired it.

I've used Robinhood Social, and I don't think what they've built as it exists today really works. If I see some random person buy $300,000 of Robinhood, that doesn't give me conviction.

There has to be some new primitive on top—maybe next-generation research or next-generation thesis-building—that actually helps people develop conviction.

I've heard anecdotally from someone who decided to buy secondaries on Augment Markets after reading a Sacra research piece before buying a big slug of a stock or a pre-IPO company. More and more, I think independent research is what will drive investing activity on these platforms.

It's not enough to say, "Nancy Pelosi bought this stock," or, "Tony Hawk bought Snowflake." No one cares what Tony Hawk is investing in. Sorry, Tony Hawk. Great guy, but he's not an investor.

How do you build a network around investment insight when the most valuable information often becomes less valuable as it spreads?

There's been a shift where a lot of hedge funds are becoming much more public. They're not necessarily pumping their bags, but they're advertising what they've invested in because people increasingly realize that retail investors matter.

Everyone is talking about their stocks. The All-In guys do it. Brad Gerstner has his own podcast.

Once you buy the stock, you have an incentive to share your investment thesis because you want other people to buy it. There's an education, sales, and marketing component to investing that matters more in 2026, which is kind of hilarious when you think about it.

Does AI ultimately make the investing market more efficient because everyone gets institutional-quality analytical capability, or less efficient because millions of agents can pursue increasingly specialized theses across a much larger universe of tradeable claims and assets?

I think we'll see a lot of solo GPs and hedge funds all over the world. There are going to be smart people in India, China, Africa, Canada, and Europe who have really good discernment and are very talented at using AI to make money.

The mission we have at Artemis is to make investing more meritocratic. You shouldn't have to be some Stanford Business School guy who went to Goldman Sachs to become a hedge fund manager. It could be some kid in Rwanda who loves investing, has a laptop, has really good worldviews and theses, makes a ton of money, and ends up on a leaderboard with returns that are competitive with Citadel. Then dollars and LPs can flow toward that person.

I think the world gets more meritocratic. Individuals around the world will be able to make much more money.

That will make things more challenging for the hedge fund industry. What separates a long-only fund from a retail investor at that point? Maybe inside information or proprietary data. The line between a very big hedge fund and an individual investor is going to become very thin, and I think that's very exciting for democratizing wealth for this generation.

If everything goes right for Artemis over the next five years, what does it become?

I'd love for us to build a network where retail investors can come to Artemis and sign up for an investing agent that monitors their portfolio 24/7 and sends them buy and sell recommendations.

Retail investors could also invest in funds on the Artemis platform using Hyperliquid vaults or some other new primitive. Then you'd have creators—people who love investing—publishing theses on our platform and raising capital from retail investors.

Anthony and I call it the "investing mecca": somewhere retail investors, hedge funds, and everyone else come together and pitch each other investments.

I want Artemis to be the investment platform that undergirds all of that.

This transcript is for information purposes only and does not constitute advice of any type or trade recommendation and should not form the basis of any investment decision. Sacra accepts no liability for the transcript or for any errors, omissions or inaccuracies in respect of it. The views of the experts expressed in the transcript are those of the experts and they are not endorsed by, nor do they represent the opinion of Sacra. Sacra reserves all copyright, intellectual property rights in the transcript. Any modification, copying, displaying, distributing, transmitting, publishing, licensing, creating derivative works from, or selling any transcript is strictly prohibited.