Here is the entire product in two lines of logic. If the market prices a company's earnings miss above 60%, the agent buys a put. If an FDA approval crosses 75%, it fires a $5,000 buy order. That is the pitch Public took to market on September 24, 2026, dressed in the phrase "the world's first agentic brokerage." No new cryptography. No new settlement layer. No token. Just a threshold and a trigger, sitting between a prediction market's implied probability and a retail investor's order book.
I have spent twenty-nine years watching this industry repackage old machinery as revelation. I audited fifty-odd ERC-20 whitepapers in 2017, and the pattern never died. The genuinely novel thing here is not the execution layer — it is the input. Public is not feeding its agent price data or earnings prints. It is feeding it Kalshi's event-contract probabilities, the crowd's real-money guess about what happens next. That is a genuinely different signal source. Whether it is a good signal source is a question nobody has answered.
Why This Lands Now
Kalshi is not a startup experiment. It holds a CFTC Designated Contract Market designation, meaning it is a federally regulated derivatives exchange, and it cleared roughly $39.7 billion in annual volume. It runs a crypto category, it has been used to price Bitcoin direction, and its rails brush against Circle and Binance on one side and Visa, SoFi, and Mastercard on the other. If you squint, the whole thing looks crypto-adjacent. If you look straight, it is fintech infrastructure with a prediction-market engine bolted to the front.
Public, meanwhile, is a licensed US brokerage. Its co-CEO and co-founder Leif Abraham put his name on the launch, which tells you this is a strategic bet, not a side experiment. The integration lets users trade Kalshi event contracts directly inside a Public account and, more interestingly, lets an AI agent read those probabilities as trading signals.
The timing is not accidental. Robinhood has been shipping Agentic Accounts that connect to third-party agents. eToro is pushing Agent Portfolios. And TrueNorth — founded by former Public executives — made the same "world's first" claim. When three or four firms race to say the same sentence, the sentence has stopped meaning anything. From ICO hype to on-chain truth, the industry's oldest habit is to inflate a feature into a category.
The Engineering Is Trivial. The Signal Is the Bet.
Let me be precise about what was actually built, because the marketing blurs it. This is a threshold-trigger conditional order engine. The execution side — conditional orders, automated puts, robo-advisory flows — has existed in traditional brokerage for years. Wiring an API that watches a number and fires an order is not hard engineering. Any competent backend team ships that in a quarter.
The difficulty lives entirely in the signal layer, and the signal layer is where the product is naked.
Here is the trap I keep circling back to. A prediction market's "60% probability" is not a probability in the statistical sense. It is a price. It reflects the marginal trade, not the aggregate belief. If a contract is thin — and many event contracts are thin — then a 60% print can be manufactured by a few thousand dollars of one directional bettor. You are not reading the crowd's wisdom. You are reading the crowd's last order, which might be one person with a strong opinion and a small position.
I have seen this movie. In 2020, during DeFi Summer, I watched liquidity migrate toward whichever pool paid the highest emission, regardless of whether the underlying asset had any depth. The number on the screen looked like a market. It was a mirage with a fee attached. Prediction-market probabilities have the same failure mode: a clean-looking percentage that hides a dirty, shallow book.
Public has not disclosed latency, slippage, or fill rates for the agent layer. That silence matters more than any feature list. When a brokerage routes an automated order off an external probability feed, two clocks have to agree. The prediction market may be pricing an event around the clock, while the equity or options market is closed. Your agent sees 74% at 4 a.m. Rome time and fires at the open, into a gap, into a completely repriced world. Timing mismatch is not a bug you patch later. It is baked into the architecture of cross-venue automation.
The author framing around this launch posed the honest question — whether these signals are truly predictive or merely well-packaged noise — and left it open. I will go further. Until somebody publishes win rates and alpha over a meaningful sample, the product is a more efficient trade-frequency generator. Brokers earn commission on activity. Every trigger is a ticket. If the signal is worthless, the user loses and the venue still gets paid. That asymmetry is the quiet story inside the press release.
The Race Nobody Can Win
Competition is where the thesis really frays. Public claims exclusivity through the Kalshi integration. That exclusivity is a contract, not a moat. Robinhood has the distribution. eToro has the social-trading DNA. TrueNorth has Public's own former leadership and an identical claim. And the loudest omission in the entire narrative is Polymarket — the crypto-native prediction market with a user base that dwarfs most compliant venues. If Polymarket opens a signal API, Kalshi's differentiator evaporates overnight.
This capability is commodity. The moment multiple firms ship the same threshold-trigger engine fed by the same kind of probability feed, the differentiator collapses to brand and licensing. Public's genuine edge is that it holds a brokerage license and a Kalshi relationship. That is a business advantage, not a technology one, and business advantages in regulated finance get copied with paperwork and patience.
Follow the value chain and you see who actually captures the upside. Kalshi sits upstream as both signal source and regulated exchange. Every Public user who trades event contracts becomes new volume and new fee revenue for Kalshi. Meanwhile Public earns incremental commission and account activity. The user earns... whatever the signal delivers, which is unproven. The pricing power in this arrangement sits with Kalshi, not with the brokerage that announced it. Public is a distribution channel dressed up as an innovator.
The Rules Are Being Written Behind the Product
Now the part the press release does not lead with. Kalshi is federally legal as a DCM, but its state-level status is a moving target. Roughly a dozen states have live litigation against it. Ohio hit it with a $5 million fine. Appellate courts have issued contradictory rulings, and there is no Supreme Court guidance to settle the question of federal preemption versus state jurisdiction over event contracts.
This is not hypothetical regulatory risk. It is realized, ongoing, and expensive.
On top of that, the CFTC issued proposed rules in June 2026 — proposed, not final. The entire legal boundary of the product sits inside an unwritten window. The author's own summary captured it cleanly: the rules are unwritten while the rails are already laid. That phrase deserves to be carved into the industry's forehead. Building execution infrastructure that depends on a legally unsettled signal source is not first-mover advantage — it is first-mover exposure.
[Continuing Core — Regulatory Mechanics]
There is a second-order regulatory question nobody is asking yet, and it is the one that will generate the lawsuits. When a broker lets an AI agent execute a trade based on an external probability feed, does the broker still owe a suitability obligation? If the agent buys a put because some thin event contract crossed 60%, and the client loses money, the client's lawyer will argue the broker failed its duty of care. The AI's decision logic is a black box by design. You cannot explain why a probability threshold is a suitable recommendation for a specific retail investor's risk profile. That is a genuine gap in the compliance architecture, and it will be litigated the first time a user loses a meaningful amount.
I want to be fair to the engineering. The execution layer is real and it works. The Kalshi integration is real. The idea of using prediction-market probabilities as a leading indicator is not stupid — it is genuinely interesting. Prediction markets sometimes aggregate information more efficiently than analyst coverage. But "sometimes" and "sometimes more efficiently" is not a foundation. It is a hypothesis. And a hypothesis wearing a product launch is exactly the kind of thing that gets retailed to people who need certainty.
The Contrarian Read: This Is a Distribution Story, Not an Innovation Story
Everyone is reading this as an AI story. It is not. It is a distribution story wearing an AI costume.
Strip the marketing and what Public actually did was negotiate access to a novel data feed and expose it to a retail base. That is a partnership, a contract, a go-to-market move. The technology is a conditional order engine. The innovation is a supply agreement. The reason the industry is calling it "agentic" is that "agentic" is the word that moves valuation multiples in 2026, the way "decentralized" moved them in 2017 and "AI-powered" moved them in 2023.
I have watched vocabulary inflate before. In 2017 I flagged economic flaws in the token models of Golem and Bancor days before their launches, and the flag was not sophisticated — it was simply looking at what the code actually did versus what the whitepaper promised. The same discipline applies here. Read the mechanism, not the label. The mechanism is a threshold and a trigger. That is fine. That is useful. It is not a paradigm shift, and pricing it as one is how retail gets hurt.
The counterintuitive angle the consensus is missing: the crypto side of this story is almost entirely decorative. Kalshi has a crypto category. It has been used to forecast Bitcoin. Circle and Binance appear as rails. Meta and Muse appear in an agent-commerce context. But none of these connections is load-bearing for the actual product. The press release lists them because adjacency signals credibility. The real coupling is between a US brokerage, a CFTC-regulated exchange, and a retail order book. Chasing the alpha while the market sleeps means noticing that the crypto narrative here is a coat of paint, not a foundation.
And if anyone thinks this displaces DeFi prediction markets, look again. It does the opposite — it validates the category while steering retail demand toward a compliant, KYC'd, fee-charging venue. Polymarket is the real competitor, and its absence from the conversation is the loudest silence in the room. If a signal-aggregation layer ever proves out, the demand will not stop at one brokerage. It will become middleware — prediction-market data APIs sold to every broker in the market. That is the actual future this launch is pointing at: a new signal-vendor industry, with the current product as its first tentative proof of concept.
What to Watch Instead of What to Buy
The thing that settles this story is not another launch. It is a number. The first venue to publish agent win rates, alpha, and drawdown over a real sample — say twelve months of live execution — will either crown this category or bury it. Until that number exists, every claim of predictiveness is a marketing artifact.
Watch the CFTC's proposed rules move from proposal to final. That document will define the legal perimeter of the entire prediction-market signal business. Watch whether the Supreme Court takes up the state-versus-federal question, because one ruling could turn a dozen lawsuits into settled law overnight. Watch whether Robinhood or eToro signs a prediction-market deal of their own, because the moment they do, Public's exclusive becomes a footnote. And watch Kalshi's volume after the integration matures, because that is the only hard evidence that any of this is real.
Scanner's note: the signal is unproven, the legal ground is moving, and the moat is a contract, not code. Human faces behind the blockchain code rarely appear in the press release — but they are the ones holding the risk when the threshold triggers.