While the market sees another AI chatbot wrapped in a startup announcement, the ledger shows something far older: a decades-old gap in America's financial infrastructure that no fintech app has managed to close. Every large corporation in the country hedges its cost of capital. More than 85 percent of the Fortune 500 use derivatives to pin down interest rates, commodity prices, and currency risk. The roughly 33 million small businesses in the United States — 99.9 percent of all firms — barely touch them. A 40-seat restaurant does not have an ISDA master agreement. A construction firm does not have a treasury desk. A hurricane warning does not wait for the budget review.
Kalshi just launched Blanket, an AI-powered assistant that turns a business owner's plain-English description of a fear into a hedge on a CFTC-regulated event contract. The announcement itself was a restrained three paragraphs. The story underneath is anything but restrained. For the first time, a prediction market is being positioned not as a betting venue or a political polling instrument, but as the safety rail underneath the daily economy. That is a new category, and it deserves more than a product roll-out cycle to understand.
I have been covering these markets since the Iowa Electronic Markets era, through the collapse of Intrade, the death of the Pentagon's infamous terrorism futures exchange, the rise of Polymarket, and the slow, methodical build-out of Kalshi. In all that time, nobody has tried to build what Blanket is attempting: a bridge between the institutional risk-management stack and the family-owned business that has been excluded from it for a century. Bridging the gap between code and community is exactly the kind of work that changes a financial system, not just a user interface.
Context: The Long Road to Regulated Prediction Markets
Kalshi was founded in 2018 by Tarek Mansour and Luana Lopes Lara with a deliberately narrow and distinctly non-crypto thesis: take the Commodity Exchange Act, the law that has regulated grain futures since the New Deal, and use it to legalize a new kind of product called an event contract. The exchange launched in 2021 as a federally licensed derivatives clearinghouse. Each contract is binary: it settles at $0 or $1. Each settlement is keyed to an official, public metric — a Bureau of Labor Statistics CPI print, a Federal Reserve statement, a National Weather Service temperature reading, the official list of named storms. There are no disputable oracles, no subjective judges, no manual claims process. The data does the deciding.
The regulatory plumbing is worth understanding because it is the foundation of everything Blanket is trying to do. Kalshi self-certifies new contracts under its CFTC license, which means it files its own rule and contract terms with the commission, gives regulators a chance to object, and then lists the product. The CFTC sued Kalshi in 2023 over political event contracts; the exchange fought back in court, won, and solidified its right to operate across the broadest possible range of event categories. By 2025, it had become something rare in the prediction market business: a licensed venue with deep legal clarity, real market making, and an active order book.
The market backdrop matters as much as the legal one. Today's market is a sideways, choppy grind. Inflation has been sticky in services, rates have stayed elevated, the weather is breaking records every few months, and geopolitical events collide with supply chains at random. For small businesses, this is a nightmare combination: macro volatility with no direction, enough shocks to hurt but no trend large enough to justify large-scale repositioning. The Fortune 500 spends this period adjusting hedges. The rest of the economy simply absorbs the punches.
Blanket is Kalshi's answer to that asymmetry. It is an AI layer that translates the language of a business owner — "I worry that a freeze in February will kill my citrus crop" — into the language of a contract specification. The tool then sizes the position, places the order on the Kalshi order book, monitors the position until settlement, and explains the outcome in plain English. The promise is nothing less than the democratization of risk management.
Core: The Mechanics, The Gap, and The Inflection Point
1. The Hedge Gap Is the Real Market
Let's start with the data that nobody in the mainstream press quotes, because it reframes the entire product. The U.S. Treasury's borrowing advisory committee has long reported that about 85 percent of large non-financial corporations use derivatives to manage macro risk. Research from the academic derivatives literature puts the figure for firms with fewer than 500 employees at under 4 percent. In a country where small firms generate 44 percent of private GDP and employ 46 percent of the private workforce, this disparity is not a footnote. It is the defining structural risk of the American economy.
Why do small businesses not hedge? The standard answer is cost, and it is true, but it is incomplete. Consider what a real hedge requires today. An interest rate swap requires an ISDA Master Agreement, which ordinarily means legal counsel, credit support annexes, and a negotiation process that can take months and cost tens of thousands of dollars — before a single swap is executed. The minimum notional for an institutional interest rate swap is often $5 million or more. A family-owned factory with $8 million in revenue does not need to hedge a $5 million notional; it needs to hedge a $300,000 line of credit tied to SOFR.
Commodity futures offer a similar version of mismatched geometry. A hardware distributor in the Pacific Northwest cannot efficiently use a lumber futures contract tied to 110,000 board feet when its quarterly purchases are half that. An options strategy on the S&P 500 is accessible to any retail brokerage account, but it doesn't hedge the specific risks of an event-driven business owner — an unexpected freeze, a port strike, a tariff announcement. Traditional insurance covers property and casualty, not macro exposure. There is no policy for "my input costs rose 9 percent and my contracts were fixed." The market has simply never been built for businesses of this scale.
I have been inside the alternative: I led a rapid-response team in 2017 that audited three ICO projects within 48 hours of their token launches, tracking tokenomics against smart-contract logic. The most striking finding was not the governance flaws we uncovered — it was the asymmetry in treasury management. The projects run by experienced financial operators hedged their treasury exposure; the projects run by strangers on Telegram treated volatility like a virtue. Two years later, only the hedgers were still standing. The same lesson scales to the broader economy, and the lesson begins with a single observation: the instruments exist, the infrastructure does not.
Based on my audit experience, the small business hedge gap is not a data problem. It is an access problem. The solution has to be an interface thick enough to protect a beginner from the complexity of derivatives, while still sitting on top of real, regulated infrastructure. Which is exactly what Blanket claims to be.
2. The Anatomy of a Modern Event Contract
Before Blanket can be evaluated, the instrument itself must be understood. Event contracts are the direct descendants of the first financial derivatives — the standardized grain contracts introduced at the Chicago Board of Trade in 1848. Before standardization, a farmer and a grain elevator negotiated every shipment as a unique handshake; there was no reference price, no resale value, no efficient transfer of risk. Standardization solved this: one bushel of #2 corn in a commodity contract is the same as another, and therefore transferable. The contract became a language for risk — and each leap in financial infrastructure since has been a new layer of standardization.
Kalshi's event contract standardizes a different thing: not a commodity but an event. Take a contract like "Will CPI print above 3.0 percent year over year at the May 2026 release?" The contract trades in a band between 0 and 1. If the market prices the contract at 0.25, it implies a 25 percent probability of a high CPI print. A small business owner who is hurt by inflation buys the contract; if the print comes in high, the contract pays $1 and offsets the real-world cost. The contract is cash-settled, requires no physical delivery, and settles on a data point that no party can manipulate: the Bureau of Labor Statistics release schedule.
The mechanics are more like a futures exchange than a crypto AMM. Orders rest in a central limit order book. Market makers quote two-sided prices. Collateral is segregated in accounts under CFTC supervision. Settlement is deterministic. There is no counterparty hopping, no one holding your collateral can run away with it, and the CFTC has the power to suspend or reject products it considers problematic. While I have spent the better part of a decade writing about smart-contract custody and the spectacular failures that occur when code and treasuries collide, there is a blunt truth worth stating: the regulatory plumbing of a licensed U.S. exchange is the most reliable settlement layer that exists in this country today. The ledger remembers what the hype forgets, and the ledger of Kalshi has never lost a settlement because of a hack.
For the small-business user, the genius of the event contract is simpler. On the settlement date, the outcome is public and automatic. Either the event happened or it did not. No one needs to call a claims adjuster, no one needs to prove damages, and no one needs to fight for an exception. This is the insurance product that turbocharges the speed of a derivatives contract with the determinism of a weather forecast. And it is the foundational building block on which Blanket is built.
3. How Blanket Actually Works: The Five Layers
The announcement of Blanket was deliberately lean, which means most of the real architecture is interstitial — available from Kalshi's developer-facing materials and the pattern of how the tool behaves on the live platform. What follows is a reconstruction of the stack, layer by layer, because the details determine whether this is a toy or a turning point.
Layer one is natural-language parsing. A business owner types, "I run a food truck fleet in Houston and a freeze or a flood in February would hit my revenue hard." The parsing engine isolates the entities: geography (Houston), hazard class (freeze/flood), business type (outdoor-events-dependent, weather-sensitive). This is not a novelty search; it is a structured extraction of risk-relevant semantics, trained on data from insurance claim files and financial planning documents rather than on generic internet text.
Layer two is risk-to-contract mapping. Kalshi maintains a matrix of listed markets spanning CPI, Fed decisions, inclement weather, natural disasters, energy prices, elections, and crypto prices. Blanket scores the correlation between the user's described exposure and each market using both historical data and contemporary option analytics. A Houston food truck fleet maps strongly to a freeze/winter storm contract; it maps not at all to a Fed decision. The model understands that a freight-based manufacturing firm might need a layered position across energy and CPI, while a regional builder needs only a temperature contract for the construction season.
Layer three is position sizing. This is where retail trading tools usually fail, because they answer "what should I buy" but ignore "how much." Blanket computes a suggested hedge based on the user's stated revenue exposure and the current contract pricing. It scales the position so that the premium, or the collateral at risk, stays within a small single-digit percentage of the business's monthly operating expenses. In effect, it is running a miniature value-at-risk model from a single conversation, with a built-in risk ceiling that prevents the user from treating a hedge as a lottery ticket. The position sizing logic, in my view, is the most underrated part of the product. In my years on trading desks and in protocol audits, I have watched more retail accounts lose money to overdosing on position size than to the wrong underlying instrument.
Layer four is execution. The order is sent to the Kalshi central limit order book as a limit order, not a market order. The tool waits for the price or uses a quoted spread that respects the current depth. In a market with a 3-cent bid-ask spread, a market order on a 0.20 contract can cost you half your premium if you land on the wrong side. Every cent of slippage is a cent of hedging cost, and Blanket's quiet automation of patient limit orders is the difference between a viable consumer product and a re-indexed fee engine for market makers. These are details nobody writes press releases about, and they are exactly the details that determine long-run survival.
Layer five is the monitoring and explanation loop. After the trade is placed, Blanket stays on. It tracks the approaching settlement date, alerts the user to key pricing moves, and explains the position in human terms while it matures. At settlement, it identifies the payout, estimates the tax consequence, and frames the outcome honestly. A hedged loss is presented as the cost of an insurance premium that expired unused — not as a failed bet. This is not just pedagogy. The language we use to describe risk has real economic consequences, and a tool that teaches fear-management is building compound utility.
Core insight: The innovation of Blanket is not the AI generation — it is the abstraction layer over the regulatory plumbing of a CFTC-licensed derivatives exchange. The magic is that you do not have to know what a binary option is to hedge like a CFO.
4. The AWS Moment for Risk
Now that we understand the stack, the strategic thesis deserves a crisp restatement. Blanket is not a product. It is the prototype for a category that economics has long assumed could not exist: retail-accessible, contractually complete hedging for the long tail of the economy.
The analogy that best captures this is not a financial one; it is Amazon Web Services. AWS did not invent computing. It invented provisioning — taking the brutal operational problem of buying, racking, and maintaining servers and turning it into an API call. The consequence was that a three-person startup could access the same infrastructure as a multinational bank, not because it owned that infrastructure, but because the complexity was abstracted away on the other side of an interface.
Blanket is attempting the same settlement for risk. It turns "managing macro risk" into a prompt: type your fear, get a hedge. The underlying infrastructure is a regulated derivatives exchange; the abstraction layer is an AI that understands you. The abstraction is the product. The complexity is the moat.
The market will underestimate the compounding effect of this. Every small business that hedges adds liquidity to the event contract order books. Deeper books tighten spreads. Tight spreads lower hedging costs. Lower costs attract more hedgers. Kalshi becomes the only regulated venue operating this loop, and it owns the entire flywheel — the order book, the settlement engine, and the AI layer that feeds it. That is a powerful flywheel, and it is why the phrase "essential infrastructure" is not hyperbole; it is positioning, and positioning is the real battleground.
Narratives move markets faster than blocks, and the narrative here is doing something interesting. The market has spent a decade treating prediction markets as entertainment or political insight. Blanket is trying to rebrand them as a utility, like a thermostat or a payroll system. If that narrative grips, the valuation of Kalshi stops being a function of trading volume and starts being a function of the number of small businesses that count it as part of their monthly operating routine.
5. AI, Fiduciary-Adjacent Technology, and the Trust Problem
Here we reach the part of the announcement that worries me most, and the part where responsible coverage must draw a bright line. Blanket is an AI tool that gives financial advice. It tells an untrained business owner to buy a specific contract, at a specific size, at a specific price. That makes it fiduciary-adjacent technology: not a fiduciary, but close enough that a hallucination has real consequences.
AI hallucination is not a theoretical sports bet. It is an error mode that follows from the architecture: a model does not "know" the answer; it generates the most probable sequence of words. When uncertain, it can confabulate confidently. A user tells Blanket "I'm worried about gas prices next winter." The model maps that to a heating-degree-days contract in a different region, because the semantic similarity is high but the statistical correlation is low. The user pays a premium for a hedge that is actually not correlated with their exposure. This is not a bug in the parsing layer; it is an epistemological limit of language models.
The mitigation is well understood. Model validation, backtesting of recommended hedge ratios, adverse-example testing, and fail-safe disclosures that the tool is not a substitute for professional advice. Based on my work convening the Consensus Protocol for AI Trust in early 2026 — a roundtable with ten leaders from exchanges, regulators, and AI labs — the emerging standard is clear: any tool that recommends a financial transaction must maintain a complete decision trail. There must be a logged audit of the input, the model version, the recommendation rationale, and the execution. Otherwise, when a small business loses money and asks 'why did your AI tell me to do this?', there will be no honest answer.
The liability question is genuinely unexplored. If Blanket gives a bad recommendation and a coffee shop in Austin loses its rent, who is sued? The exchange? The AI vendor? The user who clicked "accept"? In traditional investment advice, the advisor has a fiduciary duty. Blanket carries no such duty — it is a tool, not an advisor, and the terms of service will say so, loudly. That could be the product's biggest exposure.
Crypto-native prediction markets never had to answer this question, because they never claimed to serve the unhedged economy; they served sophisticated traders who knew the risks. Blanket is walking into a far more dangerous territory: the frontier between code and community, where a single bad recommendation can destroy a decade of trust. Transparency is the only consensus that lasts, and if Blanket is not transparent about its own model errors, it will lose the small-business segment just as it starts to win it.
6. Three Case Studies in Unhedged America
The abstraction becomes real when you attach human names and business plans to it. Let me walk through three composite cases, drawn from conversations I've had over my years in this beat, because they represent the pattern of risk exposure that traditional financial infrastructure never serves.
The restaurant owner in Miami. Maria runs a 40-seat eatery a block off the waterfront. A named storm can shut her down for three weeks when the power grid fails, employees can't commute, and tourism evaporates. She cannot buy business-interruption insurance at an affordable premium, because her historical claims record is too thin for the underwriter. Blanket maps her fear to a hurricane contract that settles on whether a named storm makes landfall within a specific coastal zone. The contract trades at 0.12. She buys it for a few hundred dollars, an insurance-equivalent premium. The season ends quiet; the contract settles worthless; she is out the premium. To an observer, the hedge failed. To Maria, it bought a winter without panic, a spring without the temptation to hoard cash in a way that would have hurt her regulars' hours. The ledger remembers what the hype forgets: managing fear is an economic output, even when the contract never pays.
The construction firm in the Northeast. David bids fixed-price contracts in January for projects that will not start until August. If steel prices jump, or a tariff is announced mid-construction, his margin vanishes. He can't easily use steel futures with their large notional. Blanket maps him to a contract on a broad commodity index or a tariff-escalation event. The correlation is far from perfect, but it converts a 100 percent tail exposure into a 60 percent tail exposure. For a firm of his size, cutting the tail in half is the difference between surviving the year and closing the year down.
The coffee roaster in Austin. Priya buys green coffee from two origins and sells under fixed contracts to wholesale clients. Her input costs track shipping, labor, and packaging, all of which move with CPI. Kalshi's CPI contracts are the closest available instrument. She hedges a portion of her six-month exposure by buying a yes position on a high CPI print. If inflation roars, the contract pays out, offsetting exactly the cost increase she could not pass on in her fixed-price contracts. Culture is the new collateral: because her hedge is in place, she can keep her wholesale prices stable through an inflationary year, reinforcing her community relationships rather than shredding them. That is the kind of invisible underwriting that makes small business communities cohere.
These cases are not the product of a salesman's imagination. They are a direct mapping of a family of needs that has existed for fifty years and never found a supplier. For as long as the cost of risk management exceeded the risk itself, this market segment was unserved. Blanket is the first serious attempt to serve it at a price point that fits.
7. The Competitive Ledger
Kalshi is not racing alone. The competitive field rewards the disciplined, and it punishes the unprepared.
Polymarket remains the liquidity king of prediction markets, with volumes that dwarf Kalshi's in elections and sports. But Polymarket operates outside U.S. regulatory reach, has no small-business onboarding, no compliance wrapper, no fiat infrastructure for a 60-year-old restaurant owner, and no obligation to protect a retail user from themselves. It is a trading venue, not a utility.
The decentralized protocol world — Augur, Azuro, and a dozen smaller protocols — remains technically interesting but productively irrelevant for this specific thesis. They offer self-custody and open order books, but they fail at every point where Blanket succeeds: they have no abstraction layer for non-technical users, no legal clarity for U.S. consumers, no stability of settlement, and no duty of care. Decentralization is a mindset, not just a metric, and a protocol that lets you lose your entire collateral because you misconfigured a slippage setting is not infrastructure; it is a hobby.
There are also adjacent products from the insurance-tech world. Parametric insurance providers like Arbol have built branded weather- and climate-index products for agriculture. That is a direct analogue, and worth watching, but the product is not truly event-contract based; it is still packaged insurance with a policy document, claims infrastructure, and underwriting overhead. Blanket's advantage is speed and granularity: an exchange with dozens of products listed and instant settlement at a click.
The bigger threat is not a competing prediction market; it is a distribution partnership. Kalshi has the exchange, the AI, and the regulatory license. What it does not yet have is the embedded channel. If a payroll company like Gusto, a banking API like Plaid, or an accounting platform like QuickBooks integrates a "hedge your risks" module powered by Kalshi, the flywheel turns much faster. The next twelve months of Blanket's life are less about the product and more about who signs distribution deals.
8. What This Means for Crypto
The crypto world should be paying especially close attention, because Blanket exposes something uncomfortable about the ecosystem's priorities.
Blockchain-based prediction markets were supposed to be the killer app of decentralized information aggregation. Oracles, on-chain settlement, permissionless access — everything was in place, technically. But the industry spent its talent on building more complex financial Legos rather than articulating the user. Uniswap V4's hooks turn the DEX into programmable Lego, but the complexity spike will scare off 90 percent of developers and 100 percent of normal business owners. The crypto-native abstraction layer never arrived; it was replaced by a development abstraction layer, which is entirely different.
Kalshi, a traditional CFTC-regulated company, just built the abstraction layer that crypto promised. It did so not with a token, but with an AI model and a central order book. This is a humbling fact for the decentralized finance community. The output — event hedging for small businesses — is not itself censorship-resistant or permissionless. But for the unhedged economy, a tool that works, is legal, and is understandable beats a protocol that is elegant, decentralized, and unusable.
The deeper convergence is the one I spend my waking hours on: AI plus crypto plus financial inclusion. The decentralized AI crypto stack has been stuck on agent-to-agent payments and compute markets, which are real but small. Blanket shows an alternate path: AI as the interface, not the asset. The prediction market is the source of truth less because it is on-chain and more because it is transparent, fast, and settled. The underlying ledger could be a CFTC database for years before it is a blockchain. And that is okay. The sprint ends, but the chain remains — and the chain that will remain here is the order book's integrity, not the immutability of the settlement layer.
Contrarian: The Dangerous Side of Essential Infrastructure
Now the contrarian view, because predictably, there is one hiding in the ledger.
First, the infrastructure story has a single point of failure, and its name is the CFTC. Every regulatory victory Kalshi has won can be reversed by a shift in commission leadership, a political scandal, or a court decision. Essential infrastructure that depends on a single regulator's permission is not essential; it is contingent. This is the paradox of building utility within regulated walls — the walls can close at any moment, and prediction markets have a long history of being regulated into nonexistence. The Pentagon's terrorism futures market was killed by a Senate vote in 2003, days before launch. Prediction market history is littered with graveyards of legal venues that no longer exist.
Second, an event contract is not a hedge; it is a substitution of one risk for another. The small business owner who buys a CPI contract substitutes her real, idiosyncratic cost risk for a basis risk — the risk that the official CPI print does not match her actual cost structure. The tool is honest about this in its user interface, but the mathematical basis risk in an illiquid event contract can be worse than the original exposure. In thin markets, the "hedge" is actually a bet against a few professional market makers. And professional market makers have information advantages, because they see the full order flow while the business owner sees only her own balance sheet. A retail hedger in a thin market is not a farmer buying grain futures; she is a tourist buying a trinket in a pricing environment controlled by locals.
Third, the announcement's biggest unstated risk is data. If Kalshi becomes essential infrastructure, its order books become a real-time map of the unhedged economy's anxieties. The CFTC, law enforcement, and tax authorities can access that data. A tool that democratizes risk management also becomes a surveillance mechanism for businesses too small to have treasury departments. The AI that listens is also an AI that records. Prediction markets are celebrated for aggregating information; the flip side is that the information aggregated contains the commercial secrets of its least sophisticated participants.
Fourth, and most uncomfortably: the tool could be too successful too quickly, attracting a wave of small businesses that do not understand that a hedge is an expense, not an investment. A hedge that expires worthless feels like a loss, and a small business owner who feels fleeced will default to exactly the panic behavior the tool is designed to prevent. Adoption could curdle into resentment, and the phrase "hedge as insurance" could become the new "NFT as investment." The ledge doesn't care how carefully you walked it; it cares about the fall.
None of these risks kills the thesis. But they mean the thesis must be pursued with an almost monastic commitment to transparency and user education. The tool must teach its users what a basis point is, what a hedge is not, and what the CFTC can and cannot do, before it shows them a single button. The first wave of users will determine whether Blanket becomes a utility or a cautionary tale.
Takeaway: What to Watch Next
The sprint ends, but the chain remains. Over the next twelve months, the test for Blanket is not trading volume or app downloads — it is a single question: can a 40-seat restaurant owner, one month after buying her first hedge, explain to a skeptical spouse what the hedge did and why she bought it? If the answer is yes, prediction markets have crossed the line from bet to infrastructure, and Kalshi has become the default settlement layer for the unhedged economy. If the answer is no, Blanket will be remembered as an elegant product that could not teach its own value proposition.
Watch the retention charts, not the volume charts. Watch whether the AI's explanations are honest about basis risk. Watch whether the small business owner understands that a hedge is a premium paid, not a lottery ticket. Trust takes a decade to build and a single bad contract to destroy; the ledger remembers what the hype forgets. The unhedged economy has waited fifty years for this tool. It deserves an interface that is as honest as the ledger it sits on. Transparency is the only consensus that lasts — and for the first time in a generation, the consensus is within reach.