Over the past seven days, another DeFi protocol lost 40% of its liquidity providers. I don’t know which one. The point is that in this bear market, there is always a candidate. We chase TVL, we chase APRs, we chase the next hook in Uniswap V4, and we keep missing the real signal. The signal is that the most interesting experiment this month is not a chain, not a rollup, and not even a token. It’s Blanket, an AI-powered risk-analysis tool launched by Kalshi on August 7. We didn’t need another L1 to see where crypto was heading. We needed a small tool that doesn’t pretend to be infrastructure.
Blanket does one narrow thing. It asks a small business about its big headaches: snowstorms, gas bills, tariff surprises, maybe an election result that flips policy. Then it recommends an event contract on Kalshi that could partially offset the financial damage if that headache materializes. It does not execute the trade. It does not touch the money. It does not require a wallet. It is a referral layer, a matching engine, a digital risk consultant that lives on top of a regulated market. And because it refuses to be a protocol, it might teach crypto more about distribution than any new consensus mechanism ever could.
The product was built by Lauris Zminsky, an independent fintech entrepreneur, not by Kalshi’s internal team. That detail matters more than any benchmark. It tells us Kalshi is not just opening an API; it is opening a platform. It tells us that the regulated exchange understands something many crypto projects still refuse to accept: the next wave of adoption will not come from infrastructure. It will come from tools that hide infrastructure completely.
The End of the Chain Fantasy
Let me start with the confession of an recovering infrastructure maximalist. In 2017, I was a junior consultant in Chicago, and I spent a disproportionate number of nights reading Vitalik’s ZK-SNARKs papers. I built a crude proof-of-knowledge demo in ZoKrates. I wrote a Medium post called “Why Mathematics is the New Social Contract” and somehow got invited into a DAO that was trying to build decentralized identity. I was convinced that the future would be determined by which base layer could prove the most truth with the fewest trusted parties. I was wrong about the layer, but right about the craving.
The market does not reward cryptographic novelty. It rewards the reduction of uncertainty. ZK proofs reduce computational uncertainty. Prediction markets reduce event uncertainty. Blanket takes the second kind of uncertainty and wraps it in the first kind of language, but the language is not the product. The product is a tiny bridge between an abstract event contract and a concrete business pain. That is why I keep coming back to Blanket, because it is the anti-protocol. It is not trying to become the settlement layer. It is trying to become the layer below a decision, the layer above a fear.
We have spent ten years arguing about whether decentralized rails are better than centralized rails. The argument is beside the point. Identity isn’t about a wallet address; it’s about the obligations that address accepts. The same logic applies to risk markets. A small bakery in Ohio does not care whether the storm hedge settles on a sovereign chain or a CFTC-regulated exchange. It cares whether the bakery can keep its doors open after a tornado. Blanket is built for that bakery. It is not built for us, and that is exactly why it matters.
The Kalshi Context
Kalshi is a designated contract market, which means it operates under CFTC oversight. It lists event contracts on things like weather in specific cities, monthly energy prices, inflation prints, tariff decisions, and election outcomes. During the 2024 election cycle, its volume exploded. Traders used it to express views on the White House race, on congressional seats, on policy probabilities. It became the more sober cousin to Polymarket, still experimental but with a legal wrapper that institutional users could bite into.
The election boom created a narrative, but also a problem. A market that lives and dies by the election calendar is a market with a severe seasonal dependency. Once the votes were counted, Kalshi had to find a way to stay interesting in an off-cycle year. The obvious answer was to expand from speculation to risk management. Instead of asking “who wins,” the product could ask “what does that win mean for your business?” Instead of attracting retail prediction junkies, it could attract a completely different audience: procurement managers, fleet operators, food distributors, independent pharmacies, construction firms, anyone whose margin depends on weather and policy.
Blanket is the first public expression of that expansion. It is a third-party tool built on Kalshi’s architecture, but it is also a test. Kalshi is testing whether its event contracts can be reimagined as hedging instruments for Main Street. It is testing whether an independent developer can build a viable vertical product without needing to become an exchange, a custodian, or a clearinghouse. It is testing whether the platform play works in a heavily regulated environment, where every recommendation has a legal smell to it.
For crypto people, the easy reaction is to dismiss Blanket as a centralized chatbot on top of a regulated order book. That reaction is lazy. The more honest reaction is to recognize that this is exactly what mainstream risk adoption will look like: not users on-chain, but users on apps that happen to be connected to an exchange that happens to be transparent. The underlying rails might be slower and permissioned, but the user experience is gentle. That gentle experience is the hardest thing crypto has never been able to ship.
The Architecture of Blanket
Let’s get technical, because the details reveal the strategy. Blanket sits at the application layer. It is not a protocol. It has no consensus mechanism, no native token, no validator set, no bridge, no sequencer, no fraud proof. It is a service that consumes data, applies a recommendation model, and returns a structured list of Kalshi event contracts that seem correlated with the user’s stated exposure. This is a decisive design choice. Blanket is deliberately boring.
Based on the public information, Blanket likely pulls two kinds of inputs. The first is Kalshi’s market data, either through the Embedded API that Kalshi promotes or through public market endpoints. That gives Blanket a real-time catalog of live contracts, bid-ask spreads, historical prices, and implied probabilities. The second input is external macro and weather data: temperature forecasts for a city, heating degree days, storm warnings, tariff announcements, perhaps election polling aggregates. The tool then runs a matching service that maps the user’s inputs to a set of contracts that look, statistically, like they could behave as a hedge.
The phrase “AI” in the marketing copy is doing a lot of work. My experience with AI-labeled crypto products is that “AI” often means a small rules engine with a chatbot skin. Blanket almost certainly follows that pattern. An LLM gives the user a friendly conversational interface, but behind the interface there is probably a deterministic mapping: risk factor in, contract category out, confidence score attached. That is not a criticism. That is engineering discipline. The most valuable tool in financial risk management does not need to be a giant neural network. It needs to be an honest calculator that can clearly explain why a specific contract is relevant. In a regulatory environment, the ability to explain a recommendation is more important than the ability to optimize it in ways no human can understand.
People often ask me where the magic is in decentralized systems. I used to say it was in the cryptographic proof. Over time, I learned that the magic is in the social contract that makes the proof matter. The same lesson applies to Blanket. The magic is not in the LLM. The magic is in the decision to locate the safety boundary at Kalshi. Blanket does not hold funds, so it cannot be hacked for funds. Blanket does not execute trades, so it cannot be accused of manipulating a market. Blanket only recommends. By separating advice from execution, the product dramatically narrows its own attack surface. It also radically simplifies its compliance story, though not as much as its creators might hope.
The technical design is deliberately boring, and that boredom is the most bullish signal in the entire announcement. The innovation here is combinatorial, not foundational. It takes a regulated exchange, a language model, some weather data, and a small-business problem, then snaps them together like LEGO bricks. That is how innovation spreads after the infrastructure wars end. The winners stop inventing new protocols and start inventing new workflows.
The AI Behind the Curtain
Let me be blunt about the AI risk. The original announcement does not include benchmark numbers. There is no statement about recommendation accuracy, no confidence interval, no backtest, no independent audit, no model card. For all we know, Blanket’s first version is a thin wrapper over a handful of hand-coded correlations between weather zip codes and Kalshi’s existing city-level contracts. That could be useful, but it is not what most people picture when they hear “AI-powered risk analysis.”
This opacity is uncomfortable, but not surprising. In 2022, during the deepest part of the bear market, I published a report on underappreciated builders. I identified about fifteen projects that had high code activity and low price correlation. The main thing I learned from that exercise is that most crypto teams overfit their narrative to what VCs want to hear. They say “AI” because AI commands attention. They say “decentralized” because decentralization commands political support. But the actual software is often a dashboard plus an API. Blanket might be the same: a dashboard plus an API plus a cleverly designed conversation flow. That does not make it bad. It makes it normal.
What should worry us is not the absence of a sophisticated model. What should worry us is the absence of a mechanism to challenge the model. If Blanket recommends a weather hedge and the hedge fails because the contract’s underlying index does not match the user’s actual location, the user can lose a meaningful premium. There is basis risk here. A snowstorm in the city center might not trigger a contract indexed on a particular airport weather station. A tariff event might cover one product category but not the exact input the business imports. These are not black-box failures; they are ordinary contract mismatches. But when an AI chatbot recommends a contract, users assume a degree of intelligence that a rules engine cannot deliver.
The first real update from Blanket should not be a feature; it should be an audit. I would love to see a public file that lists every risk category, every model input, every recommendation logic rule, and every known failure case. That is the kind of transparency that would separate Blanket from the endless stream of crypto AI vaporware. Without that file, Blanket is a marketing experiment that happens to be legal, not a trustworthy risk tool.
Still, I have to keep myself honest. Crypto has never been a model of transparency when it comes to recommendation algorithms. We throw “non-custodial” around as if it were a synonym for trustworthy, but then we ask users to click buttons on random interfaces that can approve unlimited token allowances. Blanket does not have that problem because it does not have access to funds at all. In a weird way, the tool’s lack of ambition is its greatest safety feature. It cannot rug you because there is nothing for it to hold. It cannot execute a bad trade because it cannot execute anything. Its worst case is that it gives you poor advice, and you are always free to ignore it. That is a better safety model than most crypto protocols can claim.
Tokenless Economics
Let’s talk about the most uncomfortable part of Blanket for crypto native readers: there is no token. No BLANKET. No staking. No yield farms. No governance proposal that sends rewards to treasury wallets. No unlock schedule. No venture round with a 12% token allocation. Nothing.
This is not a missing feature. It is a structural decision. Kalshi is a regulated exchange that makes money from transaction fees. It does not need a token to coordinate its market makers. Blanket is a third-party tool, and its business model is undisclosed. It could charge a subscription. It could take a referral commission from Kalshi. It could give advice for free while collecting lead-generation fees from insurance brokers. All of those models are invisible to the public right now, and that ambiguity is a signal in itself: this is not a crypto-native startup trying to pump a token. It is a fintech experiment trying to prove a distribution thesis.
The absence of a token eliminates a whole class of crypto risks. There is no inflation tax on users. There is no governance attack surface. There is no incentive to farm liquidity that will exit when the emissions stop. There is no liquidation cascade. The product’s survival depends entirely on whether it can create enough value to retain users or enough attention to attract a strategic partner. That is a cleaner economic model than most protocols I analyze as a DAO governance architect. It is also a harder model to succeed in, because nobody can dump the failure onto token holders.
In a bear market, no token is a feature. We always say we want protocols to reduce financial abstraction and focus on real usage. Then we reward tokens that distract from real usage. Blanket reminds us that value capture can be simple. Kalshi captures value through fees. Blanket captures value by making those fees more likely. If a user follows a Blanket recommendation and buys an event contract, Kalshi earns a fee. If Blanket eventually earns a cut of that fee or a fixed subscription, it becomes a SaaS business. That is not exciting crypto. It is boring fintech. Boring fintech is what a small business wants to buy.
Liquidity isn’t about TVL. It’s about whether a business can close a hedge before a storm hits. Blanket is a liquidity introduction service, not a liquidity pool. In a world where we are desperate for sustainable revenue, that might be the most important clarification we can make.
One more point about tokenomics: because there is no token, there is no way for the market to price the future adoption of Blanket. That makes the product almost invisible to crypto investors. They cannot buy a piece of it. They cannot speculate on its user growth. They can only watch, from the outside, to see whether a non-tokenized fintech product on a regulated exchange can outlive the crypto hype cycle. That is a useful experiment. It tests whether adoption can happen without the financialized incentive structures that crypto projects default to. If Blanket succeeds, it will quietly prove that the “Token as incentive” doctrine is optional. If it fails, we will blame the distribution channels, not the lack of a token.
The Ecosystem Play
Blanket’s biggest strategic significance is not the product itself; it is what the product reveals about Kalshi. Kalshi is becoming a platform. By allowing an independent developer to build a vertical AI tool on top of its contracts, Kalshi is adopting an App Store model. That is a massive shift. A designated contract market no longer has to build every user-facing product internally. It can open its rails to entrepreneurs who understand industries that Kalshi does not understand.
This is the same move that made traditional exchanges powerful. Think about the growth of app ecosystems on top of payment rails. Venmo did not need to re-invent the bank; it integrated with the bank. Stripe did not need to be a bank; it wrapped banking rails with developer tools. Blanket is doing the same thing to Kalshi. It is wrapping regulated event contracts with a small-business risk layer. If this works, Kalshi becomes not just a marketplace but a developer platform for financial risk. That would be a moat that token-based prediction markets would struggle to match, because the trust layer is not in the smart contract; it is in the regulatory frame.
From Kalshi’s perspective, Blanket is an option-type ecosystem investment. Kalshi pays nothing upfront, takes some reputational risk by allowing its name to be associated with a third party, and gains the upside of a new vertical without the cost of building a sales force for small business. If Blanket fails, Kalshi loses essentially nothing. If Blanket succeeds, Kalshi gains a channel that feeds volume back into its own event contracts. That is a smart optionality trade. Crypto protocols rarely do this because they are too busy trying to become conglomerates. Every DAO wants to build everything: a social token, a debit card, a metaverse, a bridge. Very few are willing to let outsiders build on their rails and share the result.
The platform analogy extends further. Blanket might be one of many future Kalshi Embedded apps. A developer could build a tool for farmers. Another could build a tool for energy traders. Another could build a tool for logistics companies that move freight through tariff-affected corridors. Each of these tools is small. Each of them needs deep industry knowledge. Kalshi cannot staff all these verticals, but it can provide the API, the clearing structure, and the compliance umbrella. The third-party developers provide the relationships, the language, and the go-to-market motion. This is how a regulated exchange can evolve without losing its regulatory anchor.
Crypto infrastructure has looked for network effects by trying to become the base layer. Kalshi is trying to become the base layer by being the most reliable settlement point underneath a thousand small applications. It is a different path to the same kind of domination: not through trustlessness, but through trust minimization and ease of use. Blanket is the first visible evidence that this path has real momentum.
The Compliance Question
The hard part is not technology. It is the legal distinction between providing information and providing advice. Blanket may not execute trades or handle funds, but it does recommend specific event contracts. That is where the Commodity Futures Trading Commission starts to pay attention. A person or firm that gives advice about commodity futures, swaps, or options may be considered a commodity trading adviser, a CTA, and may need to register with the CFTC or qualify for an exemption. Event contracts are not futures in the traditional sense, but they live in the regulated world of the CFTC. The boundary is messy.
The biggest regulatory risk for Blanket is not that it will be treated as a security; it is that it will be treated as an unregistered commodity trading adviser. This is the kind of nuance that crypto-native analysis frequently misses. The Howey test is easier to clear. There is no common enterprise in a bilateral event contract. The profit comes from the outcome of the event, not from the efforts of Kalshi or Blanket. So the security label is unlikely. But the CFTC’s rules around recommendations are broader and more ambiguous. If Blanket charges a fee or receives referral compensation for its recommendations, it is doing something that looks an awful lot like advisory activity.
The fact that Blanket does not execute orders and does not custody funds is likely a deliberate compliance isolation design. It keeps the tool closer to the territory of “publishing information” than “providing investment services.” But publishing information that is tailored to a specific user, for a specific identified risk, and that recommends a specific contract, is not the same as publishing a newsletter. A court or regulator could look through the framing and say that this is individualized risk advice. If that happens, Blanket would need to register, restructure, or limit its language to avoid crossing the line.
This is not a hypothetical. In the 2025 AI-Governance work I did with an AI ethics lab in Chicago, we spent months thinking about the line between automated decision support and automated advice. The line matters because responsibility attaches to advice in a way it does not attach to a generic map. A map can show you a road. But if you point to a specific road and say “take this one because I know where you live and where you need to go,” you have changed the contractual relationship with the user. Blanket is doing the latter. It knows the user’s risk profile and it points to a specific contract. The label on the product does not change the legal nature of that interaction.
There is also the political landmine of election contracts. Kalshi fought the CFTC over election contracts and eventually won access to list them. Blanket lists election-related hedges as a category for small businesses. That makes sense in a world where tariffs, trade policy, and regulatory decisions can swing on an election result. But election contracts are politically sensitive. A tool that encourages small businesses to hedge elections could be framed as financializing democracy. It could attract media scrutiny and maybe a round of regulatory attention. Losing that attention would not kill Kalshi, but it could slow the platform ecosystem down right when it is trying to grow.
My hope is that Blanket publishes a detailed compliance risk statement. I would love to see it lay out whether it has obtained legal advice about the CTA question, whether it charges a fixed subscription rather than success-based fees, and whether it has a process for refusing a recommendation in ambiguous situations. That is the kind of transparency that would raise the bar for every crypto AI project. But if Blanket has done that work, it has not made it public. The silence is a risk. It may mean the team is relying on a “we are just an informational tool” interpretation that could be challenged later. It may also mean the team has not yet hired the right compliance lawyer. Both possibilities are unsettling.
The Adoption Mountain
Let’s talk about the market that really matters: small business owners. They are not traders. They do not know bid-ask spreads. They do not want to learn what a basis point is. They wake up thinking about payroll, insurance, inventory, and the weather. A tool that asks “what are you worried about?” and then answers with a clear, low-friction recommendation is something they might actually use. But the distance from “might use” to “will trust with real money” is enormous.
First, there is an education problem. Event contracts are legally still new to most people. The word “contract” can scare a procurement manager. The phrase “you might lose the premium if the event doesn’t happen” is a cold bucket of water. Blanket has to explain this without killing the mood. It has to turn abstract event probabilities into something that feels like buying insurance. Insurance is a product that small businesses understand. They pay a premium, and if a bad thing happens, they get paid. Event contracts can be similar, but they are not identical. There is basis risk. There is no actuarial guarantee. And there is no agent on the other side who will handle the paperwork.
Second, there is a trust problem. Crypto-native people say “trust the code.” Small business owners say “trust my accountant.” The fastest way to reach them is not a digital ad; it is through a trusted intermediary. Insurance brokers, independent financial advisors, and accountants are the bridge. Blanket should be building for them, not directly for business owners. The actual user might be a broker who uses Blanket to quickly structure a storm hedge for a client. The business owner just signs off. This is the channel thesis. If Blanket can embed itself into brokerage workflows, it can grow without ever having to teach a bakery owner how to use Kalshi’s interface.
The distribution path is the hardest part of this entire project. Kalshi can provide liquidity and clearing. Blanket can provide the recommendation interface. But neither Kalshi nor Blanket has a sales team in every region, knocking on doors of small businesses. That requires partnerships with insurance agencies, trade associations, commercial banks, maybe even utility companies. The technology stack is easy. The last mile is brutal.
During the 2020 DeFi summer, I ran weekly Governance Jam sessions with people who were excited about liquidity mining but not about governance. We managed to increase voter turnout by 40% in one quarter. The lesson I took from that is that participation is not an interface problem; it is an incentive and habit problem. The same is true for a small business using Blanket. The interface can be perfect, but the user will not open it unless there is a clear moment of need: a forecast of a deep freeze, a tariff notice, a quarterly budget review. Blanket needs to be present at that moment. That means integrations, notifications, calendars, reminders, and a human relationship. None of those things are blockchain infrastructure.
In 2022, I studied projects that kept building during the crash. The one pattern that predicted survival was not the technology; it was whether the team had a distribution channel that existed before the protocol. The teams that had a community, a consultancy, a media platform, or a sales partnership were the ones that survived. Blanket has a strong platform partner in Kalshi, but it still lacks a distribution channel into the small-business world. That is the gap I would be watching. If the next announcement from Blanket is a partnership with an insurance broker network, I’ll be more excited than if they add 10,000 new AI features.
The Contrarian Read
Now I have to turn the lens on my own community, because that is what contrarian thinking requires. The crypto response to Blanket will sound like this: it is centralized, it is regulated, it is not a smart contract, it is not censorship-resistant, it is just an API on top of a CFTC exchange. All of that is true. But the more important truth is that Blanket might succeed anyway. And if it succeeds, it will not validate crypto; it will challenge crypto to grow up.
Prediction markets were supposed to be one of the killer applications of decentralized governance. We called them truth engines. We believed that on-chain event contracts would outperform their centralized cousins because they would be global, transparent, and free from jurisdictional limits. Then the election cycle came, and Polymarket captured the attention while Kalshi captured the regulatory cover. Blanket is now trying to extend Kalshi’s reach by going vertical. If that works, the prediction market narrative will move from “open protocol” to “trusted platform with trusted assistants.” That is the opposite of what most crypto idealists want. Yet it might be exactly what the market wants.
Freedom isn’t the absence of gates; it’s the presence of consent. Blanket is gated by Kalshi, by the CFTC, by a legal structure, and by language models that are not open source. But the businesses that use Blanket are not choosing to compare it to a permissionless alternative. They are choosing to compare it to nothing. They have no hedge. That is the true competition. A grocery store that cannot hedge a banana price shock is not going to say “Polymarket offers a better odds display.” It is going to say “something is better than nothing.” Blanket is that something.
This is where the contrarian angle gets uncomfortable for me. I would love to see a fully decentralized version of Blanket, with an open risk model, auditable output, and a DAO that governs the recommendation logic. But that version would still have no distribution into the insurance broker channel. It would still struggle with the same last-mile problem. The regulated version has a head start because institutions already trust Kalshi. The legal wrapper is a feature, not a bug, when your customers are small businesses that do not want to be the first to test an unregulated experiment.
The real threat to crypto is not that Blanket will kill prediction markets. The real threat is that Blanket will set the standard for what users expect from risk analysis. If users learn to expect polished, regulated, explainable AI recommendations, they will not be satisfied with a pseudonymous interface that issues a token and calls it governance. They will demand accountability. The tool with the best accountability layer, not the best consensus layer, will win the next era.
Takeaway: A Risk Layer, Not a Settlement Layer
We are in a bear market. Survival matters more than gains. The protocols that survive will be the ones that can show actual usage without depending on token emissions. Blanket is not a protocol, but it is a perfect test case. It has no token, no inflation, no speculative premium. It is betting that the combination of Kalshi’s regulated liquidity and an AI-driven interface can convert event contracts into a real risk-management product for people who would never open a DeFi dashboard.
I don’t know if Blanket will succeed. There is no benchmark for its recommendations, no audit trail, no disclosed business model, and a dangerously thin public profile for its only developer. The compliance risk is real, especially around the CTA question. The distribution mountain is steep. But I do know that the crypto industry has been over-committed to infrastructure and under-committed to distribution. Blanket is a reminder that adoption is not about the biggest rollup or the strongest validator set. Adoption is about a tool that appears exactly when a small business feels afraid of next winter.
The questions I want us to sit with are not about Blanket’s token price. They are about ownership and consent. Who should control the recommendation model that tells a farmer how to hedge their crop? Who should be accountable when that recommendation is wrong? What happens when Kalshi’s event contracts are the only available hedge, and the business has no choice but to accept the platform’s terms? These are governance questions, and they will not be solved by a whitepaper. They will be solved by boring things like disclosure standards, review processes, and complaint channels.
So let me end with an invitation, not a summary. The next time a crypto project calls itself the foundation of a new risk market, ask it one question: what would Blanket do if it had to serve a real business on a rainy Tuesday afternoon? If the answer involves a token, you might be looking at a speculative game. If the answer involves a clear explanation, an honest limitation, and a legal path, you might be looking at the future. The future will not begin with a block. It will begin with a conversation between a machine and a frightened business owner. We didn’t need another L1 to get there. We needed someone to make the risk feel close, small, and answerable. Blanket is trying to be that someone. The rest of us are still arguing about consensus, while the sky is filling with storms.