You Were Right About The Sector. You Still Lost.

Being right is the part everyone focuses on, and it is also the easy part. Say you decided two years ago that semiconductors were going to run. You were correct. The sector ran. Now answer the questions that actually decided whether you made money: which eight names did you buy, at what weights, and what did you do when one of them tripled and quietly became sixty percent of your position? Most people never get that far. They buy the one name they have heard of, size it by feeling, and never rebalance. The thesis was right. The execution turned it into something else entirely, and the gap between those two things is where retail money goes to die.

Look at what actually happens when someone tries to express a view. They open eight tabs. They buy three of the eight names because the other five felt expensive that day. They pay a spread on each one. Now they hold eight positions, eight cost bases, eight things to check, and a portfolio whose shape they did not choose and cannot easily see. Six months later one winner dominates and the position is no longer a bet on semiconductors at all. It is a bet on one company, held by accident. The professional version of this job has a name and a full time staff behind it. Index providers publish a rulebook defining what belongs in the basket, how much of each, and when to reset it. Fund managers execute against that rulebook with a trading desk. None of it is genius. It is process, applied without exception, at a cost per trade that an individual cannot match. This is why the familiar observation that most active stock pickers trail a broad index over long horizons is less about intelligence than it sounds. The index is not smarter. The index simply never forgets to rebalance, never gets attached to a winner, and never skips a name because it felt expensive on a Tuesday.

Put numbers on it. Take the semiconductor basket Prism runs today: eight names at target weights of Nvidia twenty percent, TSMC sixteen, Broadcom fourteen, AMD eleven, Qualcomm ten, Intel ten, ASML ten, and Micron nine. Now assemble that yourself. Eight buys, eight spreads, eight approvals, eight gas fees. Then do it again every time the weights drift, which they will, because a twenty percent weight that outperforms the basket by half becomes a twenty six percent weight without you touching anything. To hold the shape you actually intended you would need to trim the winner and top up the laggards, repeatedly, forever. Almost nobody does this, and not because they are lazy. They skip it because on any single occasion the work costs more than it visibly returns. The cost of skipping it compounds quietly instead, which is exactly the kind of cost human beings are worst at noticing.

Prism's answer is one number instead of eight. One token, pSEMI, whose price is the value of that entire basket divided by the tokens outstanding, currently sitting around two hundred thirty eight USDG. You buy the number. The eight positions underneath are held for you, at weights you can read on chain, reset when they drift outside a defined band. But a token that claims to represent a basket is easy to promise and hard to make true, and there are two ways the promise usually breaks. The first is a wrapper that simply asserts exposure: you send money, a contract issues you a token, and somewhere a manager claims to hold the underlying. That works right up until the moment you want out and discover the promise was the only thing backing it. A token whose value rests on trust is a credit instrument wearing an index costume. The second is a token that trades on an open market with no way to create or destroy it, so when demand rises the price floats above what the basket is worth and buyers pay a premium that benefits nobody, and when demand falls holders are trapped below fair value. Without a mechanism to arbitrage that gap, the price is a popularity contest.

Prism is built so neither can happen, and the mechanism that prevents both is the same one. The token is not a claim on a promise. It is a claim on a basket you can take delivery of at any moment. Anyone can create new pSEMI by delivering the underlying assets in the exact proportions the vault currently holds, and the contract mints the corresponding amount in return. Redemption is that operation in reverse: burn pSEMI and the contract returns your pro rata slice of every asset in the basket, not the cash value of it but the assets themselves. That single design choice defends the price without us lifting a finger. Suppose pSEMI trades at two hundred fifty while the basket it holds is worth two hundred thirty eight. Anyone can buy the eight underlying assets for two hundred thirty eight, mint one pSEMI, sell it for two hundred fifty, and keep the difference, and they will keep doing it until the premium is gone, because free money attracts a crowd. Run the same logic on a discount and redemption closes the gap from the other side. The peg is not maintained by a treasury or by a market maker we pay. It is maintained by everyone who wants to be paid for noticing.

Everything the vault holds is visible while that happens. Net asset value is computed from oracle prices on chain and readable by anyone, and so is the full composition: every asset, its current weight, its target weight, and how far it has drifted from where it should be. This matters more than it sounds. If you cannot verify what a basket holds, you do not own an index. You own a rumor. The whole point of buying a theme rather than a company is that you know what you are exposed to, and that knowledge has to survive contact with a system you did not build.

There is one problem specific to what Prism is doing, and the answer we chose is not clever but careful. These assets do not keep the same hours. Tokenized equities follow markets that close on Friday afternoon and stay shut through the weekend. Silver closes. Memecoins never sleep, not for a minute and not for a holiday. A basket that mixes all three has to survive being half awake. The wrong answer is to guess: use the last known price, extrapolate from a correlated asset, apply a model. Every one of those approaches lets someone mint or redeem against a number no market is currently willing to honor, and the loss lands on everyone else in the vault. So Prism refuses to quote. Every price feed carries a freshness requirement, and if any leg of a basket has no fresh defensible price, the vault names the asset that went stale and pauses minting and redemption for that basket entirely. Not degraded, not approximate. Paused. Existing holders keep their tokens and their exposure. Nobody transacts against a price nobody can defend. Refusing to quote is worse than quoting for exactly one thing, which is convenience, and better for the only thing that matters, which is not losing your money to someone with a faster clock.

The other guardrails follow the same philosophy of bounding the system rather than trusting it. Every asset carries a maximum weight it may never exceed, so no single name can quietly swallow a basket, and the Degen index carries the tightest caps of all, because a memecoin basket is a different animal and pretending otherwise is how people get hurt. Rebalancing runs only when weights drift outside a set band, because trading every wiggle burns more value than it captures. Any rebalance that would cost the vault more than a defined amount of net asset value reverts instead of completing. A cooldown prevents the vault from thrashing between two assets that keep trading places. None of these are promises in a document. They are conditions in a contract that reverts when they are not met, which is a meaningfully different kind of assurance.

What comes out the other side is four things. You get one position instead of eight: one thing to buy, one thing to sell, one line in your portfolio, one cost basis. You get the weights you actually intended, maintained without you remembering to maintain them, which over a long enough horizon is most of the distance between a good thesis and a good outcome. You get a token you can verify and exit, because the composition is public and redemption is open to anyone, so if you ever doubt what you hold you can simply take delivery of it. And you get the discipline of a rulebook without hiring anyone to run it, which until now was the part reserved for institutions.

The three indexes live today are deliberately a beginning. They were chosen to prove the machine against three different kinds of asset: equities that close, a metal that closes, and memecoins that never do. The product is not the three baskets. The product is the machinery that makes any basket possible, and that is where this goes next. We widen the shelf first, with AI infrastructure, energy, dividend payers, and the themes people keep trying to assemble by hand, each one a configuration rather than a rebuild because the vault does not care what it holds. Then we hand over the pen, so anyone can define an index, publish it, and earn a share of the fees when other people hold it, because there is no reason the list of available themes should be limited to the ones we thought of. Three baskets curated by a team is a product. Ten thousand baskets curated by a market is an asset class. Two of those steps are already done: there is now a route that mints straight from USDG, because delivering the basket is the correct primitive to build on but not what a first time buyer should have to do, and a swap underneath the mint turns eight approvals into one. Rebalancing runs on a keeper against a solver the vault computes itself, so it happens on schedule rather than when someone remembers. And all of it goes into a phone, because the person who most needs to stop babysitting eight tickers is not sitting at a desk. The ambition underneath every one of those steps is the same: having an investment idea should not require you to become an operations department to act on it. You should be able to buy the idea. One token, the whole theme. One light in, a full spectrum out. Prism Capital.