ai16z Trades in 2026: The Ghost Token and the $SLOP Auction

A venture fund that promised an artificial intelligence would pick its trades is, by every official measure, dead. Its founder has disavowed the token. Its treasury was handed over to settle a class action. Its market value has fallen roughly 99.9% from a peak near 2.4 billion dollars. And yet, on any given day in November 2026, somebody is still trading ai16z.

On 3 November 2026 the migrated token, now called ELIZAOS, changed hands to the tune of about 1.2 million dollars in 24 hours at a market capitalization of roughly 1.98 million dollars, according to CoinGecko. A corpse with a live tape. But the trade that actually matters this week is not the ghost token at all. It is the $SLOP auction, a brand-new tokenized funding rail built by ElizaOS creator Shaw Walters, which opens for pre-bidding on 5 November and closes on 10 November, per TokenPost. This is the story of where the real ai16z trade went: up a layer, out of a disavowed fund’s token and into a fresh token with a 400,000 dollar target and a payment contract that has not yet been independently audited.

Making sense of that means pulling apart three different things the phrase ai16z trades has meant over two years: the trades the fund claimed its AI would make, the trades its ghost token still makes today, and the trade its founder is asking the market to make this week.

What ai16z trades was supposed to mean

ai16z launched on the Solana platform daos.fun on 24 October 2024. The pitch was simple and, for a few weeks, intoxicating: an autonomous AI portfolio manager, cheekily styled after the venture firm Andreessen Horowitz, would trade a crypto treasury on behalf of the people who bought in. The gauntlet moment came when the real Marc Andreessen gave the project a social-media shoutout, and the fund rocketed toward a 96.6 million dollar market cap within hours, as The Block reported at the time. By late December 2024 the fund token was worth more than 22 million dollars, and on 2 January 2025 it peaked around 2.39 billion dollars with roughly 291 million dollars in daily volume, figures later recapped by CoinDesk.

It is easy to forget how genuinely novel the idea felt in late 2024. The persona at the center, an AI styled as Marc AIndreessen, posted investment theses, bantered with traders, and appeared to run a book in public. For a market hungry for the next story after the first wave of AI tokens, a fund you could literally watch think was irresistible. The reflexive loop was immediate: attention lifted the price, the rising price drew more attention, and the attention was taken as proof of the thesis. None of that required the underlying trading to be autonomous, or even good.

The daos.fun structure gave the word trades a specific meaning. Each fund runs a one-week raise, after which the designated trader controls the treasury while the fund token trades freely on an automated market maker. At expiry, holders can burn their tokens for a pro-rata share of whatever the treasury holds, plus any SOL, with proceeds split between the fund creator and investors. So ai16z trades, in the original sense, meant the on-chain positions the treasury took: the buys and sells an AI was supposedly choosing. That framing is exactly where the trouble started.

The autonomy gap, or who was actually pressing the button

Within days of the launch, the crypto outlet Protos poured cold water on the central claim. No fund on daos.fun, it reported, was actually operated by an AI; the trades attributed to the ai16z agent were in practice made by, in its memorable phrasing, “a non-AI human named Shaw.” The agent could post, could talk, could generate the appearance of a thesis, but a person still signed the transactions. Protos called it AI-washing, and the label stuck.

This gap between pitch and practice is the hinge the whole story turns on. The market did not price a chatbot with a social account; it priced an autonomous capital allocator that did not, in any strict sense, exist. The problem of proving that an AI actually did the thing it was paid to do remains unsolved at the infrastructure level, which is why a cottage industry of verifiable-compute projects has grown up trying to answer it, a subject we cover in our look at zk-ML and proving the AI you paid for actually ran. ai16z never had that proof. It had a narrative, and for a while the narrative was enough.

The pattern was not unique to ai16z, which is part of why it mattered. The autumn of 2024 produced a whole crop of so-called agent funds and personality tokens that borrowed the same trick: wrap a human-run operation in the language of autonomous intelligence, let the branding do the work, and let the price chart supply the credibility. ai16z was simply the biggest and loudest example, which made it the one that drew both the capital and, eventually, the lawyers. The reckoning that followed was less about one project lying than about a market that wanted to believe faster than it wanted to verify.

From a meme to 2.4 billion dollars to a settlement

The arc from here is a near-perfect parabola. On 28 January 2025 the project rebranded from ai16z to ElizaOS, repositioning the token as the native asset of an open-source agent framework rather than a VC fund. On 6 November 2025 it executed a token migration, a 1-to-10 redenomination of supply that let holders swap their old AI16Z for new ELIZAOS at a 1-to-6 ratio. That migration later became the core of a legal complaint.

In April 2026, Burwick Law filed a class action in the Southern District of New York. The suit alleged that the project traded on confusion with the Andreessen Horowitz name, overstated the autonomy of its trading, and diluted existing holders through the migration, which distributed roughly 60% of new supply to holders and 40% to insiders. The case went before Judge Jed Rakoff that May. Rather than fight it to the end, the team settled, transferring the foundation’s remaining treasury to a group of holders. On 5 August 2026, Walters told CoinDesk the token was “dead, completely,” with no further foundation support, buybacks or supply measures, and vowed he was “never letting a token come close to Eliza again.”

The daos.fun clock that never rang

There is a quieter detail inside the collapse that explains a lot about why the token lingers. Every daos.fun fund is born with a clock. ai16z was created on 24 October 2024 with a one-year life, which meant the treasury was due to wind down around 24 October 2025, at which point holders would burn their tokens for a pro-rata share of whatever remained. That is the mechanism that is supposed to return capital and close the book.

The clock never rang cleanly. The January 2025 rebrand and the November 2025 migration converted a time-boxed fund into an open-ended token, short-circuiting the orderly expiry the daos.fun design promised. Instead of a clean redemption, holders got a new ticker, a diluted supply and, eventually, a lawsuit. The structure meant to protect investors by forcing a wind-down was simply overwritten, and the token drifted on with no terminal date at all. A fund that was supposed to end became a token that cannot quite die.

Two ghost tickers, one dead fund

Here is the first trap for anyone who types ai16z into an exchange search box in 2026: there are two tokens, and they are not the same. The migrated token, ELIZAOS, is the one speculators still chase. As of 3 November 2026 it traded at about 0.0002653 dollars for a market cap near 1.98 million dollars, ranked around 2,398th, with a 24-hour volume just over 1.2 million dollars, a circulating supply of 7.482 billion against a max supply of 11 billion, and a price some 97.5% below its all-time high of 0.01078 dollars, per CoinGecko. The legacy token, still listed under the old AI16Z ticker, is a near-fossil: about 0.0003395 dollars, a market cap near 373,000 dollars, ranked around 3,965th, with barely 8,600 dollars of daily volume and a price essentially 100% below its 2.47 dollar peak, according to CoinGecko.

Add the two together and the entire on-chain estate of what was once a 2.4 billion dollar phenomenon is worth roughly 2.36 million dollars, a decline of about 99.9%. The lesson in the split is behavioral: speculation picks the migrated ticker and ignores the fossil, which means a trader who buys the wrong one is holding an asset with almost no liquidity and no narrative. Always verify the contract address, not just the three or four letters on the screen.

Reading the tape on a token nobody will defend

A disavowed token trades differently from a normal one, and the difference is worth spelling out because it governs the risk. There is no foundation left to buy back supply, no market maker of last resort, and, by the founder’s own statement, no buyback or supply management of any kind. What remains is a thin float that can be shoved around by a single motivated buyer or seller. The token therefore moves on three things: narrative, sector beta, and reflexivity, in roughly that order.

The narrative moves are the sharpest. ELIZAOS has a documented habit of spiking violently on any flicker of attention, then round-tripping the entire move within days. A speculative rally in October 2026 that briefly lifted the token well into triple-digit percentage gains had, by early November, fully unwound; the 30-day change sat around negative 67% on 3 November, per CoinGecko. These are not trends; they are impulses against an empty order book. Thin books are also where searchers hunt, since every large market order on a shallow pool is a sandwich waiting to happen, a dynamic we unpack in our piece on MEV strategies in 2026.

The practical consequence is exit asymmetry. Entering a position in a thin token is easy because your buy is the thing lifting the price; exiting is hard because your sell is the thing crushing it. The round-trip cost, spread plus slippage plus priority fees, can quietly eat a double-digit percentage of a position before any price move at all. And because reported volume on micro-cap tokens is frequently inflated by wash activity, the only honest way to size a trade is to assume the real, two-sided liquidity is a fraction of the headline number.

None of this is unique to ELIZAOS, but the token is an unusually pure specimen because it has been stripped of every support a normal asset enjoys. There is no team talking up a roadmap, no treasury defending a floor, no fresh narrative except the occasional reflected glow of the wider AI-agent trade. What is left is close to a laboratory sample of how a token behaves when the only thing moving it is other traders’ expectations about other traders.

Where the ghost token actually changes hands

If you decide to trade ELIZAOS, the next question is where, and the answer shapes the cost. The token lives natively on Solana, where the bulk of genuine liquidity sits in automated market maker pools rather than on a deep central-limit order book. A handful of centralized exchanges also list it, but listing is not the same as liquidity; a venue can show a price while offering almost no depth behind it.

On a concentrated-liquidity AMM, the practical reality is that a few hundred dollars can move the quoted price of a sub-2-million-dollar token by a noticeable margin, and that the pool’s depth is wildly asymmetric around the current price. Traders chasing a green candle tend to arrive together, which thins the book on the buy side exactly when everyone wants in, then evaporates on the sell side exactly when everyone wants out. That is the structural reason the October spike round-tripped so completely: there was never enough two-sided depth to hold a higher price once the narrative faded. For a token like this, the venue analysis is really a liquidity analysis, and the liquidity is thin everywhere.

The funding layer moved up a level

Kill the token and a hard question remains: who pays the developers of an open-source framework that no longer has a treasury? The answer taking shape through the autumn of 2026 is Slop Cash, a non-custodial funding rail that Walters built with Drew Pierson, a veteran of Arweave, Orderly Network and Ava Labs. Its design is deliberately narrow. Slop Cash pays contributors, whether human or AI agent, only for work that gets merged, not work that gets submitted. It scores accepted pull requests and code reviews deterministically and in public, freezes each month’s results into a proposal with a 14-day review window, and then pays out in USDC from wallets it does not itself control.

The numbers behind it are modest but real. Across its monthly cycles since July 2026, Slop Cash has scored thousands of merged pull requests and code reviews; its September payout came to 6,005 USDC across 72 transfers, a figure reconfirmed by ChainCatcher. The crucial nuance, and the one that separates honest reporting from hype, is this: the Eliza framework itself stays tokenless, but Slop Cash does not. It launches its own token, $SLOP, through a platform called Umia. The tokens did not disappear when Walters declared the old one dead; they moved up a layer, from the product to the rail that funds and governs it.

The problem Slop Cash is trying to solve is older than crypto. Open-source maintainers have always carried critical infrastructure on volunteer time, and the field has tried many fixes: corporate sponsorships, quadratic funding, retroactive public-goods rounds, bug bounties. What Slop Cash adds is a tight, legible loop between merged work and payment, with the scoring done in public and the money moving only after a review window. Whether that loop is durable depends entirely on the thing funding it, which is where the token, and its auction, come back into the picture.

Umia’s chief executive, Francesco Mosterts, a former Point72 portfolio manager, frames the whole exercise as a fix for the incentive problem that killed the last token. As he told The Block, “a liquid token can push teams toward short-term decisions aimed at increasing its price.” Umia itself raised about 6.1 million dollars at an 18 million dollar fully diluted valuation in an August-to-September 2026 auction, the same reporting noted. The irony is not subtle: the man who called his token completely dead is midwifing a new one, structured so the token governs a treasury rather than standing in for the code.

Inside the $SLOP auction, 5 to 10 November

Slop Cash is Umia’s first external project, and its auction is the live event of the week. Deployment is on Robinhood Chain. Pre-bidding opens at noon United States Eastern time on 5 November (17:00 UTC), formal bidding begins at the same hour on 7 November, and the auction closes at noon Eastern on 10 November, according to TokenPost. Bids are denominated in USDG, a stablecoin, and the stated target is 400,000 dollars; at the time of announcement, the committed amount was zero. If the target is not met, the structure provides for a refund.

The supply is fixed at 100 million $SLOP, with 56 million circulating at the token generation event and 45 million, or 45% of supply, sold in the auction itself. The remaining allocations, detailed by ChainCatcher, spread across ecosystem and contributor programs, the team, performance reserves, protocol-owned liquidity, a treasury, contributors and the initiator. Governance over the treasury, above a fixed monthly allowance, runs through decision markets rather than a simple multisig vote. The table below lays out the distribution.

Two choices in that structure are worth flagging. Denominating bids in USDG rather than a volatile asset means bidders commit a fixed dollar value, which makes the 400,000 dollar target a clean threshold rather than a moving one. And deploying on Robinhood Chain, rather than a general-purpose layer-1, ties the launch to a consumer-finance brand’s infrastructure, a notable choice for a project whose entire premise is funding open-source work in the open. Neither is a red flag on its own; both are signals about the audience Slop Cash is courting.

The audit gap in the new trade

For a system whose entire pitch is trustless, deterministic payouts, one disclosure deserves far more attention than it has received. According to ChainCatcher, Umia’s own contracts have been audited by the formal-verification firm Certora, but the Slop Cash payment contract itself had not yet been independently audited heading into the launch. In other words, the component that actually moves contributors’ money is, at the moment of the token sale, the least verified piece of the stack.

That is not a prediction of failure; it is a statement of where the risk sits. Smart-contract history is littered with projects whose peripheral logic was solid and whose money-moving core was not. Anyone weighing a bid should treat the unaudited payment contract as the headline risk, not a footnote, and should read the auction’s terms with the same skepticism they would bring to any new token sale. The broader compliance questions around who is liable when an autonomous funding rail misfires are only beginning to be mapped, a theme we track in our coverage of DeFi compliance in 2026.

What Walters is actually building now

The most telling signal about where ai16z trades is headed is what its founder spends his time on, and in early November 2026 that is emphatically not a token. A Principia Research analysis cited by ChainCatcher described Walters and the elizaOS community reverse-engineering Hyperliquid’s closed node software on GitHub: a project six days old, already drawing 46 contributors, though none of its 96 verification checkpoints had yet been passed. The framing was pointedly about decentralization, noting that five foundation nodes controlled some 47.4% of the staking share and the top three validators together held around 36%.

Whatever becomes of the effort, the shape of it is the point: ship open-source code that scrutinizes a closed system, with no token bolted on. It is the anti-ai16z. The contrast with Hyperliquid’s own model, where a live exchange funnels real fees into buybacks, is instructive, and we examine that design in depth in our analysis of Hyperliquid and the question of real yield versus a billion-dollar buyback. Walters’ public posture has hardened into a line he first offered at Token2049 in Singapore and repeated to Decrypt: “you probably do not want to give an AI agent a bunch of money and expect it to make you more.” The elizaOS repository, meanwhile, keeps shipping, with a large open-source following and a permissive MIT license, a living codebase orphaned from a dead ticker.

The reverse-engineering project also says something about the maturity of the critique. In 2024, the ai16z pitch asked the market to trust an opaque claim of autonomy. In late 2026, the same founder is pointing a crowd of contributors at someone else’s opaque system and asking them to prove, checkpoint by checkpoint, how centralized it really is. The posture has flipped from make a claim to verify a claim, which is, in miniature, the arc the whole sector has been forced through.

The sector it trades inside

ELIZAOS does not trade in a vacuum; it is a micro-cap inside the AI-agent category, and the category itself has cooled from its early-2026 froth. On 3 November 2026 the CoinGecko AI Agents sector showed a combined market cap of about 3.06 billion dollars on roughly 164.5 million dollars of daily volume, led by Venice Token, Virtuals Protocol, the Artificial Superintelligence Alliance, Kite and OriginTrail, per CoinGecko. Against those names, a sub-2-million-dollar token is a rounding error in its own sector, which is why ELIZAOS moves far more on its own narrative than on sector beta.

The gap between ELIZAOS and the sector leaders is also a gap in kind, not just size. Venice, Virtuals and the Artificial Superintelligence Alliance all have running products, revenue of some sort, and teams actively defending their valuations. ELIZAOS has a respected codebase and a founder who publicly wants nothing to do with the token. When the AI-agent category rallies, capital flows to the names with a story to tell and a team to tell it; a disavowed ticker is the last thing bid and the first thing sold. That is why the token can post a triple-digit monthly move and still end the quarter deep in the red: the sector tide lifts it briefly, then leaves it stranded.

The regulator that never showed up

One striking feature of the entire ai16z saga is who did not intervene. A token rose past 2 billion dollars on a claim of autonomous AI trading, fell more than 99%, and the reckoning came not from the Securities and Exchange Commission but from a private class action. Under the current leadership of SEC Chair Paul Atkins, the agency has signaled a lighter touch toward on-chain markets, with Atkins framing the regulator’s job, as quoted by CoinDesk, as being “to set the rules of play and referee the game, not to pick the winning team.”

That posture leaves the liability question squarely on humans. An AI agent has no legal personhood, so when an autonomous trader loses money or an automated payout misfires, responsibility traces back to a deployer, a developer or a user, never to the software. The $SLOP auction will test the edges of this again: a stablecoin-denominated token sale on a corporate-adjacent chain, with a refund mechanism and a governance layer built on prediction markets. Whether any of that attracts scrutiny is an open question, and it sits inside a broader enforcement environment we map in our review of SEC crypto enforcement in 2026.

For holders of the ghost token, the regulatory vacuum cuts both ways. No enforcement action means no formal finding that anyone was wronged beyond the private settlement, but it also means no process that might have recovered value or clarified what the token actually is. The asset trades in a gray zone: too small to interest a regulator, too tainted to attract a serious bid, too liquid to disappear entirely. That is a strange, stable equilibrium, and it may persist for a long time.

How to think about trading, or funding, any of this

Strip away the drama and two distinct decisions remain. The first is whether to trade the ghost token. If you do, treat it as what it is: a thin, disavowed asset with no backstop, no buyer of last resort, and a demonstrated tendency to round-trip violent spikes. Size positions for total loss, verify the ELIZAOS contract address rather than trusting the ticker, and assume the exit will be worse than the entry.

The second decision is whether to participate in the $SLOP auction. Here the checklist is different: read the distribution and note that 45% is sold while more than half the supply sits with the team, ecosystem, reserves and treasury; weigh the unaudited payment contract; understand that bids clear in USDG with a refund if the 400,000 dollar target is missed; and remember that Walters owns the Eliza intellectual property and has sworn off ever letting a token near the framework again, which means $SLOP governs a funding rail, not the code itself.

The through-line across both is the same lesson the sector keeps teaching. The ai16z trade was never really a bet on an AI’s stock-picking; it was a bet on a story, and the price followed the story up and then all the way back down. The new trade, $SLOP, is a cleaner structure with a candid founder and a real funding problem to solve, but it is still, at bottom, a bet on a narrative. Price the narrative accordingly, and never confuse the quality of the code with the price of the token; on this project, above all others, the two have almost nothing to do with each other.

Frequently Asked Questions

Is ai16z the same as ELIZAOS?

They are two stages of the same project. ai16z began as an AI fund on daos.fun in October 2024, rebranded to ElizaOS in January 2025, and migrated its token in November 2025, letting holders swap old AI16Z for new ELIZAOS at a 1-to-6 ratio. Both tickers still trade, but ELIZAOS is the migrated, actively traded token while legacy AI16Z is a near-dormant fossil worth a few hundred thousand dollars.

Why does a dead token still trade?

Because declaring a token dead is a statement about official support, not a delisting. The founder ended foundation backing, buybacks and supply measures, but the ELIZAOS contract still exists on-chain and remains listed on exchanges, so anyone can still buy and sell it. With a thin float and no market maker, it moves sharply on narrative and tends to round-trip those moves within days.

What is the $SLOP auction and when is it?

$SLOP is the token for Slop Cash, an open-source funding rail from ElizaOS creator Shaw Walters, launching through the Umia platform on Robinhood Chain. Pre-bidding opens 5 November 2026 at 17:00 UTC, formal bidding starts 7 November, and the auction closes 10 November, with bids in the USDG stablecoin and a 400,000 dollar target.

Did an AI actually trade the ai16z fund?

Not in the way the marketing implied. Reporting from Protos in October 2024 found that the trades attributed to the ai16z agent were in practice made by a human, with the AI handling communication rather than autonomous allocation. The gap between the autonomous-AI pitch and the human-operated reality became a central allegation in the later class action.

Is $SLOP safe to buy?

This is not investment advice, and the risks are specific. The Slop Cash payment contract had not been independently audited heading into launch, even though Umia’s contracts were audited by Certora; the token sale accounts for 45% of supply while the rest sits with insiders and reserves; and the project is brand new. Read the distribution, understand the USDG refund mechanism, and size any participation for the possibility of total loss.

By Marcus Okafor, senior editor at HOGE Wire, covering artificial intelligence and crypto markets.