It may not be a human
AI-operated wallets handled 8% of daily memecoin trading volume on Solana's decentralised exchanges during March.
By June, that number was 34%.
More than 120,000 agent-only wallets were active on the network, each executing an average of 11.4 transactions a day. Human wallets averaged 2.1.
On peak days during token launches, automated agents account for over 70% of all DEX volume on Solana.
The majority of trading activity on the network now comes from software, not people.
This is not limited to memecoins or to Solana. Kraken, Binance, OKX and Coinbase have each shipped native toolkits specifically for agent developers in the past year.
MetaMask launched a dedicated self-custodial wallet for AI agents in June.
The biggest exchanges and wallet providers are rebuilding the infrastructure of crypto trading around a single assumption: the next wave of volume will come from machines.
Perpetual futures, the most actively traded derivative in crypto, sit directly in the path of this shift.
Hyperliquid, the dominant venue for on-chain perps, processed $633B in volume during the first quarter of 2026 and now handles roughly 70% of all decentralised perpetual futures activity.
Open interest on the platform hit $11B at its mid-year peak.
This is the venue where Deploy.Finance's Super Perps agent operates: an autonomous agent that reviews the market, decides whether to go long, short or stay flat, and manages each trade from entry to exit with no human involvement.
Super Perps is the path towards the question that applies across crypto derivatives: what happens when the participants on both sides of a leveraged trade are increasingly machines?
What agents do differently
The advantage an autonomous trading agent has over a human trader is not speed.
The agent does not panic-sell during a drawdown or widen its stop because it "feels" confident about a position.
It does not take revenge trades after a loss. And it does not get tired at 3am. Every cycle starts clean: fresh data, no memory of what happened last time, no attachment to a prior view.
During backtest, the agent identified 2,298 potential trades and chose to do nothing on 1,350 of them.
A human trader with a screen open and money at risk almost never chooses to do nothing.
The risk framework reinforces this.
Position sizing comes from a fixed loss budget. The stop is set before entry and never widened. Winners are managed with a trailing stop that locks in profit as the trade moves.
These are not new ideas.
Institutional systematic funds have operated this way for decades. What is new is that these mechanics are running inside a retail self-custody wallet, on a decentralised exchange, without a fund manager in between.
The Wintermute OTC desk reported that institutional investors accounted for 72% of spot trading volume in the first half of 2026, the highest share on record. Realised volatility in crypto has dropped from roughly 70% in earlier cycles to about 45% in the current one.
The market is professionalising from the top down through institutions and from the bottom up through agents. The space left for a human retail trader making discretionary leveraged bets is getting smaller from both directions.
The hard question
If autonomous agents are structurally better at trading perpetual futures than most humans, what exactly is the retail trader doing in this market?
One answer is that products like Super Perps are a response to this reality.
If the median retail derivatives trader loses money (and exchange data consistently suggests they do), then delegating to a disciplined system with a fixed risk budget might be a better outcome than trading manually.
The backtest shows a 52.3% win rate with a 13.6% maximum drawdown over thirteen months.
The other answer is that delegation introduces its own risks.
The user cannot evaluate the agent's individual decisions in real time. They cannot override a trade they disagree with.
There is also a market-level question that gets less attention.
When a large enough share of perpetual futures volume comes from agents running similar signal sets (positioning data, liquidation zones, funding rates, order-book depth), the agents may begin trading against each other rather than against mispriced human flow.
At that point, the edge that comes from disciplined execution against emotional retail traders starts to compress.
What's next
Automated agents are becoming a larger share of crypto derivatives volume every quarter. The biggest names in the industry are building the infrastructure to support them.
Products like Super Perps are packaging autonomous execution for users who would otherwise be trading manually, and probably trading worse.
The question that nobody has a confident answer to yet is what the equilibrium looks like. A market where most participants are machines running on overlapping data will behave differently from a market where most participants are humans with uneven information and uneven discipline.
For now, the retail trader faces a practical choice: compete against autonomous systems that do not sleep, do not tilt, and size every position from a risk budget rather than a feeling, or hand the job to one of those systems and accept that you cannot fully see how it works.
Neither option is comfortable.
But pretending the market still belongs to people watching charts may be the least comfortable option of all.
This is why Deploy Finance built Super Perps, your own AI trading agent to take directional bets on decentralized perpetual markets.
The agent is now live on deploy.finance and open to >$10,000 deposits as part of the initial rollout.
Learn more about Super Perps in our previous read.