AlphaNet Brings a Complete AI Trading Process to Hyperliquid

The platform combines strategy signals, position management and execution, letting users choose a trading approach without building the system behind it.

AlphaNet, a platform for automated quantitative trading strategies, went live on Hyperliquid in September 2026. Users can now allocate to more than 50 strategies listed on its site for the perpetual futures exchange, rather than building and running trading logic themselves.

Having a view on where Bitcoin is headed is one thing. Turning that view into a trading process means deciding when to enter, how much exposure to take and how to execute as conditions change. Those decisions continue long after the first order.

For individual traders, the commitment can be substantial. A strategy needs attention when volatility rises, liquidity thins or a position moves against it. Automating order placement helps, but someone still has to develop the logic governing those decisions.

AlphaNet is built around taking on that work. Its users select a quantitative strategy and an allocation, while the platform connects model signals with position management and execution.

The attraction is access to a developed trading process. Users decide which strategy to run and how much capital to commit, without having to research every signal, configure every entry rule or operate the underlying software themselves.

From configuring a bot to choosing a strategy

Trading automation can leave very different amounts of work with the user.

A configurable grid bot, for example, executes orders around parameters the trader chooses. Copy trading delegates the initial trading decisions to another trader, whose activity the service follows. AlphaNet’s Autopilot mode instead lets users allocate to a developed AI quantitative strategy.

According to the platform’s user guide, an Autopilot deployment begins with an asset, a strategy and a USDC allocation. The strategy submits orders in response to model signals, while the user can monitor activity, adjust the allocation or stop it. AlphaNet also offers a Standard mode for users who want to specify their own orders.

That division of responsibility gives the user a different job: choosing trading behaviour that fits their objectives.

AlphaNet’s Hackworth family makes these differences explicit. Trend favours long exposure while retaining some short-side activity. Prime combines long and short signals. OptimaShort places greater emphasis on short-side and mean-reversion opportunities. These profiles give readers a starting point for understanding why strategies trading the same asset may behave differently.

Between late August and late September, ZEC was one of the best-performing crypto assets, and the choice of strategy on it mattered almost as much as the choice of asset. From 23 August to 22 September, Hackworth Trend ZEC returned 92%, while Hackworth OptimaShort ZEC returned 59% over the same days. A user with a bullish view who picked the trend-following variant captured considerably more of the move. The same pattern held on HYPE, where Trend returned 48% and OptimaShort 27%.

Why the whole trading process matters

An attractive prediction is only the beginning of a trade.

Suppose a model anticipates a price rise, but volatility increases before the position is opened. The direction of the forecast may remain unchanged while the appropriate position size becomes smaller. If liquidity deteriorates as well, the system faces another decision about how quickly to execute.

AlphaNet’s Hackworth V3 documentation describes this connection between predictions and trading actions. Position sizing responds to signal strength, volatility and the available risk budget. A separate trading layer uses estimates of market conditions to determine how model outputs are acted on.

The practical point is that the trading system has more to decide than whether to buy or sell. It must determine how much to trade and how to carry out that decision under the conditions available.

Execution can materially change the outcome. AlphaNet’s Hackworth V3 documentation reports that, in backtests, its execution engine cut average slippage from 4.2 to 2.8 basis points. On a $1 million trading volume, that 1.4-basis-point difference is $140. Repeated across a strategy’s activity, small differences can consume part of its predictive advantage.

AlphaNet’s published research on dynamic time-weighted average price, or TWAP, addresses this problem. It examines how to distribute orders over time, comparing traditional optimisation methods with reinforcement learning approaches that respond to changing liquidity.

Trading immediately can consume the available liquidity at favourable prices. Trading more slowly can reduce that pressure, but exposes the remaining order to a moving market. Integrating execution into the strategy allows that trade-off to be considered alongside the signal that generated the order.

This helps explain AlphaNet’s product approach: the user allocates to a process encompassing several connected decisions, rather than assembling separate tools for each stage.

Making an automated strategy visible to its user

Delegating individual trades still leaves a need to understand what the system is doing.

AlphaNet’s documented marketplace includes Metrics, Analytics and Trades views. Users can examine a strategy before deployment, then review allocations, positions, pending orders and completed fills through the Portfolio interface.

Those distinctions help turn a return chart into something a user can assess. A strategy’s intended exposure explains what it is trying to do; its history shows how that approach behaved over a particular period. Individual trades provide another level of detail.

AlphaNet publishes strategy performance on its website in real time. Its public descriptions also encourage comparisons across directional exposure, drawdown and market conditions.

What the user continues to decide

Automation changes the work involved in trading; capital allocation remains a consequential choice.

Users decide which strategy fits their objectives, how much capital to commit and whether its behaviour continues to justify the allocation. For traders interested in systematic methods but unwilling to build and maintain the infrastructure themselves, AlphaNet offers a concrete route from strategy selection to automated execution on Hyperliquid. Its proposition rests on bringing research, position management and order execution into one process that users can choose and monitor.

The resulting decision is larger than selecting a bot setting. It is choosing a trading method, understanding the role it serves and deciding how much capital to put behind it.