Some crypto trading bots can be profitable over particular periods, while others lose money. Automation alone creates no guaranteed edge. The outcome depends on the strategy, the market environment, the risk taken and the full cost of trading.
Automation is not the source of profit
A bot can execute an ineffective rule more consistently than a human. That improves consistency, but not necessarily the result. Similarly, an attractive historical chart can reflect favourable conditions, selective reporting or a backtest that understates trading costs.
A simple illustration shows why costs matter: if a hypothetical trade earns 0.10% before costs but incurs 0.14% in total costs, its net result is negative. Repeating it more frequently repeats the loss.
What evidence is useful?
Results should distinguish live trading from simulation and should show the observation period, losses and costs—not only winning trades. Strategy changes also matter: a track record from an earlier design may not describe the current version.
AlphaNet’s AI-agent research emphasizes that sizing, exits and cost control are separate from a model’s ability to produce a convincing market explanation.
A profitable period is evidence about that period, not a guarantee of future income. Neither an “AI” label nor a high historical win rate removes the possibility of substantial loss.