diff --git a/crates/ml/src/cuda_pipeline/experience_kernels.cu b/crates/ml/src/cuda_pipeline/experience_kernels.cu
index 52eb55c00..baa78facb 100644
--- a/crates/ml/src/cuda_pipeline/experience_kernels.cu
+++ b/crates/ml/src/cuda_pipeline/experience_kernels.cu
@@ -473,7 +473,13 @@ extern "C" __global__ void experience_env_step(
     float tx_cost = 0.0f;
 
     if (delta != 0.0f) {
-        tx_cost   = fabsf(delta) * raw_close * tx_cost_multiplier * spread_scale;
+        /* Market impact: larger trades cost more (quadratic in size).
+         * A 1-lot order fills at bid/ask. A 4-lot order moves the market.
+         * impact_scale = 1 + |delta/max_position|^2 — ranges from 1.0 to 2.0.
+         * This teaches the model that scaling to max position is expensive. */
+        float size_ratio = fabsf(delta) / (max_position > 0.0f ? max_position : 1.0f);
+        float impact_scale = 1.0f + size_ratio * size_ratio;
+        tx_cost   = fabsf(delta) * raw_close * tx_cost_multiplier * spread_scale * impact_scale;
         cash     -= delta * raw_close;
         cash     -= tx_cost;
         position  = target_position;