diff --git a/config/gpu/h100.toml b/config/gpu/h100.toml
index ef0aab8d6..fd5e87a1f 100644
--- a/config/gpu/h100.toml
+++ b/config/gpu/h100.toml
@@ -1,6 +1,6 @@
 # H100 PCIe/SXM (80GB VRAM) -- full production
 [training]
-batch_size = 1024
+batch_size = 0  # 0 = auto-compute from VRAM via AutoBatchSizer
 num_atoms = 51
 buffer_size = 500_000
 hidden_dim_base = 256
diff --git a/crates/ml/src/trainers/dqn/trainer/constructor.rs b/crates/ml/src/trainers/dqn/trainer/constructor.rs
index 5692acf02..8ab5f4501 100644
--- a/crates/ml/src/trainers/dqn/trainer/constructor.rs
+++ b/crates/ml/src/trainers/dqn/trainer/constructor.rs
@@ -30,14 +30,6 @@ use super::DQNTrainer;
 
 impl DQNTrainer {
     pub(crate) fn new_internal(mut hyperparams: DQNHyperparameters, debug_logging: bool, override_device: Option<MlDevice>) -> Result<Self> {
-        // Validate batch size is non-zero
-        if hyperparams.batch_size == 0 {
-            return Err(anyhow::anyhow!(
-                "Batch size must be greater than 0, got: {}",
-                hyperparams.batch_size
-            ));
-        }
-
         // WAVE 26 P2.2: Validate gradient_accumulation_steps > 0
         if hyperparams.gradient_accumulation_steps == 0 {
             return Err(anyhow::anyhow!(
@@ -105,6 +97,13 @@ impl DQNTrainer {
             }
         };
 
+        // batch_size == 0 is the sentinel for auto-compute from VRAM.
+        // Let AutoBatchSizer drive the batch size, capped at 8192.
+        if hyperparams.batch_size == 0 {
+            hyperparams.batch_size = max_safe_batch.min(STATIC_MAX_BATCH_SIZE);
+            info!("AutoBatchSizer: batch_size auto-computed to {}", hyperparams.batch_size);
+        }
+
         // Cap to VRAM ceiling from AutoBatchSizer (no separate scale-UP —
         // AutoBatchSizer already accounts for model size and available VRAM)
         if hyperparams.batch_size > max_safe_batch {