Severe Performance Stall during GLB Extraction with Low CPU & GPU Utilization

I am to report a significant performance bottleneck I've encountered during the "Extract GLB" step on a high-end desktop system. While the initial 3D asset generation is incredibly fast and properly utilizes the GPU, the subsequent GLB extraction process is extremely slow and appears to be underutilizing the available hardware.
System Configuration
Here is a summary of my hardware configuration:
CPU: AMD Ryzen 9 9950X3D (16-Core / 32-Thread)
Memory (RAM): 190 GB DDR5
GPU: NVIDIA RTX 5090 D (32 GB VRAM)
OS: Windows 11, running the application via WSL2
Observed Behavior
Low CPU Utilization: Total CPU usage remains very low, typically under 10%. Per-core analysis in the Task Manager shows no sustained single-core or multi-core load.
Low GPU Utilization: The GPU utilization is also extremely low, often near 0-3%, indicating it is not actively engaged in any heavy computation.
No Disk/Memory Bottleneck: Both RAM and disk I/O show no significant activity.

Start by reproducing the “Extract GLB” step in the reported Windows 11 and WSL2 environment, comparing it with the fast 3D asset generation step. Since the issue names no files, tests, or entry points, trace the extraction path and profile where time is spent. Done means identifying the bottleneck and documenting or verifying a fix for the severe stall.