Calibration against an IT8 target only tells you each scanner is reasonably accurate against that reference; it does not guarantee two different real-world samples will scan identically.

Based on the answers, the most likely causes are:

1. The paint samples may not truly match. Different paint batches, substrate color, and application thickness can all shift the scanned color.
2. Comparing hue alone is misleading. Hue differences are not very meaningful, especially for low-saturation colors. Delta E is the better metric for judging whether two colors are actually far apart.
3. One scan may have a white-balance/profile bias. From the target comparison, scanner #2 appears warmer in the greys, which suggests a white-balance issue. Its stronger saturation across multiple colors suggests there may also be an additional profiling or processing difference.
4. Different scanner gamut by itself does not explain a mismatch if both are being viewed and managed correctly.

Best next step: compare scans of the same authentic IT8 target and evaluate them with proper color management and Delta E, rather than relying on a painted sample or hue values alone.