If big-data applications rely on robust error handling on narrow floats the most, then why the languages they use disregard exceptions?

error-handling artificial-intelligence floating-point big-data gpu