
A recent survey by VentureBeat reveals that enterprises are investing heavily in AI infrastructure without fully understanding the costs involved. While many organizations rely on hyperscalers and model-provider APIs, there is a growing interest in AI-specialized clouds, which few currently use. The survey found that only 21% of enterprises have AI running at scale, and most report low GPU utilization. This indicates a disconnect between investment and operational maturity, as companies plan to switch or add infrastructure providers within the next year.
Read original