Microsoft is reducing the number of neocloud partners it relies on for AI capacity, focusing on a smaller set of suppliers and its own data centres as the industry buildout shows strain, CNBC reported. The shift suggests hyperscalers are becoming more selective after a scramble for GPU hosting. Neocloud specialists had benefited from overflow demand when Microsoft and peers needed extra training and inference capacity. Consolidation could pressure smaller hosts while large remaining partners gain volume. Executives across the sector have also been reacting to a post-midterm sell-off in AI infrastructure shares.
Source: CNBC