Will AI investing move beyond the hype, towards substance, in 2025?

For two years now, venture capitalists have been at a critical cross-road – invest in new, rapidly evolving AI technologies and run the risk of backing something that doesn’t work out in the long term, or wait and see, and run the risk of missing some outsized investment returns.  
 
While most investment in 2023-24 has gone into the foundational layer, in 2025, which segment(s) – model makers (foundation layer), application layer, operation layer, computing layer – are likely to see more allocations, which segments expect to see a slowdown in allocations. And as AI infra costs hold back profits, what might change in the roadmap-to-value for each of the segment(s) to confidently generate a targeted ROI?