As we wind down 2025, I’m doing what just about everyone else is doing—thinking about 2026.
For the private markets, this means thinking about more AI, all the time. That said, I do think next year the rubber is going to meet the road for AI startups and giants alike. High compute costs, compressed margins, and soaring valuations and expectations will inevitably collide with reality. And for some, this will mean even more acquisitions and more acquihires than perhaps we’ve seen so far in the AI boom.
I started asking around: Which startups would make smart acquisition targets for a tech giant in 2026?
“To unlock ‘real world’ AI like robotics, autonomous vehicles, smart factories, spatial computing, and embodied AI, tech giants need models that can reason about the real world in real time,” said Aidan Madigan-Curtis, Eclipse Ventures partner, via text. “Startups like Wayve, Physical Intelligence, WorldLabs, Bedrock Robotics, The Bot Company and GenesisAI, are already building simulation engines, sensor fusion stacks, and world models that learn from physical interaction—capabilities that would take incumbents years to replicate internally.” (Eclipse is an investor in Wayve.)
Madigan-Curtis gets at an essential question: In AI, when does it make more sense to acquire rather than build? Shensi Ding, CEO and cofounder at AI integration infrastructure startup Merge, points out an unconventional idea around finance (a widely touted AI use case): “Large AI players should acquire boutique investment banks and use historical financial models to train them. This work is highly specialized and requires domain expertise to really break through and build trust.”
