5 insights that caught my attention in just published BCG 2026 M&A Report:

​New AI alliances:

2022: 356

2023: 493

2024: 662

2025: 1,310

2025 number alone beats 2023 and 2024 combined.

Enterprises want AI access with room to adapt

For most AI capability decisions, BCG recommends partnering or building. Models and tools move too fast to justify owning them outright in many cases.

“use the lightest structure that provides the access or control required to gain the advantage while preserving the freedom to switch as the technology evolves…. Usually, the decision will lean toward partnering for or building custom solutions."

My take: speed, tailoring and freedom to switch are now becoming an expectation.​

Hyperscalers connect an otherwise fragmented AI stack

Most segments of the AI stack still develop separately. Hyperscalers are the exception.

That makes the big three the main route across the AI stack. And BCG warns that positions taken in the next two years will be hard to reverse.

Industrial partnership patterns point the same way.

"Operators typically combine broad cloud partnerships with a smaller number of specialist alliances, making only selective acquisitions."

Distribution command an extraordinary premium

OpenRouter, a platform routing AI models, went from a $1.3B valuation in May 2026 to an $7B+ acquisition agreement price a few months later. That's a 5x+ repricing up.

It's an exceptional case, but it shows what sitting between AI models and the workflows that use them is worth.

For me, that's the strongest argument showing how central distribution is to AI strategy.

Ecosystem position now has a price

Most AI acquisitions still buy a team or a technology. The only theme growing is buying a position in the ecosystem: from 3% of AI companies' deals in 2020 to 9% in 2026 so far. That's ~3x in share.

My take: ecosystem position has a price now. Alliance teams build it every day.​

Execution still decides whether an alliance delivers

BCG cited a study that found ~9 in 10 partnerships between industrial and technology companies missed their goals. Partner teams have lived with that problem for years.

The advice is to have clarity on what you want to achieve, data and IP ownership, secure portability and benchmarking rights, and a way to exit when value falls short. Expect those terms in your next AI partnership negotiation.​

For alliance and Marketplace leaders, I'd connect these signals:

• Leverage the cloud relationships that give you access to your target customers.

• Prioritize integrations that put your product inside a recurring customer workflow.

• Measure adoption and commercial value continuously

​​Source​