Ask HN: What Changes When Adtech Moves from Keyword to Context-Based Targeting?

Adtech is seeing a major shift—from keyword-based targeting (search, bid, intent) to more context-driven approaches (content, behavior, semantics, attention signals). With increasing privacy regulations, deprecation of third-party cookies, and rise of LLMs, the ability to interpret context in real-time is becoming central to modern ad targeting.

I'm curious:

What are the key technological shifts when moving from keyword to context-based ad targeting?

How are platforms adapting—especially DSPs, SSPs, and publishers?

What impact will this have on existing ad revenue models (e.g., CPC, CPA, CPM)?

Who is actively working in this space right now—either startups, research labs, or open-source communities?

If you're working on this, experimenting with new approaches (e.g., semantic embeddings, on-device inference, or federated contextual targeting), or looking to collaborate—I'd love to connect and potentially partner.