Nvidia-backed Nebius to acquire AI agent search firm Tavily

Nebius, a NASDAQ-listed AI cloud company based in Amsterdam, announced an agreement to acquire Tavily, a leading agentic search provider serving Fortune 500 enterprises and top AI companies.

The deal aims to integrate Tavily’s real-time search infrastructure into Nebius’s AI platform, supporting the growth of enterprise autonomous AI systems.

Tavily’s technology will complement Nebius’s Token Factory by providing real-time web access for improved factual accuracy.

This acquisition advances Nebius’s strategy to build a unified platform for building, tuning, and deploying autonomous AI agents.

Tavily’s CEO Rotem Weiss and team will continue developing the product under its current brand, leveraging Nebius’s global infrastructure.

The transaction is expected to close in the next few weeks, subject to customary closing conditions.

🔗 Source: Nebius

🧠 Food for thought

Implications, context, and why it matters.

This acquisition goes beyond search technology

Nebius runs as an AI infrastructure company, with multi-billion-dollar agreements with hyperscalers like Microsoft and Meta 1.
It lifted its 2025 capital expenditure outlook to about $5 billion to grow its global data center footprint 1.
Tavily fills a gap with a “web access layer for AI agents.” These tools let AI systems pull current information from the public internet. Tavily says 1 million-plus developers rely on it 2, 3.
With Tavily built in, Nebius can offer a fuller stack for autonomous AI systems, beyond raw GPU capacity.

The AI cloud fight is moving toward integrated platforms

AI cloud competition now centers on packaged agent platforms, alongside access to the latest NVIDIA GPUs.
Amazon Web Services offers broad infrastructure plus Bedrock (a managed service for building and deploying generative AI applications) 1. Nebius is taking a narrower approach with an all-in-one platform for agentic AI.
Owning Tavily’s agentic search layer lets Nebius deliver tighter integration and potentially higher performance than a setup built from separate pieces on a general-purpose cloud.
The deal raises pressure on cloud providers to ship similar agent-building environments. It may also speed up acquisitions of agent tooling and data-grounding companies (providers that help AI systems stay accurate by connecting responses to up-to-date sources).