Cisco acquired Galileo: what it means for customers

What Happened, in Order

Cisco's announcement framed the deal around trust in AI systems, describing Galileo as "purpose-built to solve one of the hardest and most consequential problems in AI: Trust", and placed it in the Splunk division, to extend the AI Agent Monitoring capabilities in Splunk Observability Cloud. No financial terms were disclosed. Cisco said it expected the deal to close in the fourth quarter of its fiscal year 2026.

Splunk's post of 28 May said Cisco had completed the acquisition and named three capabilities Galileo brings: customisable evaluations, real-time guardrails that block harmful outputs within 200 milliseconds, and lower-cost evaluation on Luna-2 small language models.

What Galileo Brought to Splunk

Galileo built an AI evaluation and observability platform with more than 20 built-in evaluations for RAG, agents, safety and security. Its products included Luna-2 evaluation models, CLHF for tuning judge metrics from a small labeled set, and runtime controls through Protect and later Agent Control.

What Has Visibly Changed

• The name. Galileo's release notes: "As of August 7, 2026, Galileo is now Splunk Agent Observability." Splunk's product page describes it as a way to "observe, evaluate, and secure the agents, models, and infrastructure behind your AI".
• The docs. docs.galileo.ai covers customers who onboarded before 7 August 2026; customers after that date are pointed to agent-observability-docs.splunk.com.
• The packages. Splunk's Agent Control setup page lists a splunk-ao package (0.2.1 or later) alongside agent-control-sdk, agent-control-evaluators and agent-control-evaluator-galileo.
• The guardrail product. Protect was deprecated in June 2026, and new users are told to set up Agent Control. The timing overlaps with the acquisition; Galileo's docs give no reason for the change, and this page does not supply one.

What Has Not Visibly Changed

• Published plans. galileo.ai/pricing still lists Free ($0, 5,000 traces a month), Pro ($100 a month billed yearly, 50,000 traces) and Enterprise (contact sales).
• Luna-2's gating. Galileo's docs and Splunk's new docs both say "Luna-2 is only available in the Enterprise tier."
• Agent Control's license. The open-source repository is still Apache-2.0, with Python and TypeScript SDKs. Code already released under that license stays usable under it, whoever owns the company.

Several customer-impact details remain unpublished: how existing contracts change at renewal, whether galileo.ai continues as a separate storefront and how the product will be packaged beside Splunk Observability Cloud. Those details need confirmation from the account team.

What a Customer Should Ask

If Galileo is in your production path, get these answers in writing before renewal:

1. Which contract and price list apply at renewal: Galileo's or Splunk's?
2. Do the Free and Pro plans on galileo.ai continue, and where will they be sold?
3. Where is our data processed and stored now, and has the subprocessor list changed?
4. Which docs, SDK packages and API endpoints should we build against after 7 August 2026?
5. If we use Protect, what is its end date, and does our tier include help moving to Agent Control?
6. Does anything in our current plan now require a Splunk Observability Cloud subscription?
7. Who is our support contact, and do existing SLAs carry over?

Questions to ask an agent eval vendor covers the rest of a vendor review, and migrating between platforms covers what moves if you ever do leave.

What It Means for Your Decision

The acquisition alone is not a reason to replace a Galileo deployment that works. The product continues under the Splunk Agent Observability name, and previously released Apache-2.0 Agent Control code remains available under that licence. Review the contract, data terms and migration commitments before the next renewal.

For a team choosing a tool now, Galileo still offers 20+ built-in evaluations, Luna-2 and hosted guardrails on Enterprise, with Agent Control available as an open-source project. Galileo alternatives compares it with other platforms. Failproof AI vs Galileo is the closer comparison if you are deciding between Galileo's built-in evaluation catalog and Failproof AI's automated failure discovery, code-based and LLM-based evaluations, alerts and workflow for turning findings into tested fixes. Existing Splunk Observability Cloud customers should also ask how Agent Observability is licensed alongside their current plan.