Investing in Galileo, the Leading Gen AI Testing and Monitoring Platform
Generative AI (Gen AI) presents an unprecedented opportunity for enterprises to make their workforces more efficient, personalize their products and services for their customers and innovate within their fields. No wonder, then, that Gartner estimates that by 2026 over 80% of enterprises will have implemented one or more Gen AI-enabled tools — adding up to $318 billion in software spending by 2032, per a recent Bloomberg report.
As enterprises around the world work to deploy Gen AI within their tech stacks, one issue looms especially large: how to test, measure and monitor AI systems for accuracy, performance and safety. Due to their non-deterministic nature, large language models (LLMs) and the applications built upon them can be unpredictable and difficult to oversee, leaving them susceptible to both hallucinations and malicious actors who seek to reveal sensitive data or affect their outputs. These concerns make it hard for enterprises to adopt Gen AI at scale, particularly in industries such as financial services where accuracy and security are paramount.
This makes Gen AI evaluation, observability and protection both a critical emerging field and one that offers significant market potential for automated solutions. And the leading provider of such solutions is Galileo, the newest addition to the Citi Ventures portfolio.
Founded by an impressive roster of AI experts who have all experienced the AI measurement problem first-hand — CEO Vikram Chatterji, COO Yash Sheth and CTO Atindriyo Sanyal previously led AI efforts at Google AI, Google Brain and Uber AI, respectively — Galileo has built a scalable, AI-driven Evaluation Intelligence platform that helps companies ensure the safe, effective performance of LLMs and Gen AI applications in both development and production.
The platform is powered by Galileo Luna, the company’s proprietary family of Evaluation Foundation Models (EFMs). EFMs are Gen AI foundation models purpose-built for evaluations and trained on smaller, more specific datasets than LLMs are — making them more accurate, faster and lower cost for particular tasks such as hallucination detection. On top of Luna lie four modules:
- Galileo Fine-Tune, which algorithmically improves fine-tuned LLMs and training data
- Galileo Evaluate, which empowers rapid, collaborative experimentation and testing of Gen AI applications. Evaluate offers tracing and visualization features, guardrail metrics, experiment management tools and more.
- Galileo Observe, which provides real-time monitoring and rapid debugging of Gen AI applications in production and alerts users about issues with the application
- Galileo Protect, which can intercept malicious LLM prompts and outputs to prevent unwanted behaviors, secure sensitive data and avoid hallucinations.
Taken together, Galileo’s GenAI Studio provides a user-friendly, low-cost and low-latency LLM evaluation and observability stack that is perfect for enterprises, mid-market companies and more. Thus, it is no surprise that Galileo’s revenue has grown by an astonishing 834% since the beginning of 2024 and that the company has quadrupled its number of enterprise customers in that time — including six Fortune 50 companies. As Gen AI adoption increases and advanced AI methods like retrieval-augmented generation (RAG) and agentic workflows become the norm, we expect demand for Galileo’s platform to continue to grow at a rapid pace.
“As several AI experts from Citi and industry practitioners noted during the recent Citi® Gen AI Summit, RAG-based Gen AI applications cannot be deployed in production without a best-in-class evaluation framework that can measure the quality of the applications pre- and post-production. We are very excited about the adoption of Galileo’s ground-breaking Luna models by top enterprises across financial services, media, consumer products, industrial, high-tech and many other industries to lower costs and increase accuracy.”
– Vibhor Rastogi, Head of AI Investing, Citi Ventures
With accurate, secure Gen AI core to many enterprises’ long-term strategies — including Citi’s — we are delighted to announce our investment in Galileo’s Series B funding round. The round was led by Scale Venture Partners and also included Databricks Ventures, ServiceNow Ventures, Amex Ventures, SentinelOne Ventures, Battery Ventures, Walden Capital, Factory and AI leaders like Clement Delangue, CEO of HuggingFace and Ankit Sobti, CTO of Postman.
Our congratulations to the Galileo team on their tremendous success thus far. We look forward to working with you all well into the future!
For more information, email Matt Carbonara at matt.carbonara@citi.com, Vibhor Rastogi at vibhor.rastogi@citi.com, Cagla Kaymaz at cagla.kaymaz@citi.com or Olivia Zhang at olivia.y.zhang@citi.com.
To see Citi Ventures’ full portfolio of companies, click here.