For a team spending $499 a month on cloud infrastructure, the cost of finding waste is often the waste itself. This is the wedge for CostGraph, a pre-seed FinOps platform that offers its core analysis for free to any customer whose total monthly cloud spend is under $500 [costgraph.ai, retrieved 2024]. The bet is that by embedding early with cost-conscious engineering teams, it can build a relationship that scales with their cloud bill.
Founded in 2025 and backed by Precursor Ventures, CostGraph is entering a crowded field with a specific technical focus [upskillist.com, retrieved 2026]. Its dashboard combines cost metrics from multiple providers with performance insights, aiming to give infrastructure teams a single pane for spotting and fixing inefficiencies across Kubernetes, virtual machines, and on-premises systems [costgraph.ai, retrieved 2024].
The Kubernetes-first wedge
While the platform casts a wide net, its initial technical depth is in Kubernetes cost visibility. The documentation details how it analyzes historical container utilization to recommend specific CPU and memory requests and limits, a common source of overspending in containerized environments [docs.baselinehq.cloud, retrieved 2024]. It aggregates pricing across clusters and supports Prometheus-based utilization data, positioning it as a tool built by engineers for the granular, real-world problems of running production K8s.
Recent development has focused on expanding provider support, with first-class integration for AWS launching in August 2026 [LinkedIn, August 2026]. This move is critical for broadening its addressable market beyond purely Kubernetes-centric shops.
The team and the traction signal
The public face of CostGraph is 21-year-old co-founder Umegbewe "Great" Nwebedu, an infrastructure engineer who previously led blockchain infrastructure at Botanix Labs and has been a keynote speaker at CNCF events [Techpoint Africa, November 2025] [thecondia.com, retrieved 2026]. The company's sole product designer, Shater Tsavsar, joined in 2025 to help build the platform from the ground up [shatermt.com, retrieved 2024]. The $500,000 pre-seed round provides runway to develop the product and test its freemium conversion motion [upskillist.com, retrieved 2026].
Public traction is early. The website features a testimonial from a founding engineer at a startup called Revyl, but verifiable enterprise customer announcements or detailed case studies have not yet surfaced [costgraph.ai, retrieved 2024]. The current go-to-market strategy appears entirely product-led, relying on the free tier to attract users who will eventually hit the spend threshold.
The competitive landscape
CostGraph is not entering a greenfield. It faces established, well-funded competitors, each with a different go-to-market angle and technical focus.
| Competitor | Primary Focus | Key Differentiator |
|---|---|---|
| Kubecost | Kubernetes cost monitoring | Market leader, deep K8s integration, enterprise sales. |
| OpenCost | Open-source cost standards | Free, community-driven specification, vendor-neutral. |
| Vantage | Multi-cloud cost analytics | Strong visualization, historical trend analysis. |
| CloudZero | SaaS and cloud cost intelligence | Focus on unit economics and cost per customer. |
CostGraph's differentiation rests on its combined freemium model and its promise to not just identify but also explain and recommend fixes for waste. Its challenge is to move beyond being a feature-alternative to these incumbents and establish a unique workflow.
The technical breakdown and scale risks
The platform's architecture, as described, follows a logical path for cost optimization. It pulls billing data from cloud provider APIs, correlates it with utilization metrics from Prometheus, and applies rules to flag anomalies and suggest rightsizing. The promise of "AI-powered" insights likely refers to pattern recognition across this dataset to surface non-obvious waste, like orphaned resources or underutilized reserved instances.
The real test comes at scale. For the free tier to be sustainable, the cost of serving a customer must remain below the lifetime value of the fraction that converts to paid. More critically, the analysis that works for a $500-a-month startup can break down for a complex enterprise with $500,000 in monthly spend, distributed teams, and custom discount agreements. The recommendations engine must evolve from suggesting simple container rightsizing to navigating intricate committed-use discounts, cross-region cost arbitrage, and application-specific performance trade-offs that finance teams will scrutinize. If the platform's AI is trained primarily on data from small-scale early users, its advice may not generalize upward, creating a ceiling on its utility just as a customer's needs become most valuable.
Sources
- [costgraph.ai, retrieved 2024] CostGraph - Cloud savings you can act on | https://costgraph.ai/
- [docs.baselinehq.cloud, retrieved 2024] Introduction - CostGraph Documentation | https://docs.baselinehq.cloud/concepts/how-it-works
- [Techpoint Africa, November 2025] How this 21-year-old is helping global companies build resilience | https://techpoint.africa/feature/21-year-old-building-global-infrastructure/
- [shatermt.com, retrieved 2024] Shater Tsavsar Portfolio | https://www.shatermt.com/work/costgraph
- [LinkedIn, August 2026] LinkedIn post by Ott Salmar | https://www.linkedin.com/posts/ott-salmar_aws-billing-and-cost-management-introduces-activity-7494992411346214912-9fdQ
- [thecondia.com, retrieved 2026] Umegbewe Nwebedu built his first website at 9 | https://thecondia.com/umegbewes-exciting-career-path/
- [upskillist.com, retrieved 2026] CostGraph funding information | https://upskillist.com/
- [startups.com, retrieved 2026] CostGraph funding information | https://startups.com/