The Timing Problem Nobody Talks About
Stacklet launched in August 2020 with $4 million in seed funding and a thesis: enterprises needed an administrative layer for cloud governance that didn't exist. The founders, Travis Stanfield and Kapil Thangavelu, had spent years inside Capital One's engineering organization watching teams build one-off scripts to keep cloud resources from spiraling. They packaged what they learned into a commercial product.
The problem? The market didn't care yet.
In 2020, companies were still in full cloud-migration mode. Spend was going up, but nobody was losing sleep over it. The macroeconomic environment that would eventually make cost control a board-level topic — pandemic aftershocks, rising interest rates, a general end to easy money — hadn't landed yet. Stanfield told TechCrunch that most companies weren't "all that concerned with constraining cloud costs" when the company came out of stealth.
Three years later, that had changed completely. Revenue tripled year over year in 2023, and in June 2024 Stacklet raised a $14.5 million round ($12 million from new investor SFE Equity, $2.5 million from existing backers) bringing total funding to $36.5 million. The company now counts more than 100 customers, including Lenovo, Petco, Western & Southern Financial Group, and three-quarters of the top ten investment banks. Their average annual cloud spend: roughly $1 billion.
This is a case study in what happens when a developer tool finds its moment.
Open Source Roots: Cloud Custodian at Capital One
The whole thing starts with a rules engine. Around 2013, Capital One was in the middle of migrating its entire data center to AWS — the first major financial institution to attempt that at scale. The move created governance headaches that no existing tool solved cleanly. Teams were writing custom scripts to enforce policies, hunt for misconfigurations, and keep compliance from falling through the cracks.
Thangavelu, who had joined Capital One from the consulting world, built Cloud Custodian to handle this. It started as an internal tool, then got open-sourced in 2015. The approach was deceptively simple: policies written in YAML, enforced automatically across whatever cloud infrastructure you were running. AWS, Azure, GCP, Kubernetes — if it spoke the API, Cloud Custodian could police it.
"We wanted to be able to take a policy and write it in a simple syntax, like YAML," Thangavelu explained at launch, "and then automate the enforcement of that policy across whatever the cloud platform you're using."
By the time Stacklet launched commercially in August 2020, Cloud Custodian had about 1,000 GitHub stars and was pulling in 20,000 monthly downloads. The community had grown organically — engineers at other banks and financial services firms had picked it up, and non-financial companies were starting to use it too.
The commercial thesis was straightforward: give companies a management interface on top of the open-source engine. Dashboards. Role-based access control. Audit trails. The stuff that makes a tool safe for an operations team to trust at enterprise scale.
What Stacklet Actually Does
Strip away the marketing and the platform automates a handful of practical problems that cloud teams deal with daily:
Rightsize over-provisioned resources. Clean up idle infrastructure. Enforce tagging policies so teams can actually see where spend is going. Flag security misconfigurations — public S3 buckets, databases with open security groups, before they become incidents. And enforce governance guardrails so engineers don't accidentally spin up resources in the wrong region or violate a compliance requirement.
Stanfield positions this as a unified category. Cost tools like CloudZero or Mavvix optimize spending. Security tools like Enso or Orca scan for vulnerabilities. Stacklet claims to be the only vendor covering all three angles, cost, security, governance, under one product.
That positioning is convenient, whether or not it holds up in a feature-by-feature comparison. The real insight is that these problems share infrastructure. A misconfigured database that costs you money is also a security risk. The same YAML policy that stops an idle VM from burning cash also prevents an engineer from bypassing tagging requirements that make cost attribution possible.
Scaling When the Market Finally Caught Up
Stacklet's revenue trajectory tracks almost perfectly with the macro story.
The $18 million Series A in January 2021 came at a moment when cloud spend was still climbing and the pain was real but not yet urgent for most enterprises. Addition led the round; Foundation Capital (which also led the seed) participated, along with Liam Randall, a veteran of enterprise software who joined as VP of business development.
Then the environment shifted. Interest rates rose. Budgets tightened. CFOs started asking pointed questions about cloud line items that had previously been invisible.
By June 2024, Stacklet was reporting 3x revenue growth year over year. The company had acquired Cloud Kente's management platform, one of the few vendors that had actually built something on top of Cloud Custodian with real enterprise features, in April. That acquisition signals consolidation: if you're building a governance platform, the smart move is to absorb the competition rather than watch them poach customers with a more polished product.
The investor base tells its own story. Addition and Foundation Capital backed the company from day one. SFE Equity wrote the largest single check in the most recent round. All three saw a market transition from nice-to-have optimization to mandatory governance infrastructure.
AI Workloads and the Next Cost Wave
The TechCrunch interview with Stanfield in June 2024 flagged something that will define the next chapter for companies like Stacklet: AI workloads are expanding the cloud spend footprint.
GPU compute is expensive. Training and inference on large models creates consumption patterns that traditional cost management wasn't designed to parse. When your infrastructure bill includes line items for model serving across multiple regions, the governance challenge gets qualitatively different from "stop the idle VMs."
This connects directly to what many engineering leaders are asking right now: how much does an LLM cost? The answer depends on model size, inference volume, whether you're self-hosting or using a managed API, and dozens of other variables. But the broader point is that AI compute is becoming a significant slice of cloud budgets, and most cost management platforms aren't built to attribute or optimize those workloads at the granularity enterprises need.
Scalable AI compute introduces a governance problem that mirrors what Cloud Custodian solved for traditional infrastructure: you need policies that are declarative, enforced automatically, and consistent across heterogeneous environments. The YAML-based approach that made Cloud Custodian workable for AWS resources in 2015 has a parallel in how you'd govern GPU clusters and model endpoints today.
For Stacklet specifically, the AI angle isn't a pivot, it's an extension of the same problem. More spend categories, more teams touching infrastructure, more surface area where a missing tag or an idle resource creates waste. The company that positioned itself at the intersection of cost and governance is naturally positioned to absorb the AI spend line item into its product.
Lessons From Stacklet's Trajectory
A few things stand out from watching this company's path from open-source tool to $36.5 million in venture funding.
Developer adoption isn't a commercial strategy, but it's a powerful starting point. Cloud Custodian gave Stacklet something almost no other cost management startup had: organic trust from the engineering community. When you're selling to CTOs and platform teams, "we built this because it worked at Capital One and thousands of engineers use it for free" is a credibility shortcut that no amount of marketing spend can buy.
Category timing matters more than product superiority. Stacklet had a real product in 2020. The market just wasn't ready for the urgency. Revenue "tripled" in 2023 not because the product suddenly got 3x better, it got better, sure, but because the macro environment made cloud cost control non-negotiable. The same company, with the same technology, in 2019 or 2020, would have been a much harder sell.
Consolidation signals category maturation. The Cloud Kente acquisition, absorbing a competitor that was building on the same open-source foundation, is the kind of move companies make when the easy land grab is done and growth comes from squeezing efficiency out of the existing player pool.
If you're evaluating the enterprise case studies in cloud cost governance, Stacklet's story is instructive. The company didn't invent cloud cost management. It didn't invent cloud governance. What it did was find a single open-source tool that solved a real problem well enough to build trust, then wrap the enterprise features around it at precisely the moment market pressure made those features mandatory.
The next test is AI. Cloud spend was already growing fast before the current wave of AI infrastructure pushed it further. The companies that build governance layers for that spend, the ones that extend the same policy-as-code approach to GPU clusters, model endpoints, and the sprawling compute footprints that LLM inference creates, will be the ones writing the next round of case studies in cloud economics. Stacklet has the community trust and the product foundation to compete. Whether the open-source approach translates cleanly to AI infrastructure governance remains to be seen.
The company's own roadmap will tell us. With $36.5 million raised, a 100-customer base averaging $1 billion in cloud spend, and a market that Stanfield describes as still early-stage for governance tooling, Stacklet has room to find out.
