Yes, I remember it well. My friend Jyoti Bansal, fresh off Cisco’s acquisition of his AppDynamics for over $2 billion, was looking for his next adventure. I remember speaking to Jyoti when they first announced Harness. He said when they launched AppDynamics, he saw that the APM market was outdated and ripe for disruption. It wasn’t being done right, given current technology. He looked at the then-current state of CI/CD, the beating heart of DevOps, and said, “This should work better.” And so was born Harness.io.
Reflecting back on this in light of today’s announcement of a $240 million round on a $5+ billion valuation, I am not surprised. Jyoti stuck to his vision and executed relentlessly. Where before the big three in DevOps platforms were GitLab, CloudBees and JFrog, Harness didn’t just sneak into the club — they busted the door down.
And today, they didn’t just bust down a door; they blew the hinges off.
Harness announced a $240 million Series E financing, including a $200M investment led by Goldman Sachs and a $40M tender offer with participation from IVP, Menlo Ventures, and Unusual Ventures. This raise values Harness at an impressive $5.5 billion. That’s not just a number — that’s a statement. A statement that the market believes the future of software delivery will be driven by AI, and Harness is the company building the platform to get us there.
AI is Breaking and Remaking the Delivery Pipeline
Let’s level with each other: AI is rewriting the rules of development. We all know the trope “AI writes code.” Cute. But writing code is only 30–40% of the engineering lifecycle. The other 60–70%? That’s what we lovingly call the “outer loop”:
Testing, verification, deployments, security, compliance, governance, optimization.
The messy parts.
The expensive parts.
The parts no one brags about on stage at conferences.
But here’s the real kicker: AI is making that outer loop bigger and more demanding. Code volume is skyrocketing — up to 4x in many orgs — and every single line still needs to be tested, secured, deployed, governed and observed. More code doesn’t magically mean more quality. Usually, it means more opportunities for risk to pile up.
This is where Harness has been playing — and winning — long before generative AI made Code With Me the new normal. And this is exactly why today’s raises land with such force.
Harness isn’t building “AI for developers.”
They’re building AI for everything after developers write (or use AI to write) code. It’s AI for software delivery AtoZ.
That may be the most important part of the entire AI transformation.
Harness AI: A New Architecture for a New Era
Harness is packaging this idea into something they’re calling Harness AI, a unified system specifically designed for the outer loop. And unlike the typical marketing mush we see in this industry, this architecture is actually crisp, coherent, and here’s the magic word, useful.
Harness AI has three foundational layers:
1. AI Agents Built for Software Delivery
Not generic LLM wrappers.
Not “AI copilots” that politely suggest things.
Actual focused agents that perform delivery, testing, verification, security, governance, and operational tasks.
2. The Software Delivery Knowledge Graph
This might be the real crown jewel.
A context engine that maps:
- code changes
- services
- deployments
- tests
- environments
- incidents
- policies
- cost signals
In other words: the entire delivery universe, stitched together so AI decisions aren’t guesses — they’re grounded.
3. Enterprise-Grade Orchestration Engine
Because insight without execution is trivia.
Harness’s orchestration engine turns AI-driven recommendations into safe, repeatable automation across pipelines and environments.
When you combine those three layers, you get something that has been missing from DevOps for a long time:
precision.
Not hype.
Not “maybe this works.”
Actual precision aligned to a company’s architecture, workflows, and risk profile. This is the embodiment of Jyoti’s vision of doing CD better from all those years ago, made fresh today. It is also true to the DevOps core, release quality software faster.
Receipts? Harness Has Them.
I’ve covered enough platforms to know that architecture diagrams are the easy part. Value is proven in customer outcomes. And this is where Harness’s story gets compelling — real enterprises are already seeing material, measurable gains.
Harness cites:
- United Airlines accelerated deployment times by 75% and migrated 80% of workloads to the cloud.
- Morningstar achieved 5x faster builds, consolidated 36,000 pipelines to 50 templates, and moved from weeks-long releases to daily deployments.
- Keller Williams increased deployment frequency 6x, saving three weeks in every release cycle.
- National Australia Bank cut build times by 67% and improved troubleshooting efficiency by 85%.
These aren’t hobbyist metrics. These are outcomes that affect regulated industries, global teams, real money, and real risk.
And scale? Harness is operating at a level most vendors in this space only dream about:
- 128 million deployments and 81 million builds in the last 12 months
- 1.2 trillion API calls protected
- $1.9 billion in cloud spend optimization
- Trusted by 1,000+ enterprise engineering teams
- A 1,200+ person global team across 14 offices
- On track to exceed $250M ARR in 2025, with 50%+ YoY growth
No matter how you slice those numbers, that’s momentum on an entirely different tier.
Why This Raise Matters — Beyond Harness Itself
For all the talk about AI rewriting the rules of software, one undeniable truth remains:
Software still needs to get to production — quickly, safely, and with governance baked in.
Today’s funding announcement validates something the DevOps community has known, but sometimes forgotten in the AI hype cycle:
DevOps is alive and well.
Not because the term refuses to die. Not because vendors cling to it. But because the problems DevOps was created to solve — bottlenecks, risk, reproducibility, resilience — are getting bigger in the AI era, not smaller.
Harness’s raise tells us:
- The market believes in platform consolidation, not tool sprawl.
- Enterprises are ready for AI-native delivery, not bolt-on gimmicks.
- Investors are backing companies that solve the outer loop bottleneck, not just code generation.
And yes, this raise puts Harness at the top of the DevOps platform hierarchy. For those of us who’ve been watching the rise of platform engineering, AI-native pipelines, and governance-at-scale, this is the logical next chapter.
The IPO Question — Let’s Be Honest
I’ve said it before: Jyoti regretted not getting AppDynamics to IPO. The company was ready. The market was ready. Then Cisco made an offer too large to refuse.
Founders don’t get many chances at a do-over.
Everything about this raise — the amount, the investors, the valuation, the timing — is a bright neon sign pointing toward a public offering. I’d be shocked if Harness doesn’t file in the next 12–18 months.
And unlike the last time, Jyoti holds the pen.
The Shimmy Finish
I’ve had a front-row seat to Harness’s evolution, from a CI/CD disruptor to a global platform defining the next era of software delivery. And today’s news doesn’t just feel like a win for Harness — it feels like a win for the entire DevOps ecosystem.
Because if AI is the crucible reshaping software delivery, Harness is one of the companies actually forging something new inside it.
So to Jyoti, the Harness team, and the thousands of engineers counting on this platform:
Congratulations. This is well-earned — and just the beginning.
Because if history has taught us anything, when Jyoti sets out to fix something that “should work better,” he doesn’t just fix it.
He changes the market around it.
Just for the nostalgic, here is the article and podcast from when Harness first announced on DevOps/com. Have a listen
