StackOne Raises $20 Million to Power AI Integration Infrastructure After Hitting 1 Billion API Calls
A New AI Infrastructure Contender Emerges: StackOne Bets Big on Rewiring the Integration Layer
As AI agents flood the enterprise, StackOne positions itself not as another connector tool—but as the backbone for a new generation of software. With $20 million in fresh funding and 1 billion API calls already processed, can the startup outmaneuver billion-dollar incumbents and become the AI-native control layer for enterprise workflows?
The Integration Arms Race Enters Its Next Phase
At a time when software giants are racing to retrofit their stacks with AI, a lesser-known London startup is quietly laying claim to one of the most valuable choke points in the enterprise: integration. StackOne, a two-year-old company founded by former Google and Oracle engineers Romain Sestier and Guillaume Lebedel, today announced a $20 million Series A funding round led by GV (Google Ventures), with participation from Workday Ventures and other deep-pocketed backers. But the money is just the headline. The ambition is more audacious.
StackOne is building the infrastructure layer that AI agents will run on.
Its pitch is as sharp as it is technical: while legacy vendors stitch together APIs with templates and manual rules, StackOne’s AI agent generates and maintains deep integrations autonomously—promising a 5x drop in deployment errors and an exponential gain in speed. The company has already processed over 1 billion API calls, signed up enterprise customers like Drata and Mindtools, and launched more than 200 production-grade connectors ranging from HR systems to IAM.
For enterprise tech watchers, this is not just another platform story. It’s a signal that integration—long relegated to middleware plumbing—is becoming the most strategic surface area in the age of AI.
The Bigger Battlefield: Control Layers for Autonomous Workflows
Enterprise integration used to be boring. Not anymore.
As companies deploy AI agents to automate increasingly complex workflows—customer support, financial operations, employee onboarding—the ability to move data instantly and securely across software boundaries has become existential. Analysts expect the integration platform market, already at $8 billion in 2023, to surge past $17.5 billion by 2033, outpacing the core SaaS market itself.
“Every agent needs real-time access to business systems,” noted one market analyst. “And every system now exposes APIs. Whoever controls that layer controls the stack.”
But control won’t come easy.
The incumbents—Boomi, MuleSoft, Workato, Zapier—are well-entrenched, with vast customer bases and hundreds of millions in annual revenue. Boomi, for instance, boasts $500 million ARR and over 20,000 customers. MuleSoft’s revenue feeds into Salesforce’s $5.2 billion integration and analytics segment. Workato grew revenue from $86 million in 2022 to an estimated $150 million+ in 2024, serving 11,000 customers.
By contrast, StackOne’s revenue remains undisclosed. And yet, its traction is hard to ignore. Customers like Drata reportedly launched 100 integrations in a matter of weeks. Mindtools won a seven-figure Workday contract by deploying StackOne’s platform in record time.
These wins, while anecdotal, underscore a critical inflection point: AI is changing not just what software does—but how it connects.
From Template-Led to Agent-Native: A Technological Shift
At the heart of StackOne’s approach is its proprietary AI agent, which automatically generates, tests, and maintains API integrations. This stands in contrast to legacy platforms that rely on prebuilt templates and human configuration.
“The agent doesn’t just wire things together—it learns system schemas, maps fields dynamically, and rewires logic as APIs evolve,” said one technical advisor familiar with the company.
StackOne’s real-time, event-driven architecture also reflects a deep bet on system design. The platform supports bidirectional data flows, encrypted secret vaults, and single-tenant runtimes—a must for security-conscious enterprises. With over 3,000 actions supported across 200+ connectors, it already rivals some mid-market offerings in breadth, if not yet scale.
This technical posture is not theoretical. In production, StackOne’s infrastructure has served customers across three continents, delivering integrations in weeks that would typically take months. Still, questions remain about its scalability under enterprise-grade load and how it handles complex edge cases—especially when AI is generating code on the fly.
StackOne vs. the Field: Can Speed Beat Scale?
| Vendor | **ARR ** | Key Advantage | Exposure to StackOne’s Threat |
|---|---|---|---|
| Boomi | $500M+ | Massive ecosystem, deep SAP/Oracle hooks | High (mid-market overlap) |
| MuleSoft | $400M–$500M* | Salesforce distribution, data fabric tools | Medium (slow to evolve tech) |
| Workato | $150M+ | AI/low-code hybrid UX | High (AI-native angle clash) |
| Zapier | $310M | SMB dominance, viral growth | Low (different segment) |
*Part of Salesforce’s broader $5.2B Integration & Analytics revenue.
In a space where incumbents enjoy massive network effects—Boomi has 250,000 users; Zapier has over 3 million—StackOne’s agent-led automation could become a differentiator. But it could also be easily imitated if commoditized open-source models catch up.
The stakes are enormous. If StackOne’s agent model can compress six months of integration work into hours, it won’t just compete with other platforms—it will change how software is built.
Investor Signals: Betting on a Category Breakout
The investor syndicate tells its own story. In addition to GV and Workday Ventures, StackOne is backed by XTX Ventures, Playfair Capital, Episode 1, and angels from OpenAI, DeepMind, Microsoft, and MuleSoft. These names reflect a blend of AI, enterprise SaaS, and deep infrastructure expertise.
“What they’re building is foundational,” one investor noted. “You can’t have autonomous agents if they can’t talk to your business systems securely and reliably.”
The funding environment is also tilting in StackOne’s favor. AI integration startups raised €481 million in Europe alone during the first six weeks of 2025. While capital may become selective later this year, agent-based infrastructure remains a red-hot thesis for VCs chasing the next generational platform shift.
Risks on the Horizon: From Security to Sustainability
Despite the promise, StackOne faces several non-trivial risks:
- Security & Governance: Autonomous agents manipulating enterprise data pose new compliance challenges. Integration flows must be auditable, killable, and explainable.
- Model Risk: AI agents can hallucinate. A flawed field mapping or misrouted record could cause serious downstream errors—especially in HR, finance, or healthcare systems.
- Unit Economics: Event-driven, low-latency integrations consume bandwidth and compute. If cloud providers tighten discounts, margins could suffer.
- Sales Motion Complexity: Enterprise integration deals often require 6–12 month sales cycles, meaning StackOne must carefully manage cash burn and customer acquisition.
While its $20 million war chest gives it a 24–30 month runway by most estimates, success will hinge on converting pilots into revenue—and doing so faster than the competition can respond.
Three Futures for StackOne
| Scenario | Probability | Implication |
|---|---|---|
| Wins as a New Standard Layer | 40% | Becomes default for AI agent integrations; $100M+ ARR by 2028 |
| Niche Success / Acquihire | 35% | Specialized tool bought by Workday, Snowflake, or similar |
| Outpaced by Open-Source or Incumbents | 25% | Loses differentiation, pivots or fades |
The tipping point may come when a Fortune 500 enterprise uses StackOne to deploy a full integration suite—at scale—in under 10 days. That would mark a real test of its architecture, support model, and monetization viability.
Not Just Another API Startup
StackOne is not playing the same game as its predecessors. By treating integration as the substrate for AI-native workflows, it reframes a mature market as an unsolved frontier. And while the company is early, the ambition and initial traction suggest it could redefine what integration even means.
The question now isn’t whether integrations matter. It’s who controls them—and how fast they can adapt to a world run by machines.
If StackOne is right, they won’t just help enterprises plug into AI—they’ll become the plug itself.