Loophole Labs – investing in the future of infrastructure

As workloads became more ephemeral, and less monolithic, this problem exacerbated. So much so that the standard was such that you could solve for portions of this, by focusing on specific compute architecture, or on Storage tiering, but not on the entire portability of the workload. Developers needed to decide on location, and if the app needed to be ported, you needed an expensive, complex and time-consuming migration project. This worsened with the complexity of the application.

In 2021, Shivansh (Shiv) Vij was facing this problem when he oversaw the management and reliability of a service while he worked at IBM. Once he determined there was no adequate solution in the market, he founded Loophole Labs, which was set out to solve this exact problem, and it's why we led the company's seed round.

Portability - the problem Cloud never really solved

Over the past two decades, infrastructure evolved from monolithic, on-premises applications into a cloud-native world of virtual machines, containers, and orchestration layers like Kubernetes. Moreover, with the surge of AI and agents, there is a need for additional tooling such as sandbox environments for agents. The sprawl of agents has created complexities related to making sure apps can be trusted in production and overcome the non-deterministic behavior of components in those apps and workflows. The AI shift unlocked speed and flexibility, but it also introduced a new set of structural problems that the industry hasn't fully cracked:

• Cost: Cloud spending has become the second-largest line item on many companies' P&Ls, right behind headcount. Discount mechanisms like reserved and spot instances help, but they come with reliability trade-offs that most workloads can't tolerate. With the introduction of agents, AI token costs exacerbated the problem and added another major cost.

• Reliability: Even at "five nines" availability, providers only guarantee low latency within a region — not across regions, and certainly not across providers. Outages still happen, and the blast radius can be significant.

• Complexity: As the number of cloud services has exploded, so has the burden on platform and infrastructure teams to understand, integrate, and maintain them. When you add on the complexity of non-deterministic agent behavior, architecting the infrastructure supporting those environments is even harder.

• Sovereignty: A growing patchwork of data residency and governance requirements is forcing companies to think carefully about where their workloads physically run.

Loophole Labs stepped into that reality with a fundamental thesis, that an intelligent abstracted layer, including networking, cloud storage, computing, virtualization and smart memory caching and management, can provide the building blocks for a more reliable, durable, cheap, yet performant infrastructure.

Building the hard part first, leveraging world-class engineers

Founder Shiv Vij, first ran into this problem while interning at IBM and finding no existing networking solution fast enough to meet the SLA he needed. Rather than start with a go-to-market wedge, Shiv and the Loophole team spent several years building the genuinely difficult technical layer: a coupled stack spanning networking, storage, and orchestration capable of migrating a live, stateful workload — CPU, memory, disk, virtualization and network state — from one destination to another in single-digit milliseconds. As industry tools evolved Loophole realized that they offset traditional tradeoffs by deterministically knowing their solution could durably run services like spot-instances with guaranteed SLAs because of their blazing-fast fail-over capabilities.

That's a hard engineering problem that larger, well-funded incumbents have tried and struggled with. It's also precisely the kind of technical moat Cervin looks for in companies: something that can't be bolted together quickly by a fast-follower, and that gets more valuable as the underlying market — cloud and AI infrastructure — continues to grow.

We were drawn to the team's engineering-first culture, their willingness to spend years hardening the core technology before chasing revenue, and Shiv's ability to synthesize feedback from a wide bench of infrastructure operators and repeatedly sharpen the product's focus.

A neutral, multi-cloud fabric powering production-grade agentic services

Cervin’s investment thesis was straightforward: as vendors lock-in and cloud and token costs become bigger strategic concerns, a neutral, multi-cloud, AI-native layer will emerge to serve the enterprises which traditional and neo-clouds are structurally disincentivized to serve well. No single cloud provider can offer high-availability guarantees for agents across clouds or regions without undermining its own business model — that gap is exactly where Loophole Labs built a durable, category-defining infrastructure.

By solving the scalability and reliability challenges, Loophole’s “Architect” platform, becomes a foundational piece of infrastructure — the kind of asset that every major agentic infrastructure provider should provide to customers, big and small.

What’s ahead – Loophole’s next chapter

We're excited to share that Cervin Ventures co-led Loophole’s Lab’s Seed round in March 2024. Over the past two and a half years, we’ve gotten to know and admire the technical execution of the Loophole team. They’ve grown to gain the trust of sizable enterprise customers and service engineering-forward companies.

Today, we’re also proud to share that Loophole Labs begins the next chapter of its mission and entered an agreement to be acquired by LiveKit. This milestone reflects years of disciplined technical execution by Shiv and the team — building a genuinely novel piece of infrastructure technology from first principles and proving out its value with some of the most demanding workloads in the industry. Cervin is proud to have partnered with Loophole Labs, and we're excited to watch the technology and team continue to grow with the reach and resources of a larger, engineering-first organization such as LiveKit behind them.

As Loophole scales one of the hardest, least-glamorous and most-consequential problems in agentic infrastructure, we're grateful to have backed a team willing to spend the time getting the fundamentals right. We look forward to seeing what this technology unlocks in its next chapter.

Congratulations to the Loophole and LiveKit teams

The Cervin team