Our $100M Series B

Raising our Series B round of financing

Bryan Cantrill, Steve Tuck
5 min readadvanced
--
View Original

Overview

Oxide Computer Company announces a $100M Series B funding round led by USIT, more than doubling their total capital raised to date. The article traces the company's journey from skeptical VCs in 2019 through building a complete hardware-software cloud computing stack for on-premises deployment, shipping their first system two years ago, and reaching general commercial availability with growing customer adoption.

What You'll Learn

1

Why co-designing hardware and software together produces superior on-premises cloud computing products

2

How Oxide built a full-stack cloud computer from custom board designs to a distributed control plane

3

Why transparency (open RFDs, documentation, source code, podcasts) accelerates enterprise sales cycles

4

How a startup can challenge incumbents by rethinking on-premises infrastructure from first principles

Prerequisites & Requirements

  • Basic understanding of cloud computing concepts and on-premises vs. cloud deployment models
  • Familiarity with data center infrastructure components (servers, networking, storage, hypervisors)(optional)

Key Questions Answered

How much funding did Oxide Computer raise in their Series B?
Oxide raised a $100M Series B led by USIT with participation from all existing investors. This more than doubles their total capital raised to date, bringing it from $89M over nearly six years to over $189M total. The funding positions Oxide to address customer demands around manufacturing scale, system scale, and operations scale.
What technology components did Oxide build from scratch for their cloud computer?
Oxide built custom board designs with hardware root-of-trust, their own microcontroller operating system (Hubris) replacing traditional BMC, custom platform enablement software eliminating UEFI BIOS, their own host hypervisor, a custom network switch and switch runtime (Dendrite), an integrated storage service, and a distributed control plane (Omicron) delivering API-driven elastic compute, virtual networking, and virtual storage.
Why did Oxide choose to build both hardware and software for on-premises cloud computing?
Oxide's thesis was that cloud computing is the future of all computing, that on-premises deployment remains strategically important for many organizations, and that the entire stack—hardware and software—needed to be rethought from first principles. Co-designing hardware and software together enables capabilities like true hardware root-of-trust, end-to-end power observability, and seamless user experience while eliminating costly third-party licensing.
How does Oxide's transparency help accelerate enterprise sales cycles?
Oxide found that new customers were already familiar with the company through their podcast, public RFDs (Requests for Discussion), documentation, and open source code. This transparency shortened traditionally long enterprise sales cycles because prospects could evaluate the technology deeply before engaging in a sales conversation, leading to surprisingly quick transitions from first contact to delivered product.
Who led Oxide Computer's Series B funding round?
USIT, founded by Thomas Tull, led Oxide's $100M Series B round with participation from all existing Oxide investors including Eclipse Ventures, their original seed investor. USIT shared Oxide's vision of a big, important on-premises cloud market uniquely served by Oxide's integrated hardware-software approach, with a shared disposition toward building lasting value.
What will Oxide use the $100M Series B funding for?
The funding will address customers' most pressing questions around three types of scale: manufacturing scale (ability to produce more racks), system scale (supporting larger deployments including multi-rack customers), and operations scale (making large numbers of racks easy to operate together). It also expands their roadmap scope for future product development.
How does Oxide eliminate traditional data center infrastructure complexity?
Oxide replaces multiple third-party components with integrated alternatives: custom board designs replace generic server motherboards, Hubris OS replaces the traditional BMC, custom platform enablement eliminates UEFI BIOS vulnerabilities, their own hypervisor removes third-party hypervisor licensing, Dendrite replaces complex network switch software, and integrated storage eliminates external storage dependencies—all managed through a unified API-driven control plane.

Key Statistics & Figures

Series B funding raised
$100M
Led by USIT with participation from all existing Oxide investors
Total capital raised prior to Series B
$89M
Over the nearly six-year lifetime of the company before the Series B
Company age
Nearly 6 years
From founding through Series B announcement
Time since first system shipped
2 years
First system shipped approximately two years before this announcement

Technologies & Tools

Operating System
Hubris
Oxide's custom microcontroller operating system used to replace the traditional BMC
Networking
Dendrite
Oxide's custom switch runtime, open source on GitHub
Control Plane
Omicron
Oxide's distributed control plane delivering API-driven elastic compute, virtual networking, and virtual storage
Hardware
Lpc55s69
Microcontroller used for hardware root-of-trust implementation

Key Actionable Insights

1
Co-designing hardware and software from first principles can yield dramatically better products than integrating third-party components. Oxide eliminated UEFI BIOS vulnerabilities, third-party hypervisor licensing costs, and complex network switch integration by building each component themselves with deep integration in mind.
This approach is applicable when building infrastructure products where integration complexity and licensing costs of existing components represent significant pain points for customers.
2
Radical transparency through open documentation, public RFDs, podcasts, and open source code can significantly shorten enterprise sales cycles. Oxide found that prospects who had already consumed their public technical content moved from first conversation to delivered product surprisingly quickly.
Enterprise infrastructure companies should consider publishing technical design documents, source code, and engineering discussions publicly as a sales acceleration strategy, not just a community-building exercise.
3
When building a deeply technical product, the biggest risk investors may see is market risk rather than technical risk. Oxide experienced VCs saying 'we know you can build this but don't think there's a market,' when the founders believed the opposite—that technical risk was high but market demand was certain.
Founders raising capital for infrastructure startups should be prepared to address market validation head-on, and seek investors who understand the target market's pain points firsthand rather than those unfamiliar with the domain.
4
Building a complete integrated system requires hitting milestones in sequence: board bring-up, networking, control plane, manufacturing, regulatory compliance (FCC), first shipment, first field update, and then general availability. Each milestone unlocks the next phase of customer and investor confidence.
Hardware-software startups should plan for a multi-year development timeline with clearly defined milestones, understanding that each completed milestone de-risks subsequent stages and builds credibility with customers.
5
Eliminating third-party dependencies in infrastructure (hypervisor licensing, BIOS vendors, switch software) not only reduces costs but also removes entire categories of security vulnerabilities and operational complexity for customers.
When evaluating build-vs-buy decisions for infrastructure components, consider not just the direct cost savings but the elimination of vulnerability surface area and integration complexity that third-party dependencies introduce.

Common Pitfalls

1
Venture capitalists may dismiss on-premises infrastructure markets by claiming 'there is no market' while simultaneously acknowledging the technical feasibility. This inverted risk assessment—underestimating market demand while overestimating the team's ability to build—can lead startups to over-rotate on market validation rather than focusing on the genuine technical challenges.
Seek investors who deeply understand the target market's pain points rather than trying to convince generalist VCs. Eclipse Ventures understood the on-premises market because they invest in companies where hardware-software co-design matters.
2
Relying on third-party infrastructure components (UEFI BIOS, hypervisors, switch software, BMC firmware) introduces security vulnerabilities, licensing costs, and integration complexity that can undermine the user experience. The traditional approach of assembling servers from commodity components leaves entire categories of problems unsolved.
Oxide's approach of building from first principles eliminated the UEFI BIOS and its accompanying flotilla of vulnerabilities, rapacious hypervisor software licensing, and complex switch integration—problems inherent to the traditional component-assembly model.
3
Focusing on individual technological components rather than the integrated product experience can cause teams to lose sight of the end goal. While each technology component is important, the objective is the product, not its parts—customers care about the complete solution, not the impressive sub-systems.
Oxide explicitly notes that each component serves specific customer problems, but emphasizes that shipping a complete, integrated product was the real milestone, not completing any individual technology piece.

Related Concepts

On-premises Cloud Computing
Hardware-software Co-design
Hardware Root-of-trust
Rack-scale Computing
Api-driven Infrastructure
Distributed Control Planes
Virtual Networking
Enterprise Sales Cycles
Deep Tech Venture Capital
Open Source Infrastructure
Bmc Replacement
Uefi Bios Alternatives
Elastic Compute
Virtual Storage