Investing in Oxide: Unlocking Cloud Benefits for On-prem Data Centers

Today, we are excited to formally announce Intel Capital’s investment in Oxide Computer Company. We believe the time has come for a new system architecture that unlocks the benefits of cloud for on-prem data centers, and we feel the opportunity is massive.

The last two decades have seen the most advanced hyperscale operators and Cloud Service Providers (CSPs) arrive at the same conclusion: to materially improve their costs, performance and energy efficiency of their data centers, they needed to take a new approach: designing their own data center infrastructure from the ground up, including the vertical integration of servers, storage, networking and power supplies into custom racks. These industry leaders have captured the significant benefits of this approach which largely remain beyond the reach of others. But with Oxide, on-prem datacenter operators finally have access to the same cloud-scale economics and performance advantages.

Oxide: Bringing Cloud Benefits to On-prem DCs

Today, most enterprises (outside of the hyperscalers) are required to “rack and stack” their own data centers or colo facilities — deploying and configuring servers, storage, networking equipment, and various software stacks provided by different vendors — or move workloads to the public cloud. While opinions run deep on both sides, the reality is there is no one-size solution that works across the spectrum of workloads and business requirements. Users must assess the tradeoffs between cost, agility and complexity in determining the optimal manner to run their business, and the rapid adoption of cloud reflects a strong desire to optimize for agility. At the same time, many customers have found that over time the tradeoffs shift. This has led to some prominent examples of cloud repatriation in recent years as detailed in the Dropbox IPO registration statement. It’s a particularly important analysis for SaaS ISVs given cloud costs can make up such a large portion of their total cost of revenue. Ultimately, we feel all companies will need to get smarter in assessing where it is optimal to run each of their workloads based on ROI and other operational requirements such as data sovereignty. Many workloads will benefit from running on-prem, and those workloads deserve superior infrastructure economics combined with a developer experience that is on par with what they get in the cloud.

The appeal of Oxide is clear. Customers gain next-gen rack-scale infrastructure that is as performant and energy-efficient as what the CSPs offer while retaining the visibility and control necessary to optimize their operations, all with a developer-friendly API interface. This isn’t a panacea for all workloads, especially those that have highly elastic usage patterns or rely on specialized cloud services. But for more predictable workloads, we believe Oxide offers an incredibly compelling choice of infrastructure, preserving the ease and flexibility of cloud while fundamentally altering the build versus rent decision.

The Tightly Integrated Advantage

Oxide’s rack-scale solution includes 32 compute sleds with over 2,000 CPU cores, a network switch, up to 32TB of DRAM and 1024TB of NVMe storage, all vertically integrated alongside an advanced software stack and deployed in a 12-15 kW footprint. By tightly coupling the hardware and software components in a complete rack, users can get up and running in a matter of hours compared to months for many datacenter deployments. Further, with Oxide, customers can provision more than double the compute capacity in a lower power footprint using 1/3 less physical space, driving a 70%+ improvement in performance per watt compared to the traditional rack and stack options. Oxide’s TCO can be 50%-75% lower than relevant instance types available in the leading CSPs, while Oxide’s operational console provides significantly deeper visibility into the state and performance of the infrastructure. And in a world of constant threats, the entire system is designed from the ground up to minimize security threats across all layers. While it’s not a consumer product, Oxide’s tightly coupled approach is not unlike what Apple has done with the iPhone or Tesla has done with its vehicles. Post install, seamless software updates to Oxide racks enable new features and performance improvements in a way that’s not seen today with the traditional rack and stack approach.

The Team

If launching new systems companies were compared to Olympic diving, it would undoubtedly be assigned the highest degree of difficulty. Unlike ISVs which can get to an MVP product on modest resources, systems companies require a significant amount of upfront engineering cycles before you even know if you have a product that hunts. No surprise that many investors shy away from these types of bets. And it’s not enough to be playing in a big market, you also need a team that can deliver magic. We are fortunate to have worked with Steve and Bryan before and honored to partner again. Not only are they passionate entrepreneurs who are fun to work with, but their deep expertise in cloud and systems coupled with their ability to assemble one of the smartest technical teams we have seen compelled us to invest in Oxide. With the Series A funding in hand and the first set of customers installed and running, we’re excited to back Oxide in its quest to build the next iconic systems company.

The Future is Now

William Gibson famously said, “the future is here, it’s just not evenly distributed.” When it comes to cloud infrastructure, the hyperscalers have been living in the future while on-prem datacenters have been living in the past. Until now. Oxide is unlocking the value of highly performant rack-scale infrastructure for the rest of the world. For those interested in taking advantage, please check out Oxide.computer to learn more or add yourself to the waiting list.