Edge Cloud: A Major Refactoring is Underway

One of the themes we explored at my previous firm, NGP Capital, was the Edge Cloud. It is the thesis that we are moving past "peak centralization" of infrastructure driven by three important trends: billions of connectable devices already in the field, the availability of 5G connectivity, and the explosion of data being generated outside of the enterprise. This move to the edge is however, not a replacement for the centralized processing provided by the large cloud providers.

NGP Capital's Edge Cloud graphic

Rather, expanded edge workloads will use central cloud infrastructure and the two layers will interact in delightful ways. The explosive growth of edge and cloud infrastructure layers has created the need for a robust Access and Aggregation Edge in the middle that provides the impedance match between the diversity of device edge ecosystems and the amazing power of the cloud.

At NewBuild.VC, we continue to dive deeper into this thesis, which we expect will play out over the next decade. This market tailwind is a $150B+ revenue opportunity by 2025 for infrastructure companies. Incumbent Saas and Cloud Leaders are well-positioned to expand their footprint to the edge, but startups are driving new design points that challenge cloud dominance, expand the market, and win. This article explores some of the insights that are powering the Edge Cloud startup ecosystem.

Developers (still) Rule

The cloud has trained 25M developers worldwide to interact with sophisticated infrastructure and use the "right tool for the right job", calling resources at different levels of abstraction to optimize their applications. Developer centric infrastructure companies from Twilio to Datadog have built successful businesses identifying opportunities to package complex infrastructure into consumable services. With all this expertise out there, expect magic to happen when application developers can call edge connectivity, distributed compute, real-time data, and AI infrastructure within their normal workbenches.

As BVP memorably put it in their Scaling to $100M playbook, market leading Saas companies find a "second act" to drive ARR growth. We believe that many cloud infrastructure Saas companies will find their second act delivering edge infrastructure as services to their developer audiences. When evaluating these companies, we pay attention not just to whether there is a near term developer need for this service, but also whether the architecture elegantly supports the edge cloud. Companies we admire on this front, are Subspace, which turns low-latency CDN functionality into a callable feature for game developers, and Pulumi, whose infrastructure as code software is architecturally suited for setting up and tearing down ephemeral infrastructure.

Edge Cloud is the architecture for digital industries startups.

We expect this market tailwind to be $150B+ revenue opportunity by 2025. Incumbent Saas and Cloud Leaders are well-positioned to expand their footprint to the edge, but startups are driving new design points that challenge cloud dominance, expand the market, and win.

Digital Industries Need Edge Cloud

In 2015, McKinsey published a report on IoT where they predicted that the Internet of Things would disrupt the 85% of GDP that the internet was yet to disrupt, and unlock $11 Trillion of value by 2025. While predicting is hard, especially about the future, they were directionally right: we have seen sensors and network infrastructure expand dramaticatically in the intervening years. There are now billions of internet connected devices from treadmills to warehouse robots that are running purpose-built cloud applications. These form the left-most device edge layer of our Edge Cloud architecture. McKinsey updated their study recently, offering that 65% of the value of IoT will be in business settings and among these, factory and operations use cases are substantial drivers.

The pandemic accelerated remote work, dark stores and factories, and there is a supply chain reorganization underway, all of which drives Edge Cloud adoption. Incumbents like AWS have taken notice and plan to use this transformational moment to draw these vertical industries into their cloud orbit. But there is plenty of room for Saas startups to build new applications and vertical relevant infrastructure. My partner, Chad Bailey, writes more about this nearby.

Applications come to Data

Another reason the Edge Cloud architecture is inevitable is because most enterprise data is already generated outside enterprise walls. Streaming data allow real-time decisions that in turn drive productivity and deliver customer delight.

While generated at the edge, most data will eventually rest in the cloud. But where it is processed - right at the edge, at an aggregation point in the middle, or in a cloud data center - depends on the "3 laws of data". The law of physics asserts that low latency applications that can't wait for data to get to the cloud to be processed, will drive a move of compute, storage, performance monitoring and security infrastructure to move closer to the edge instead. The law of economics posits that when individual pieces of data, say for a predictive maintenance application, are not valuable enough to ship to the central cloud, applications will migrate close to the data to derive insights quickly and save the data slowly (or not at all). And finally the laws of the land such as regulatory limitations that require data to stay in a particular location or privacy protocols drive processing to move.

These data constraints and application needs cross industry verticals, promising the growth of new horizontal categories. Companies we admire in this mold are Esper, which brings cloud Devops to the edge, and Macrometa, which has build a global data network modeled after CDNs.

Larger Edge Metaphor

The Edge Cloud framework allows us to anticipate how a robust edge industry can interact with the cloud, yielding disruptive ways of solving challenging problems. We recently invested in a Web3 company that provides global network infrastructure, but interestingly, much of the control plane and pub-sub analytics reside in the cloud.

When companies such as Aerohive and Meraki built cloud controlled networks a decade ago, they had to build their cloud operations from scratch. Today, we see startups building control planes in the public cloud for distributed network and security appliances, dramatically improving manageability. Enterprise networking and security companies that build integrated appliances can expect competition from startups with the advantages of an Edge Cloud architecture, and indirectly see webscalers challenge today's enterprise stack.

We are incredibly excited to continue investing in the Edge Cloud at our new firm. If you are working in these areas and are looking for thought partners on your business journey, please connect with us.

written by