The edge computing software segment is expected to be one of the most rapidly growing and diverse segments within the entire edge computing market and it includes platform, AI and analytics, as well as supporting segments such as DCIM and edge OS. The platform, AI and analytics segments are three of the largest within the edge computing software space and are also the fastest growing. Our analysis shows edge analytics and AI start-ups are among the most acquired companies.
In the following table we show the different kinds of software that are used in edge networks along with their functionality and market opportunity.
Driven by AI, ML and DL, edge analytics to be one of the most important segments
Analytics is an integral part of edge computing and all types of end users—from drone operators to factory floor automation—will require analytics. As Multi-access Edge Computing (MEC) will place the computational processing power closer to the end user, we believe the challenge will be in the form of managing thousands of heterogeneous connections under strict response constraints for applications, service creation, and network administration. We believe machine learning (ML) and deep learning (DL) will be the way ahead for 5G providers to benefit from MEC yet get around this complexity. However, ML and DL are not standalone solutions, and they are meaningless unless the data is analyzed and acted upon, which is where analytics will come in. CIR believes edge analytics will be one of the most in demand software solutions at present, with use cases ranging from autonomous cars and smart buildings (with real-time security control) to factory floor cobots.
Edge platform and opportunities in the open-source space
Another software segment with high growth potential is edge platform. Significant demand will come from telcos, who will use platforms to create customer facing edge apps as new profit-making avenues. While the platform segment may mature more rapidly than both AI and analytics segments, open-source computing platform will see rapid growth. The biggest advantage of open-source edge computing platform is the freedom it gives to developers to develop and deploy apps closer to the user. This will be a win-win for all as giving the developers a great amount of freedom will only mean better and more efficient apps. For telcos, partnering with these third-party app developers will help them with customized solutions and apps depending on customer needs.
Edge platform and analytics remains key focus area for large players and VC funds
Edge AI, analytics and platform are not just demand driven segments but are getting significant push from venture capital funds as well. Key companies in this space such as Adapdix, EDJX, Edgify, FogHorn, Olea Edge Analytics, and Pensando have raised more than $200 million in combined funding.
Edge platform, AI and analytics have garnered interest from large tech majors and other large market players as well. Deutsche Telekom entered the open-source platform space through a partnership with Aricent (which is a part of Altran Group). FogHorn, a key company in the edge AI space has received funding from (among others) Bosch, Dell, GE, Honeywell, Intel. HPE was a key participant in the latest funding round of Pensando. Xnor.ai was purchased by Apple in January 2020 at an estimated price of around $200 million.
Security segment is currently undervalued
While here remains a significant current focus on edge platform and analytics (and AI), we believe security is another area to watch as device edge picks up. As large amounts of critical data will be processed on the derive device or near to the source (such as cell towers) (local processing) security may come back as software instead of service (ie, SaaS-based offering delivered from a central location or cloud) as this simply means an always-secure quicker-resolution – consider an autonomous car that does a lot of processing on-board. Having a software-based security that resides in it will be the need of the hour than a central cloud-based security SaaS that connects intermittently. This of course will be subject to the device’s capability to handle extra load. The legal requirements related to security may change as well and it will be interesting to see how GDPR or local privacy requirements impact the dynamics of this segment.
CIR believes security, in fact, will be a very interesting segment to watch based on the various data types needing to be handled. For example, autonomous cars that perform on-board calculations and connect intermittently with the central server will have a low security risk whereas IIoT devices at mission critical installations that store and calculate data at local level but maintain high levels of cloud connectivity will present significant security risks.