I have been covering spatial computing as an analyst since 2014, when I joined Moor Insights & Strategy, and even a year before that, in 2013, as a journalist. One of the companies that has always interested me in the space is Amazon Web Services (AWS), which is the world’s largest provider of cloud services and technologies. I have long been curious about its role within the broader spatial computing ecosystem, especially considering how much the metaverse and spatial computing depend on the cloud. After all, Microsoft’s entire spatial computing effort, including the HoloLens project, primarily exists to drive more utilization of its Azure cloud services, so it makes sense that Microsoft's biggest competitor in cloud computing, Amazon, would be looking to do the same.
Amazon’s approach to spatial computing
One big component of what Amazon has to offer the metaverse and spatial computing, whether via AR or VR or not in a headset at all, is that the company genuinely seems to be always looking for ways to enable customers to be successful. That focus is at the core of AWS’ message, which is why I hear Amazon’s ‘customer obsessed’ mantra repeated endlessly every time I attend an AWS event or talk with AWS leaders.
This focus extends to the company’s partnerships in the industry. AWS already publicly supports a plethora of XR companies big and small, including Campfire 3D, Meta, Magic Leap, Nvidia, and many others. AWS’ Spatial Computing unit incorporates multiple divisions, including gaming, film, simulation, and geospatial, which makes sense when you consider that all of these mediums are inherently spatial. Many, but not all, will use a game engine to render their worlds. AWS has also heavily supported open standards such as OpenXR and even turned its game engine, Lumberyard, into the Open 3D Engine in partnership with the Linux Foundation to create an open-source game engine that it regularly updates.
AWS has also positioned itself as the premier cloud provider for spatial computing by deploying not only dozens of cloud zones around the world, but also local zones and its Wavelength edge-computing infrastructure to help lower latency even further. Because so many spatial computing applications are so latency-sensitive, it makes sense that AWS wants to get as close to the user as possible. This enables better remote rendering, giving lightweight AR and VR headsets a boost to rendering quality or performance without increasing a headset's weight. Without ample connectivity and ubiquitous edge computing, AWS's developers would not be able to depend on this capability to build applications that use edge cloud resources.
AWS also had a 3-D engine called Sumerian, which it built in 2017 but has since wound down, giving the rationale that XR market growth hasn't been what many expected. I believe Sumerian didn't always match developers' workflows; Amazon has been transitioning users to Bablyon.js for authoring and publishing using AWS Amplify. Amazon still does a great job of supporting the two main engines— Unreal Engine and Unity—that most spatial applications are built on.
The approach I see from AWS today is building things in collaboration with customers as the customers need them. When a project becomes compelling enough for other customers, AWS will build it out as a product that helps reduce time-to-market and simplify the complexity of spatial computing. That leads us to AWS’ most significant spatial announcement at re:Invent 2022: AWS SimSpace Weaver.
SimSpace Weaver was announced just a few weeks ago at the AWS re:Invent conferenceby AWS CEO Adam Selipsky during his opening keynote. Selipsky led by talking about how AWS enables many of its customers to power their simulations, including one of the largest companies in the world, Siemens. Amazon sees simulations as a form of high-performance computing (HPC) that helps to anticipate a problem before it happens. Spatial simulations take things to an entirely new level by incorporating millions of real-time interactions among people, machines, and the environment. Spatial simulations are now used to simulate even entire cities; we’ve already seen that with Ericsson’s use of spatial simulations to render an entire city's 5G coverage using Nvidia Omniverse. A simulation of this complexity and accuracy is also known as a digital twin.
Adam says that the 3-D engines that can simulate these real-time interactions were never designed to scale to the size of a multi-million-person city covering hundreds of square miles. To achieve this, AWS has needed to create solutions in-house that seamlessly interconnect multiple instances of these simulations to scale to the size of a major city.
That requirement led AWS to build SimSpace Weaver, which allows for running massive spatial simulations without needing to manage the infrastructure. SimSpace Weaver takes a spatial simulation and automatically scales it across up to 10 Amazon Elastic Compute Cloud (EC2) instances. It also enables objects to move seamlessly from one instance to another, which makes simulating these massive worlds more natural. AWS already supports Unreal Engine and Unity and offers a SimSpace Weaver SDK for custom simulation engines. AWS’ focus with SimSpace Weaver is to remove complexity while building these massive simulations, which gives that time back to developers so they can work on the simulations themselves.
Clearly, AWS has a lot going on within the metaverse and spatial computing arenas. While many of the talks at re:Invent centered on game development and how to use AWS to drive social interactions and large-scale gaming environments, many of those same capabilities translate directly to spatial computing. Yet because AWS is a critical infrastructure provider for so many companies, it cannot disclose the names of its customers unless they publicly mention AWS.
This creates a challenge as we try to fully assess AWS' place in the XR industry. We know that its cloud computing powers many different virtual spaces that users now enjoy, including games, VR experiences, and more. But in many cases, consumers are entirely unaware of AWS' role behind the scenes—which presents its own challenge for AWS in spreading the word about its work.