Can we really afford the metaverse?
by Pete Hill, VP Corporate Development, Cudo Ventures
Metaverses are not easy to build, and it is worth remembering that it is unlikely that there will be one single all-encompassing metaverse. Much more likely, is that there will be myriad different metaverses supporting many diverse communities.
The vision we all have is for a fully immersive, 3D rendered, photorealistic digital realm offering us real-time interactions, supporting a vast number of users across many geographies. But that is where we are now arriving at a bit of a problem. We are a long way away from that vision today and this poses some hard questions.
Intel has already forecast that the immersive internet vision of the metaverse will require 1,000 times more compute power than we currently have available. Meta estimates that 10-20 milliseconds latency will be needed in networks to support real time interactions and Nvidia forecasts that by 2028, 50% of the internet will be powered by real-time 3D software. All of this is going to require a massive shift in how we think about infrastructure and data and most importantly how we fund this kind of development.
Traditionally the approach has been fairly simplistic – build more datacentres and lay more fibre. But, as we have seen with crypto mining, powerful compute capabilities are energy hungry. For example, before its recent ‘merge’ Ethereum combined with Bitcoin was consuming more energy than the entire country of Thailand. If metaverses aren’t going to consume even more energy than crypto mining then the industry is going to have to think differently.
Energy costs have always been one of the biggest parts of a datacentre’s operational budget, often representing around 20% of total running costs. With energy prices now at an all-time high the underlying economic model of datacentres has been fundamentally changed and needs to be revisited.
Critically, the shift in core blockchain validation approaches, moving away from proof-of-work to proof-of-stake, will make a huge difference in the energy demands of metaverse builders. We have already seen some fairly impressive claims from Ethereum.
Another factor that needs to be considered in this process is the supply chain issues in regards to access to the high-performance graphics chips that metaverse builders as going to need. Over the past few years there has been a global shortage of these chipsets and prices have skyrocketed. Whilst the peak of that shortage has reduced there is still an exceptionally high demand for these chips and metaverse builders will be scrambling for access to the compute power they deliver. For without them, the vision of an immersive 3D internet is going to remain tantalisingly out of reach.
So, what can the industry do to rebalance the disrupted economic equations underpinning the future of the metaverse?
First of all, the industry has to become better at using the existing capabilities within the global datacentre marketplace. We currently have a situation where utilisation in some data centres is standing at only 60%. If that spare compute power could be accessed and offered to users in a flexible and secure manner then metaverse builders could be starting to create those new digital worlds much quicker.
By creating a decentralised cloud compute capability, the question of latency within the network can also be addressed. Moving away from large hyperscale providers with datacentres concentrated in a small number of locations will open up metaverse development to many more players and establish a new, fairer economic model.
Metaverses are going to change how we interact with each other, companies, brands and other organisations. We have to ensure that we can do this in a measured, economically viable and sustainable manner. Otherwise, the current combination of high energy costs, chip shortages and an over reliance on centralised infrastructure is going to stifle and compromise our shared vision of a new digital future before it gets a chance to become a reality.