The Strange Case of the Hole in the 5G Core and the Missing Startups

Last week saw the Mobile World Congress event in Los Angeles – an industry jamboree to celebrate all things telecom and mobile. This year, however, behind the public announcements and excitement around 5G was a more private level of concern and even despondency laid bare in the smaller sessions away from the keynotes.

At issue is the dire state of startups in the telecom world, or more specifically startup funding. Despite last year holding the record for the largest amount of venture capital deployed, hardly any is making its way to companies that build their technology and vision around telco infrastructure. The evidence for this was writ large on the show floor where, away from the glitz of the small number of major vendors and telcos, many of the exhibitors were dedicated to cabling, connectors, recycling, or phone accessories.

This is not a new problem – VCs are very wary of the very long sales cycles and the conservative purchasing mentality of telcos. They are largely staying on the sidelines of anything to do with 5G infrastructure, preferring to place their dollars on cloud, content, and application plays that they hope will one day ride on top of this new network. Even the investment teams within telcos tend not to bet on infrastructure startups but run with the mainstream VC crowd on the next social media or business-to-business application they hope will ride on top of this new network. According to CrunchBase, the last few years have seen an average of 12 to 15 5G-focused companies per year receive funding. When you dig a little deeper, however, it turns out that few of these are actually focused on 5G infrastructure or core.

This lack of new innovation, new thought, and market speed is starting to have a ripple effect across the industry. From an investment perspective, the move to 5G is at best a zero-sum game if all we are doing is gaining a faster network but with the same number of subscribers and the same revenue-per-subscriber. The whole purpose of 5G is to embrace a level of ubiquitous connectivity for all kinds of devices from cars to washing machines to wind farms, with all the new network services this entails. In this regard, as an industry, we are off by a country mile. So, if the business case for 5G is adrift because the infrastructure technology isn’t ready, we certainly have a rocky few years ahead of us.

To understand why this is the case, you must look to the current vendor eco-system. In many ways, it resembles the mini-computer market of the mid-nineties. The behemoth vendors that dominate the industry today have built their fortresses around the conservative mindset of their customers. 5G represents a renewed opportunity for these vendors to lock customers into their own monoculture, in the same way IBM, DEC, and HP were able to do so by owning as many constituent parts as possible. Telcos, somewhat resistant to this lock-in choose dual-vendor strategies, choose different vendors for different parts of their network or end up with multiple vendors anyway through acquisition and mergers. This has created a tangle of systems that are very hard to orchestrate and manage, and this is why it can still take months to provision connectivity between two cities or order up new internet service, even when all the equipment is in place. 5G now multiplies the number of radio locations and associated equipment by a factor of ten, moving the management problem from “challenging” to “untenable”.

This challenge has not gone unnoticed by the telcos who recognized the need to move toward a much more open ecosystem and have been pushing vendors for at least seven years to move their networking products to generalized compute, in an initiative which they led, known as network-function virtualization or NFV. Vendorslike the dog that knows it’s going to the vethave been led unwillingly down this path, pricing their x86 software versions of product at the same price or higher than their proprietary equipment for the same performance. Harder for the vendors is the fact that moving their technology to x86 is difficult, much in the same way that moving mini-computer applications to the web never really happened. For these reasons, Gartner last week went as far as saying that NFV is dead. In the computer world, for every legacy mini-computer application that died, a web application company was born, because the investment community recognized the new paradigm was real. In telco there has been a deafening silence from startups, despite the best efforts of Intel to encourage such innovation on their platforms. Meanwhile, startups touting SD-WAN solutions, products that typically sell to enterprises and ride on top of the telco network, have received hundreds of millions of dollars in funding.

In parallel, the last five years have seen the green shoots of open source software make its way into telco land, with projects such as OpenDaylight (which my company Lumina Networks champions), OpenStack, and ONAP. The job of open source in telco is to unify the many disparate vendors’ products so they can be managed and orchestrated together and is the only practical way we’ll get to deliver end-to-end 5G services for the myriad of new use cases. As an open source networking vendor, Lumina has been extremely fortunate, because we existed for three years as a business unit inside Brocade before we were spun out and then received early backing from Verizon and AT&T as both customers and investors. Nevertheless, I can’t say our journey has been easy, even though I don’t see an army of well-funded startup competitors coming after us. What worries me most is that telcos will revert to form and push the burden of 5G on to their existing vendors, not because they believe they will deliver, but because they can be blamed when they don’t. After all, no one ever got fired for buying IBM (until they did).

So how does this get fixed? Well, if the telcos have created this problem for themselves, they really need to be part of the resolution. In obtaining investment dollars from telcos today, startups must pass the “will you be the next unicorn?” test; the “strategic value to my business” test; and the “you must have existing significant business with us” test. Threading a startup through this needle, to end up with a $2m check does not lead to a healthy eco-system.

Remember too, that as an industry we now have a large technology debt, born out of years of under-investment in startups, which will likely lead to a scramble to make 5G services real once all the radio is in place. Telcos really need to place significant bets on the startups that can affect this outcome now and seize control of their own destiny. With this, traditional VCs will follow when they see telco investments as not a side bet, but a stake in their future network. The telco balance sheets of 2022 will surely tell the story of telco success when the 5G infrastructure bills become due.