- Public cloud providers are retiring their proprietary private 5G offerings
- Wireless industry executives say private 5G has a future helping enterprises transport data for AI
- On its own, private 5G doesn’t usually offer a positive return, so vendors are helping companies deploy private 5G in cost-effective use cases
Private 5G is not attracting enterprise customers the way vendors hoped it would. AWS recently ended support for its private 5G service, and Microsoft announced that it will retire its private 5G core service in September.
These cloud providers will still support private 5G through telco partnerships, and those partners are saying the technology has a future.
At a J.P. Morgan conference earlier this month, Verizon CEO Hans Vestberg said demand for private 5G is finally improving.
On the same day Vestberg was speaking to investors, the Wireless Infrastructure Association was wrapping up its ConnectX conference in Chicago. Several industry veterans there described private 5G as a difficult technology to market unless it can solve a specific problem for a business.
“IT directors and CIOs and the like are not shopping for a private wireless network,” said Ken Sandfeld, CEO of open RAN private 5G vendor Revells and a panelist at ConnectX. “They have problems that they’re not able to solve with the current technologies that they have in their arsenal.”
Sometimes, private 5G is the solution. Private 5G may be able to help businesses solve problems with factory automation and AI inference. For manufacturers that want to use robots in factories, Wi-Fi can disappoint if access points interfere with one another, and cellular radios with longer range may be the answer. For businesses of all types that want to use AI models to gain insights from the data they generate, private 5G can help get data to and from the cloud quickly and securely.
All about the data
Dan Hays, principal at PwC Strategy&, said his team typically talks with customers about how they can leverage data insights instead of investing in private 5G. “When we do these types of connected solutions, we start with what can you do with the data, and then we back into what’s the right connectivity solutions, what’s the right processing solution, etc.,” he said.
Hays moderated a ConnectX panel on the evolution of the mobile edge, and he asked his panelists how private 5G fits into solutions for processing and analyzing data at the network edge. Panelists reported that it isn’t always easy to connect the dots for customers. Some use cases that can be solved by a private 5G network paired with an edge appliance can also be solved by network slicing, said Arvin Singh, former vice president of global innovation and solutions at Verizon.
Nadine Manjaro, assistant vice president at integrator Sutherland Global, said enterprise customers often don’t want to invest time in understanding the differences between network slicing and private 5G. “The operators sell private 5G separately, they sell network slicing separately, they sell IoT separately, and then you have the data connectivity,” she said. “If you are a large enterprise, you know your core business, but all of these components might sound crazy. Even though ideally you want to get to AI, you want to get to inferencing, how do you put that infrastructure together? It is very chaotic.”
Considering the cost
The chaos can be worth it if networking solutions lead to desired outcomes for the business. Panelist Awaiz Khan, principal solutions architect at AWS, said the projects that “stick” are those with positive returns. Sometimes customers don’t realize there’s no ROI until they’ve seen a proof of concept. “Typically, what happens is we go into all this complexity in building the system out, but when the customer finds out that there is no return on investment it basically tends to fail,” Khan said.
Hays agreed. “The question of private networks is almost always followed by the question of ‘How much is it going to cost me? What do I get for it?’” he said.
The cost can be a stumbling block, because in addition to the upfront capital investment, businesses may need to pay a partner to manage the network going forward. Manish Mangal, head of America's Communications Business for Tech Mahindra, spoke with Fierce about this challenge.
“What we found was that the return on investment for the private 5G network was not as clear to the enterprises, because first of all they do not have cellular expertise,” he said. “For them it was a little bit of a scary moment. They know that with Wi-Fi they can just plug and play their access points, and they are good to go.”
More brains, more bandwidth
AI is already a bit intimidating for some IT departments, and adding the complexity of private 5G to an AI initiative can increase the uncertainty. But it can also increase the chance of success. This is what network operators are banking on.
Verizon highlighted the synergy between private 5G and AI recently in an announcement with Nvidia. In a press release, Srini Kalapala, senior vice president of technology and product development at Verizon, said “We’re leveraging our network’s unique strengths including private networks and Verizon’s global industry leadership in private MEC, combined with Nvidia’s AI compute capabilities to enable real-time AI applications that require security, ultra-low latency, and high bandwidth.” These are areas where private 5G usually surpasses Wi-Fi.
The question is whether large numbers of enterprise buyers will want to pay for private 5G to enable real-time AI, and the answers will likely depend on their projected returns from those AI applications.