Mavenir says it has dropped its ambitions of becoming a maker of Open RAN Radio Units as it moves to stabilise its finances and project itself as a stable, ongoing entity to customers of its core network, IMS, vRAN baseband and messaging software.
The company has secured a debt refinancing which has eliminated its existing $1.3 billion debt, and given it a new $300 million in private debt financing, which The Mobile Network understands does not fall due for several years.
CEO Pardeep Kohli said that some of the company’s lenders have agreed to convert their debt into equity – with others receiving payments. The equity swap means that existing shareholders such as Siris and Koch Strategic Platforms – which invested $500 million into Mavenir in 2021 – have had to accept a dilution of their equity. Mavenir’s existing investor, private equity company Siris, whose executive partner Hubert de Pesquidoux is Mavenir’s Executive Chair, made its backing for the deal public.
“At the end of the day, the net-net situation is we have a $300 million of debt versus $1.3 billion, and the company has more cash to go forward as well,” Kohli said.
RU business shifts away from hardware
The company said it would be concentrating solely on its software business from now on, which means that its ambition to design, build and sell Open RAN Radio Units is now over. Instead, it will serve to licence designs to ODMs who will be responsible for buying and holding the inventory required to build the units. It could also licence directly to a mobile operator who would then contract with an ODM to build the design. That releases Mavenir from the working capital demands of running a hardware RU business.
“In the hardware space you need you need volume, you need working capital and you need inventory – and our DNA was as a software company. We never actually did hardware except for this [RU] purpose. So we have moved back to our software origins, and now we’ll have baseband (CU-DU) software and we’ll have radio software, but we’ll work with ODMs to actually get our software working on the ODM hardware.”
Kohli also pointed out that Mavenir never directly manufactured its Open Beam RUs, although it was responsible for designing, sourcing and buying the hardware.
Kohli describes a situation in which a start-up RU developer like Mavenir is just unable to compete with the incumbents, which are able to hold inventory to meet demand the wide variety of frequency bands, power output, sizes and weights and MIMO designs that are required by a diverse global market.
“Ericsson, Nokia can carry billions worth of inventory quarter to quarter. Which other startup is going to be able to do that? Of course, if the government helps then it’s different. For example in India the government gave Tejas a $900 million credit line.”
Like NEC, Parallel Wireless and Fujitsu, Mavenir was tempted to enter the RU business to bring more diversity to the RU landscape as the major incumbents – especially Ericsson – were sluggish to open up their own RUs. The lack of Open RU options was in turn was putting something of a brake on Open RAN deployments as operators feared a lack of competitive options in the most expensive part of the network – the radios. Why move the RAN baseband to an Open RAN vendor if your installed base of RUs remained closed, or if the replacement Open RUs were not as good?
Encourage by this perceived demand, Mavenir developed its Open Beam product line for which it received positive feedback on the quality of its products and designs, many of which were highly energy efficient and featured advanced techniques in power management and cooling.
But while many operators said that they wanted to encourage vendor diversity and use multi-vendor sourcing in the RAN to secure best of breed technology and better pricing, that never materialised at scale as operators were either already well invested in 5G rollouts, or chose to stick with existing vendors when refresh tenders did materialise. When Ericsson pushed through a change in Open Fronthaul specifications that meant its existing massive MIMO RUs could connect to an O-RAN baseband, operators ran for the security and economic sense that offered. Now they could keep their Ericsson network and know that, if needed, they could plug in a different RU at some point down the road. AT&T was pivotal in adopting this strategy, but far from alone.
So instead of kick-starting demand and giving operators peace of mind that an Open RAN baseband could have good RU options with which to interface, Mavenir faced a situation in which it could not finance the working capital to bridge the gap between the investment required to develop a working RU product line at scale, and being paid back by operators. This was despite positive feedback on the quality of its products and designs, many of which were highly energy efficient and featured advanced techniques in power management and cooling, for example. Nor was the company alone in finding the going tough. NEC scaled back its own ambitions, early market-maker Parallel Wireless had to re-focus and has re-emerged with a platform it says can run on any hardware. Fujitsu has perhaps made the most headway, bouyed in part by demand in its home markets. Meanwhile Rakuten Symphony moved its vRAN baseband business to a licensing model,
Kohli said the company needed to assess demand forecasts and buy components a year ahead of a contract delivery date, for which it might not be paid until six months after delivery. An ODM would be better positioned to handle inventory and work with suppliers, he added.
He added, however, that the company would continue to be able to meet existing customer requirements, and the company is keen to stress that this news does not affect any current customer agreements or ongoing deployments. For instance, it has announced a contract to supply micro RUs to AT&T as part of that vendor’s process of upgrading and extending its network densification.
“With AT&T it is at the early stage and we can continue to work with them, we can work the ODM model with them as well. The issue of working capital is not a US problem. For AT&T I can find bank money, and the payment terms are good. The problem comes when you go to Europe, India and other places where payment terms are six to nine months, the PO doesn’t mean anything and nobody’s willing to finance. In the USA that’s not the case – if AT&T gives me a PO, Bank of America gives me a loan.”
Core ambitions
As it moves forward, Mavenir will be relying on the messaging, IMS and core network software areas which already contribute 80-90% of its revenues. It will also continue to invest in its radio baseband software, Kohli said.
As it attacks the 5G core network market I particular, it all be critical for Mavenir to portray itself as a solid, stable financial entity. MNOs will be wary of handing over their core networks to a company that looks to be on shaky financial ground. Mavenir also faces a headwind here as operators have tended to be conservative in their vendor choices for core network software – with the majority of the 5G Core announcements in non-Huawei markets so far going to the incumbents Ericsson and Nokia.
But Kohli is positive about Mavenir’s position here. “We’ve done quite well. We have awards at 80 customers, two years ago that was 35-40.”
Customers include Deutsche Telekom in Germany and other operating countries. “They had Ericsson for 4G, Huawei for 3G, and they picked us for 5G,” Kohli said. “So they are slowly adding more and more subscribers and scaling it up. Most of the new stuff, such as the network slicing announcements they have done, that’s all using our software.” The company also supplies all of Open RAN operator 1&1’s core. And Kohli named an additional European T1 that has not been announced publicly.
“Anyone who wants to go in the public cloud, it’s really Nokia, Ericsson and us. Since Affirmed (Microsoft) got out of it, we are winning a lot of former Affirmed customers – and in addition we have replaced Nokia and Ericsson in many places.”
The 5G cloud core market has seen the field for early challengers narrow down, Casa Systems has exited, while Microsoft Azure bought and then pulled the shutters down on Affirmed Networks. HPE still sells some aspects of core network functions, but never fully realised its pre-covid ambitions to exploit 5G’s Service Based Architecture. However there do remain other competitors in parts of the core. Oracle continues to target the cloud-native core opportunity, after winning deals with Boost Mobile back when it was still called Dish back in 2022. Enea is another with established products in policy and data management, making this market slightly more nuanced than the three player face-off that Kohli describes.
Another story that Mavenir is keen to tell is its integration of AI technology across its network software stack. Prior to MWC2025, Kohli explained to TMN how the company is adding value to its cloud-native software with AI capabilities.
“We have everything we need to run on cloud,” Kohli says, “What we are working on now is to make use of all the AI tools that are becoming available, and evolve the network so that it can align with the advancements which are happening in that space.”
Brandon Larson, SVP Cloud Platform & AI, described the process as getting AI into the “operational bloodstream” of its customers.
Importantly for Larson, this is always done in conjunction with telco domain experts.
“We built our own Deep Q-Network (DQN) reinforcement learning models, using neural networks to model system behaviour. But these are not generic – we created our own specific architecture, understood data collection and context and engineered solutions for specific telco use cases.
“We’re looking at problem statements, whether it be in fraud and security, RAN operations, or capacity modelling for IMS, and then we work our way back to the AI, with telco domain experts working with the data science to solve a very specific problem. That’s a very practical way to solve this problem within telco constraints, so we’re not just starting by throwing GPUs and horsepower at it.”
At MWC2025 TMN hosted a discussion on how Mavenir is developing AI capabilities in operator networks which throws further light on its AI strategy. (Disclosure: this was a discussion sponsored by Mavenir.)
Mavenir has taken the pragmatic route… The question is whether core network revenues and the limited Open RAN new opportunities will sustain its business in the next few years
An up and down story
In the pre-covid times, Mavenir’s headcount was around 2,000. By around two years ago, that had swelled to 6,000. Now it has already reduced headcount to something around 4,000, Kohli said. Posts on LinkedIn last year suggested that there has already been a major exodus from the company’s radio engineering division in India, and Kohli confirmed that everyone on the RU hardware side had already left, swapped roles, or was in the process of doing so.
As Mavenir built its debt, there were queries towards the back end of 2024 about the company’s liquidity as ratings company S&P doubted the company’s ability to pay back $133 million of its debt, which was maturing on Jan. 31, 2025. Shortly after that assessment, Reuters got hold of a story that Saudi company Aramco Digital, at that time led by former Rakuten Mobile Open RAN advocate Tareq Amin, was considering a billion dollar investment in the company. That investment never materialised, something that Kohli puts down to a change of government relations.
“It was a sovereign wealth fund, and we never actually confirmed which one. It was largely a government to government transaction, and with the election in the USA there was a change in relations and motivation. These things happen, and it was out of our control.”
Now the company holds $300 million of private debt, which will sit outside of the ratings agencies’ sight. Its investors have accepted a dilution of the business at an unknown valuation, clearly considering it is better to accept the prospect of a reduced share of future growth than suffer the consequences of a potential default. While the company is solid in the messaging software and IMS space, its growth in 5G will be a hard road, and faces some of the same risk dynamics as the Open RAN sector. It has also thrown its vRAN hat into the ring in the NTN space, although here too it will face a competitive market.
Dimitris Mavrakis, Senior Research Director, ABI Research, said, “Mavenir has taken the pragmatic route to develop Open RAN and only licensing its technology to partners while maintaining its expertise to support its existing contracts, most importantly AT&T. The question is whether core network revenues and the limited Open RAN new opportunities will sustain its business in the next few years.”
Amol Patel
Great article covering technical and business details in great details…