The upstart was Connexin – a regional wireless network infrastructure firm so small that it was exempt from publishing its full financial accounts last year. Yet it stepped up to face the big four mobile operators in an auction where they spent up to £500m each on rights over airwaves needed to carry 5G phone signals as the mobile broadband technology is rolled out for 2020. Connexin had just £5m in the bank as it prepared to bid, according to accounts it published after the auction.

Its participation was said to demonstrate how 5G would break the grip operators had over the infrastructure they used to run their mobile phone services. 5G was going to make such a fundamental change to their infrastructure that government and industry expected disruptive businesses like Connexin’s would step in to run it.

With just days to go in March, CEO Furqan Alamgir had made a formal declaration of Connexin’s unfaltering resolve to take on the big four. Then it quit the auction at the 11th hour, without offering any explanation.

A source close to the firm said it had realised the economic assumptions of the 5G roll-out were flawed. 5G was not going to disrupt operators’ power the way everyone thought it would.

The UK’s £1.4bn auction sold airwaves at a higher frequency than usual – necessary to carry data at the high capacity required by mobile broadband. Conventional mobile signals use lower frequencies that travel over an infrastructure of ‘macro cells’ with range measured in kilometres. But with 5G signals travelling less well, and its data demands being greater, industry planned to deliver them from a dense network infrastructure of ‘small cells’, with range measured in hundreds of metres.

Experts warned it would cost too much money for each of the four operators to roll out their own small-cell networks. Instead, they would rent access from independent firms like Connexin. Yet the costs were so great that it put off Connexin too.

Those same doubts have led city authorities in Hull, which contracted Connexin in late 2016 to roll out a public network of wireless broadband and 5G small cells, to question its own assumptions.

Ian Anderson, author of Hull’s Digital Smart City Strategy, said the city wanted to make sure it had the infrastructure to support 5G, so it could compete with other regions for high-tech businesses and skills. But there was doubt over the extent to which small cells would be necessary to achieve that.

Hull is considering a Connexin proposal to abandon its small-cell plans, and to rely instead on a conventional, high-capacity wireless network to offload traffic from the main 5G mobile phone infrastructure in busy city centre areas where data demands were too high for it to cope.

“If you put two and two together, you would begin to understand why having spectrum was not necessarily needed,” said the source, who had been involved in making arrangements for Connexin to roll out its network in Hull and prepare its plan to buy 5G radio spectrum.

Alamgir said Connexin had concluded that small-cell infrastructure was less economic than wireless for carrying 5G mobile signals on busy streets. But he refused to say why the firm quit the auction. “There are commercial sensitivities around that,” he said. “I can’t share with you why. We decided to withdraw for commercial reasons. There were investors who we have non-​disclosure agreements with, who were backing us.”

Connexin usually builds Wi-Fi networks across city centres and rents access to private firms. US networking giant Cisco gave it a boost in September 2017, six months before the spectrum auction, with a £10m investment for building wireless network infrastructure, which Cisco needed to support a ‘smart city’ platform it has been touting to UK city authorities. Cisco and Connexin have since worked together on smart-city pilots in Hull and Newcastle.

Alamgir said Connexin might still consider using small cells. But its strategy preferred Wi-Fi for offloading mobile traffic, in conjunction with low-power LoRa wireless networks needed to carry data for smart city systems.

Arvind Satyam, global managing director for smart cities at Cisco, said a typical smart-city system would need a high-bandwidth Wi-Fi network to support high-definition CCTV video, and a low-power LoRa network to support data transmissions from sensors the city used to measure such things as traffic, pollution and light.

Edward Oughton, a senior research associate at the University of Oxford who has studied small-cell deployment, said that having more than one network would be “ludicrously expensive” in most places, and “casts a huge doubt over whether widescale small-cell roll-out will take place”.

One ‘outsider’ company did compete in the 5G auction against the big four mobile operators (BT/EE, Telefónica O2, Hutchison 3 and Vofafone). That was Airspan Networks, which had built small-cell networks in other countries, with backing from Japanese investment giant Softbank. It budgeted £10m to get a 20MHz-wide slice of the 190MHz of spectrum up for sale. But at auction close, a 20MHz share had cost £150m.

Airspan CEO Paul Senior told E&T it had planned to buy 5G spectrum to run a network of one million small cells it would then rent to the big four operators.

“The economics of someone putting up small cell are really bad unless they do it at scale,” he said. “Even then, if [you’re] the only one using that small cell, the business case is very marginal. Hence it doesn’t happen.”

Papers Oughton has published on the economics of 5G roll-out suggest that if operators sustain the alliances they formed to build macro infrastructure for earlier generations of mobile services, and fill those with small cells in highly populated areas, it will be possible to sustain two competing national networks.

Connexin boosted its spectrum prospects in 2013 by appointing to its board Professor Edwin Candy, an industry expert who had built 3G networks in nine countries while working as technology director for Hutchison 3G in the 2000s as well as a government advisor and a technical pioneer of mobile network technology. But Candy was involved in a tragic accident in 2016 and resigned his Connexin directorship when Cisco made its investment last year. Cisco subsequently won £3.4m government funding to build a 5G proof-of-concept network for rural communities in a consortium it formed with with Strathclyde University, where Candy holds his professorship.

The upstart was Connexin – a regional wireless network infrastructure firm so small that it was exempt from publishing its full financial accounts last year. Yet it stepped up to face the big four mobile operators in an auction where they spent up to £500m each on rights over airwaves needed to carry 5G phone signals as the mobile broadband technology is rolled out for 2020. Connexin had just £5m in the bank as it prepared to bid, according to accounts it published after the auction.

Its participation was said to demonstrate how 5G would break the grip operators had over the infrastructure they used to run their mobile phone services. 5G was going to make such a fundamental change to their infrastructure that government and industry expected disruptive businesses like Connexin’s would step in to run it.

With just days to go in March, CEO Furqan Alamgir had made a formal declaration of Connexin’s unfaltering resolve to take on the big four. Then it quit the auction at the 11th hour, without offering any explanation.

A source close to the firm said it had realised the economic assumptions of the 5G roll-out were flawed. 5G was not going to disrupt operators’ power the way everyone thought it would.

The UK’s £1.4bn auction sold airwaves at a higher frequency than usual – necessary to carry data at the high capacity required by mobile broadband. Conventional mobile signals use lower frequencies that travel over an infrastructure of ‘macro cells’ with range measured in kilometres. But with 5G signals travelling less well, and its data demands being greater, industry planned to deliver them from a dense network infrastructure of ‘small cells’, with range measured in hundreds of metres.

Experts warned it would cost too much money for each of the four operators to roll out their own small-cell networks. Instead, they would rent access from independent firms like Connexin. Yet the costs were so great that it put off Connexin too.

Those same doubts have led city authorities in Hull, which contracted Connexin in late 2016 to roll out a public network of wireless broadband and 5G small cells, to question its own assumptions.

Ian Anderson, author of Hull’s Digital Smart City Strategy, said the city wanted to make sure it had the infrastructure to support 5G, so it could compete with other regions for high-tech businesses and skills. But there was doubt over the extent to which small cells would be necessary to achieve that.

Hull is considering a Connexin proposal to abandon its small-cell plans, and to rely instead on a conventional, high-capacity wireless network to offload traffic from the main 5G mobile phone infrastructure in busy city centre areas where data demands were too high for it to cope.

“If you put two and two together, you would begin to understand why having spectrum was not necessarily needed,” said the source, who had been involved in making arrangements for Connexin to roll out its network in Hull and prepare its plan to buy 5G radio spectrum.

Alamgir said Connexin had concluded that small-cell infrastructure was less economic than wireless for carrying 5G mobile signals on busy streets. But he refused to say why the firm quit the auction. “There are commercial sensitivities around that,” he said. “I can’t share with you why. We decided to withdraw for commercial reasons. There were investors who we have non-​disclosure agreements with, who were backing us.”

Connexin usually builds Wi-Fi networks across city centres and rents access to private firms. US networking giant Cisco gave it a boost in September 2017, six months before the spectrum auction, with a £10m investment for building wireless network infrastructure, which Cisco needed to support a ‘smart city’ platform it has been touting to UK city authorities. Cisco and Connexin have since worked together on smart-city pilots in Hull and Newcastle.

Alamgir said Connexin might still consider using small cells. But its strategy preferred Wi-Fi for offloading mobile traffic, in conjunction with low-power LoRa wireless networks needed to carry data for smart city systems.

Arvind Satyam, global managing director for smart cities at Cisco, said a typical smart-city system would need a high-bandwidth Wi-Fi network to support high-definition CCTV video, and a low-power LoRa network to support data transmissions from sensors the city used to measure such things as traffic, pollution and light.

Edward Oughton, a senior research associate at the University of Oxford who has studied small-cell deployment, said that having more than one network would be “ludicrously expensive” in most places, and “casts a huge doubt over whether widescale small-cell roll-out will take place”.

One ‘outsider’ company did compete in the 5G auction against the big four mobile operators (BT/EE, Telefónica O2, Hutchison 3 and Vofafone). That was Airspan Networks, which had built small-cell networks in other countries, with backing from Japanese investment giant Softbank. It budgeted £10m to get a 20MHz-wide slice of the 190MHz of spectrum up for sale. But at auction close, a 20MHz share had cost £150m.

Airspan CEO Paul Senior told E&T it had planned to buy 5G spectrum to run a network of one million small cells it would then rent to the big four operators.

“The economics of someone putting up small cell are really bad unless they do it at scale,” he said. “Even then, if [you’re] the only one using that small cell, the business case is very marginal. Hence it doesn’t happen.”

Papers Oughton has published on the economics of 5G roll-out suggest that if operators sustain the alliances they formed to build macro infrastructure for earlier generations of mobile services, and fill those with small cells in highly populated areas, it will be possible to sustain two competing national networks.

Connexin boosted its spectrum prospects in 2013 by appointing to its board Professor Edwin Candy, an industry expert who had built 3G networks in nine countries while working as technology director for Hutchison 3G in the 2000s as well as a government advisor and a technical pioneer of mobile network technology. But Candy was involved in a tragic accident in 2016 and resigned his Connexin directorship when Cisco made its investment last year. Cisco subsequently won £3.4m government funding to build a 5G proof-of-concept network for rural communities in a consortium it formed with with Strathclyde University, where Candy holds his professorship.