TowerXchange hosted another productive and insightful energy working group at the 5th annual TowerXchange Meetup Asia in Singapore. Decision makers from four of the region’s leading towercos, plus an innovative MNO, shared candid insights into their cell site energy experiences and requirements. Here is a summary from our minutes of the working group. Respecting Chatham House Rules, apart from a brief introduction to the participating buyers, subsequent comments have been anonymised.
Introducing the telecom energy buyers
edotco: Represented at the TowerXchange Meetup by over a dozen members of their leadership team, edotco is currently active in six countries (Bangladesh, Cambodia, Malaysia, Myanmar, Pakistan and Sri Lanka), with intent to launch in Laos and the Philippines. edotco currently owns over 18,000 towers and manages a further 10,000. 80% of edotco’s towers have outdoor configurations. Around 5% are off grid. edotco is a renowned innovator, and is conducting feasibility studies for renewable energy.
Bharti Infratel: Represented in the working group by their CTO, Bharti Infratel currently has over 40,000 towers. When the imminent merger with Indus Towers, in which Bharti Infratel owns a 42% stake, is complete, the combined entity will own over 164,000 towers across all Circles in India. With such a diverse footprint, power availability and load requirements vary considerably. Bharti Infratel’s historical focus has been purely on passive infrastructure, but they are increasingly involved in provision of active infrastructure, particularly within smart city initiatives. Bharti Infratel’s wish list? High return on investment solutions, which can be effectively maintained.
Ascend Telecom Infrastructure: Represented in the working group by their CEO, Ascend has 6,200 towers across India. Ascend are green energy pioneers. While energy represents 60%+ of MNO opex in India, energy efficiency will remain their number one priority. Ascend uses business analytics to gather data points that provide critical information enabling proactive interventions.
GTL Infrastructure: Represented in the working group by their VP Strategy, GTL Infrastructure owns 27,707 towers in India. They share Bharti Infratel and Ascend Telecom’s green tower objectives, and are seeking win-win risk sharing partnerships with vendors.
Indosat Ooredoo: Represented by their Division Head for Tower Commerce, Indosat Ooredoo introduced their portfolio of 30,000 sites in Indonesia, around half of which are owned, half co-located with towercos. Indosat Ooredoo’s primary divers are also cost-related, driven by falling ARPU and tariff wars, leading to downward pressure on opex, particularly cost of power. Power remains the MNOs’ responsibility in Indonesia, not the towercos’.
The continuing search for return on investment in renewables
“When will the savings justify widespread investment in renewables, particularly solar?” Asked one towerco. “While we quantify savings in diesel reductions, we remain wary when the calculation of ROI is not undertaken across the whole ecosystem: capex, maintenance opex, cost relative to grid,” the towerco continued.
“We’ve trialled renewable energy solutions where the return on investment the vendor modelled in Excel was not achieved in the field,” suggested a towerco.
“A number of factors, including actual opex, seasonal variations, and the impact of Service Level Agreements (SLAs) create legitimate deviations from models,” responded another towerco. “For us the feasibility and Total Cost of Ownership (TCO) of solar is about five factors: load, PV capacity, the batteries, fuel price, and optimising configuration.”
“No-one can justify investing in renewables on an environmental basis only, although reducing the carbon footprint is factored into our TCO calculation,” said another towerco. “What we need are smoother pathways to implementation – it’s the customisation costs that are most harming the TCO.”
Another towerco commented: “We appreciate the efforts, research and development being invested by solution providers. We’re not put off by the variance between modelled and delivered ROI as we appreciate that solutions need customising to the local environment. Our recommendation to vendors would be to spend less time on your spreadsheets and ROI models, and more time on proofs of concept generating live data, based on which you can continue to refine solutions.”
“We are willing to invest in renewables with the right return on investment on a TCO basis,” added another towerco. “The challenge used to be sizing the hybrid system, but we have a good sizing tool now. Ultimately, with hybridisation, batteries typically represent 40-50% of the capex, depending on the degree of desired autonomy. Sizing hybrid is all about energy storage capacity: if you size the solution too small, you burn more diesel, and achieve less autonomy and incur greater risk of downtime.”
“We also do our own design and dimensioning,” agreed a different towerco. “The modularity and scalability of the solution is key to controlling TCO in the long term, as well as the sheer amount of space in the stack.”
“In India, renewable energy vendors targeting cell sites are chasing a moving target,” said a towerco from that country. “As the grid is being extended, more and more off grid sites are being connected. Around 98% of India’s cell sites are now grid connected.”
The same towerco continued: “the telco load is also not a constant, as more tenants are added (or churn off sites), and as antennae are swapped. Peak and off-peak usage patterns are changing, yet MNOs don’t always dimension to reflect peak peak demand, which can lead to an upgrade requirement, even on grid.”
Why have ESCOs not taken off in Asia to the extent that they have in SSA?
One towerco’s simple response: “we have the necessary in-house knowledge.”
“Regulatory challenges remain a big inhibitor,” added another towerco.
TowerXchange research in late 2018 suggested ESCOs owned and operated the power systems on 6,414 cell sites in India – an impressive total, but a total which has not greatly increased in the last three years. While turbulence in the Indian telecom market offers one explanation, one of the towercos in the working group offered an alternate perspective.
“India’s initial ESCO contracts had room for improvement. They typically assumed a 1.5kW load, but when incremental load was added tenants demanded the same fees despite increasing costs, and the contracts did not always protect the ESCO from having to provide that. In that context, the economics often precluded investment in energy efficiency innovations, and as a result performance against SLAs suffered.”
Cost per kWh off grid
“In Southeast Asia, we’re paying around 56 cents per kWh off grid, inclusive of generation and capex,” said one towerco. “It’s probably much less in India.”
“Yes, we’re paying as little as 10 cents per kWh off grid on a plug and play basis,” added an Indian towerco.
“And we’re probably burning 300-400ML of diesel per kWh,” added another Indian towerco.
The relative merits of power pass through versus fixed energy cost models
A primer for readers that are newer to the tower industry. From an energy perspective, towerco business models fall into three categories:
- Pure ‘steel and grass’ towercos that only provide ‘vertical real estate’ – MNOs retain responsibility for primary and backup power
- Power-as-a-service towercos that take responsibility for power but pass through the cost of that power direct to the MNOs
- Power-as-a-service towercos that take responsibility for power and charge a fixed cost for that power to the MNOs, thus making a margin on energy efficiency gains
The working group discussed the relative merits, particularly of the latter two business models.
“Even with a pass through business model, the MNOs’ primary concern remains opex, and energy is the largest component thereof,” said one towerco.
“MNOs want uptime at the lowest possible cost,” said another towerco. “When power is a pass through, it can cause a misalignment between towerco investment and who it is that benefits from reducing energy costs. Hence MNOs increasingly favour power-as-a-service with a fixed energy cost. It’s a win-win: the MNO caps their rising energy costs, while the towerco can invest to create margin.”
“We operate a variant on the fixed energy model,” said a third towerco. “We deploy our own capex to improve energy efficiency, and we pass on 40% of savings to the tenants. Nonetheless, the MNO wanted more, despite us taking the risk, putting in the effort and incurring the cost!”
“Towercos are B2B business models, and with that comes an expectation of cost plus pricing,” countered another towerco. “We must stand behind and support our MNO partners’ performance.”
“We can earn bonuses for outperforming our uptime KPIs, and are hit by penalties for missing our SLAs,” said another towerco. “Yet energy vendors don’t pay the price for downtime, which makes us inclined to use only solutions that have stood the test of time.”
“Is there a win-win, gain sharing model to be explored between towercos, MNOs and vendors?” Asked one vendor.
“Vendors can offer an opex model, sharing gains and sharing risk on SLAs,” commented one towerco. “But most mature towercos have access to lower cost capital than vendors, so we tend to prefer a capex model.”
TowerXchange has also found that mature towercos are seldom receptive to gain-sharing or opex models from vendors. “We like spending money to save money,” the CEO of a large African towerco told us.
There are exceptions of course. For example, several early stage Myanmar towercos formed deep vendor finance partnerships in the first one to two years of the rollout, but in many cases those partnerships were put under strain by the relatively slow payment cycles from MNO to towerco to vendor – and exacerbated by limited access to foreign currency.
“Uptime is going to be even more critical if many 5G use cases are proven, such as telemedicine or autonomous vehicles. And 5G antenna also consume 2-3x as much power as 4G. In the 5G era, there will be a need for a ‘new normal’, where all stakeholders share a commitment to, and share the benefits of, lower cost energy,” concluded one of the energy working group participants.
Operations and maintenance
Towercos attending the working group adopted a variety of approached to operations and maintenance (O&M) from retaining skills in-house, to fully outsourcing O&M to third parties. Where did towercos feel the greatest gains were to be found from investing in O&M skills?
“A good O&M team enables effective field trials before new equipment is widely installed, so we can learn the actual performance in the field. We see those gains in monitoring and control,” said one towerco.
Another towerco agreed: “the skillset of the field force is key to bridging the gap between what the vendor promises, and field performance. For example, we use a mobile app to both deliver training modules, and to provide practical support at sites, ensuring our workforce are kept abreast.”
“Diesel pilferage is one area affected by the quality of O&M resources, particularly where you outsource,” added a third towerco. “And there’s always a risk if you award short O&M contracts that there is no incentive for the contractor to invest in training.”
Selected soundbytes from the energy working group
“We like fuel cell technology. If we bring a fuel cell to a site, and switch off the DG, it works great. It’s not the fuel cell technology that is the challenge, it’s the fuel supply and logistics.”
“We also like lithium-ion batteries. But any battery is only as good as the Battery Management System (BMS). We had a BMS showing 99% when the cell was actually at 70%, which incurred an SLA penalty, although the penalty was back to backed with the vendor.”
“We have learned the hard way that some technologies are too sensitive to be suitable for India. For example, they must be able to tolerate operation in temperatures in excess of 35°C, and be able to tolerate a lot of dust.”
“I’m sorry but I don’t believe anyone has a solar solution that requires a site visit as infrequently as once every 12-18 months. It’s not the battery cells, the air conditioning filters or anything else that need service – we simply have to clean the PV. Our current service intervals on solar hybrid cell sites are around every 500 hours – best case 1,000 hours.”
