Eltek: A business model to buy and sell energy by the kWh

Tower operators would like to transfer the responsibility, and risks, of energy provision to specialist ESCOs. However there is a lack of consensus on the business model, pricing and contractual terms for selling energy by the kWh. In this article we speak to Eltek’s Bob Hurley to understand the issues.

TowerXchange: We spoke about this in our last interview, but let’s start with a summary of the current state of energy provision to cell sites in Africa.

Today African cell sites remain predominantly dependent upon diesel generators (DG). Grid power is neither extensive nor, in many cases, reliable, leaving many towers off-grid or relying on a grid connection that may be available as little as two to three hours per day.

To compensate for the lack of reliable grid power, there has been significant investment in dual diesel generators at many cell sites, with small batteries to bridge the time between the grid becoming unavailable and the DG kicking in. Many cell sites in Africa run the DG 24/7, at considerable expense in terms of fuel costs, refuelling and maintenance visits, and maximising exposure to fuel theft risk.

Over the last 18-24 months there has been a move towards reducing opex and accelerating the installation of solutions that are not dependent on diesel. The first step tends to be the installation of CDC (charge discharge) battery solutions “ essentially a larger battery to carry the load for a longer period allowing the generator to run much more efficiently at a higher load for shorter periods of time, thus reducing opex.

The next step is to install an alternative source of energy such as photovoltaic panels to capture solar energy and deliver that as a power source to the base station. Others have looked at wind and fuel cell solutions. The shortfall of fuel cells is that it incurs similar maintenance costs to diesel power and still requires frequent site visits to keep the fuel source topped up, whereas once solar is deployed, site visits are seldom necessary, providing it’s producing power. Of course you can’t guarantee sun, or wind, 365 days a year.

When rolling out a new site, tower operators’ preference would be to rely purely on grid power if possible. But we don’t feel that grid power will be extended fast enough to be an alternative to renewable energy - grids are prohibitively costly, so those that need power beyond the grid are left to source or generate power for themselves.

Two years ago I’d estimate that 20-30% of the RFPs and RFQs we received involved some sort of secondary energy (CDC batteries or full hybrid). Now it’s more like 70% of the quotations we provide - demand has more than doubled and now exceeds demand for diesel dependent solutions.

TowerXchange: What is the decision making unit at towercos and operators when it comes to investing in hybrid energy and energy as a service innovations?

The operations team typically source, test and recommend hybrid energy solutions, then the finance team signs off on the capex.

Typically the COOs of MNOs are tasked with reducing opex. They and their team are defining requirements and specs and liaising on hybrid technologies to see what opex costs can be saved and how.

The operations team will typically then ask for proof of concept, for example “we want to evaluate installing hybrid solutions at 200 sites, tell us how much we need to pay, the time to RoI, and the impact on opex.” We’ll often be asked to demonstrate the solution on one site over a period of 90 days, typically at our own risk.

Usually the COO takes the result of that trial to the CFO and the financial committee for the approval of capex. The CFO is usually the gatekeeper with control of purse strings and finance governs how capex is deployed, comparing opex reduction initiatives with investing in the rollout of more base stations. Capex deployment has to maintain a fine balance between investing to generate more revenue and spending money to secure better performance.

When that coin is flipped, it had fallen in favour of revenue generation rather than opex reduction capex investments 90% of the time, but within last 6-12 months the coin has started to fall the other way. This is driven by an increasing concern about service levels and customer loyalty as poorly maintained, poorly powered base stations risk increasing churn. The movement toward towerco-provisioned solutions based on shared infrastructure is also part of this increased appetite to invest in opex reduction - many operators are considering divesting their towers, making them more cash positive, and strengthening the case for investment in hybrid energy.

TowerXchange: Let’s talk about the financial models for the sale of energy to cell sites. What has been your experience of developing and offering zero capex financing models where tower operators can repay the capital costs over an extended period?

The operations department want the hybrid solutions that Eltek and our competitors are offering, yet the finance department has many demands on their finite capex budget. So Eltek and several of our competitors developed zero capex financing options.

We proposed to implement our opex saving solution, and to use those savings over a lease or loan period to payback equipment costs, with length of lease or loan contract to match the opex reduction, hence a zero capex model. Eltek were able to demonstrate payback between two and a half and five years depending on various parameters. Eltek then approached large financial organisations, including the World Bank, to put together financing packages based on funding up front capital investments and recouping them through repayments based on opex savings on fixed terms over a fixed period.

Under a zero capex model, opex is stablilised but not significantly reduced over first two to three years while the capital costs are being recouped, but at the end of that period all the opex saving drops to the bottom line. For example, in a 50 base station region running DG 24/7, by installing CDC batteries the tower operator could recover US $1.5m in capex over a three year period, saving US $500k per year. And after three years that US $500k saving drops straight onto the tower operator’s bottom line.

However, the operators have not bought into the zero capex model, due to three stumbling blocks.

The first stumbling block involves credit provision. Africa’s franchised mobile operators tend to buy in to a local operator by acquiring 10-80% of the equity. That local OpCo is often at or near their lending limits, so when our finance partner bank evaluates the customer’s suitability to be part of a US $2-5m loan, the banks will often require some form of guarantee from the parental operator, which is typically much more credit worthy. However in the majority of cases this guarantee is refused by the parent operator.

The second stumbling block is that the finance department often thinks they can borrow money at a cheaper rate than the bank and equipment provider are offering. But the finance department remain unwilling to finance the capex themselves.

The third stumbling block is the difficulty competing with Chinese OEMs who bundle finance energy and network equipment and service propositions over very long terms to secure deeper client relationships. At any green field cell site, power equipment might represent 10% of the total capex, while 70% of capex goes into active network, with the remaining 20% in static asset and construction costs. The value of a power and network equipment bundled contract might be seven to ten times the value of just the power component, so the OEM has a greater value to discuss finance packages with the operator, while they are also motivated by securing their long term business model by locking out competitive OEMs for a five year contract term.

TowerXchange: So if zero capex financing models aren’t working, tell us about the kWh energy as a service model.

Business models to pay for energy by the kWh originated in India, South East Asia and Indonesia.

Today in Africa, the mobile network operator often retains ownership of power systems and is only selling the steelwork to towercos. As the market moves toward a model where the operators and towercos agree that it would be easier if a third party put in and owned that power solution, that drives the ESCO (Energy Service COmpany) concept.

In Indonesia and India three potential ESCO models were developed:

• Fixed fee operating lease. Under this model, the ESCO puts the power equipment in, and sells or leases it at a fixed cost to the tower operator over a fixed period of time. While there may be minimal opex reduction, opex is stabilised and the ESCO is able to recoup their capital investment.

• Energy saving agreement. Again the ESCO puts the power equipment in, and the tower operator pays a percentage of opex savings to the ESCO on a monthly basis, based on a comparison of opex before and after installation.

• Power purchase agreement. This is the Holy Grail of ESCO business models! Under a PPA, the ESCO installs, services and owns  the power equipment and sells power to tower operators at an agreed rate per kWh.

TowerXchange: How successful has each ESCO business model been?

The fixed fee operating lease model has been relatively successful. However, it generally involves capex and commitment over a shorter period than the other models, so long term value creation isn’t optimised.

Energy savings agreements have been the least successful, primarily due to the difficulty agreeing how to measure the verified energy cost saving. Cost savings have to be measured weekly or monthly, and fluctuate depending on the weather, state of batteries et cetera. So the model is complex and opex may be reduced, but it’s not stabilised.

Meanwhile, everyone is keen to put together a power purchase agreement that yields a reasonable return for the ESCO at an acceptable rate for the tower operator, but to date the  two parties are to be found at either end of the scale.

TowerXchange: We’ve spoken to passive infrastructure decision makers at operators and towercos across Africa, and there’s a lot of interest in the kWh pricing model - what challenges still have to be overcome?

There are still significant challenges to be overcome in terms of agreeing the price per kWh and the contract duration.

Let’s breakdown the costs and risks that would be transferred to an ESCO. At an off grid cell site that might be running DG 24/7, you need regular site visits to top-up the fuel tank. You’re probably losing a proportion of fuel to theft. Plus the ESCO will be exposed to the risk of increasing diesel prices. You’ve got to add in the cost of preventative and reactive maintenance visits. Some of these cell sites might have been running for three to five years, and equipment has a finite lifecycle - even if optimally managed, rectifiers, batteries and generators themselves will eventually need replacing.

When the ESCO looks at the risks and costs they are exposed to over an annual period, that increases their minimum price per kWh above the price point tower operators are trying to achieve through their opex reduction initiatives. To give you an example, tower operators are looking for models that enable them to pay US $1-1.50 per kWh in Africa. For ESCOs to achieve RoI within 12 months, the cost would be around three times that level.

If an ESCO were to accept a US $1 per kWh price, after a given period they would either go out of business, and the operator would lose a long term partner, or they would have to reduce their costs by cutting back on services, with less site visits, cheaper batteries et cetera. QoS would ultimately suffer, and we’re back facing the same problems we are today.

The other problem is that the ESCOs want a 5-10 year contract to give them time to recoup capital outlay, to help them secure long term business, and to reduce their cost of capital. However tower operators want to renew contracts annually, which exposes the ESCO to risk of turnover, increases the cost of capital, and therefore increases the kWh price they need to charge.

Finally, the risk is heightened if the tower operator does not want to share all his data with the ESCO, which can be a sensitive matter when assessing risks such as fuel theft. There needs to be a trust model, where tower operators share data on their real opex, and ESCOs share the true costs so we have a better chance to agree a kWh rate from which both parties can benefit. This is a significant opportunity for the tower operator to partner with a vendor he knows and is comfortable with, whose equipment he already has in his network, who he knows has the financial stability to be a long term partner.

TowerXchange: How can we overcome these challenges and accelerate the adoption of ESCO business models in Africa?

Discussions to create a scalable PPA business model have reached something of an impasse, but this is where the infrastructure sharing proposition can help. If an ESCO is selling energy to a single tenant, RoI is longer. With multiple tenants, capex can be recouped faster, RoI accelerated and kWh prices reduced. As the towerco model matures and gains wider acceptance, so the economics of the ESCO business model get closer and closer to making sense.

The other potential solution would be for ESCOs to be formed of joint ventures between different equipment and service providers already engaged in telecom towers, each sharing the risk and returns from an ESCO joint venture according to their own investment.

We’re interested in exploring the other services beyond energy provision that could add value to the ESCO proposition. ESCOs will be entrenched service suppliers with expertise on the ground “ they know the sites, they know the information that needs to be shared to optimise those sites, and they can add value by diversifying their services beyond energy.