Evidence · Fleet trials

IHS Towers — Nigeria

Across eleven telecom tower generator sites in Nigeria, the proof of concept produced a median fuel-saving result of 1.7%, with individual site results ranging from −1.6% to +5.7%. Two sites showed higher consumption and the operator recorded "no observable savings" at four of the eleven sites. The observed result is therefore positive but small and variable, and does not establish a clear treatment effect for this duty cycle.

Small positive result · no clear treatment effect established Telecom tower generators

A power-generation site silhouetted against a sunset

According to the operator, these telecom generators spent more than 90% of their running time at idle with no load. That provides a plausible explanation for the limited result, but load was not recorded in the trial workbook and the explanation therefore cannot be verified from the trial data itself.

+1.7%Median across all eleven sites

−1.6% … +5.7%Full range of site results

2 of 11Sites with higher consumption after treatment

4 of 11Sites recorded as no observable savings

IHS Towers trial — at a glance
Operator	IHS Towers, Nigeria — telecom tower infrastructure
Plant	11 generator sites, 17 kW to 50 kVA
Trial period	June to October 2025
Measured quantity	Average consumption per day
Full-period result	Median +1.7%, range −1.6% to +5.7%
Ten-day window result	Median +1.0%, range −4.5% to +3.4%
Operator assessment	No observable savings recorded at 4 of 11 sites
Conclusion	Small positive median result (+1.7%); no clear treatment effect established for this duty cycle

The result

Every measurement, not just the best one

IHS Towers — Result per site, full trial period, as calculated in the operator’s own analysis. Positive means lower consumption.
Site	Generator	Before	After	Saving
IHS_ABJ_1413C	IPT DCDG 17 kW	9.866	9.304	+5.7%
IHS_OGN_1307A	JMG 22 kVA	10.421	10.057	+3.5%
IHS_ENG_0773A	JMG 20 kVA	80.280	77.714	+3.2%
IHS_LAG_1266A	JMG 50 kVA	114.376	110.754	+3.2%
IHS_LAG_0718E	FG Wilson 20 kVA	38.989	38.042	+2.4%
IHS_ENG_0795A	JMG 30 kVA	96.156	94.507	+1.7%
IHS_OGU_0234H	Mikano 30 kVA	8.183	8.115	+0.8%
IHS_LAG_1064A	JMG 33 kVA	83.684	83.514	+0.2%
IHS_OGN_0949A	JMG 30 kVA	9.027	9.012	+0.2%
IHS_ABJ_1129C	IPT DCDG 17 kW	8.774	8.840	−0.7%
IHS_ABJ_1413C	FG Wilson 33 kVA	8.302	8.431	−1.6%
Median across all eleven sites			+1.7%

Grey marks the four sites the operator’s own analysis records as showing no observable savings. Median across all eleven sites: +1.7%.

IHS Towers — how it was measured

Method before result

Daily fuel consumption was recorded at each site before and after treatment over a five-month period, and analysed by the operator. We recalculated every site figure from the daily data and arrived at the same results.

We also checked whether a shorter comparison window would change the picture. A ten-day window around the application gives a median of 1.0% and four negative sites — slightly worse, not better.

Individual days do reach much larger numbers, up to 43% at one site. But the same daily data also swings as far as −145% on a single day. That scatter is noise, not signal, which is why the site-level average is the only figure worth quoting.

How we measure, and what makes a fuel trial credible →

IHS Towers — what this result does not prove

The limits of this trial

We publish these next to the result rather than in a footnote. If a limitation would change how you read the number, you should read it at the same time as the number.

Known limitations

The operator’s explanation is that these generators spend more than 90% of their running time at idle, with no load. That is technically plausible and consistent with our engine test-bench data, where the measurable effect appears under load and not at start-up.
But the workbook contains run hours and fuel consumption only — no load data. So the idle explanation cannot be verified from the trial data itself, and we will not present an unverified explanation as a finding.
Chart images circulating with this dataset suggest around 13% for a June–July window. That figure does not match any site result we can reproduce from the workbook, and its provenance is unresolved. We are not using it.

IHS Towers — what it means

Reading this evidence in context

A credible evidence base has to include weak and mixed results as well as strong ones. This trial marks a boundary in what we can claim: the median result was positive, but small and variable, and it does not establish a clear treatment effect across these eleven generator sites.

If the idle hypothesis holds, it also predicts where the treatment should work on stationary plant — gensets that carry real load, which is a different trial we would want to run properly rather than assume.

We are asking IHS Towers for load profiles. If they show the sites really did run at idle, this becomes a useful, well-understood boundary case instead of an open question.

IHS Towers — FAQ

Questions operators ask about this trial

Why publish a trial with such a limited result?

Because weak and mixed results are part of the evidence too. Across the eleven sites the median result was positive at 1.7%, but the spread was wide, two sites used more fuel and the operator recorded no observable savings at four sites. Publishing that alongside stronger results shows where the evidence is limited rather than selecting only favourable trials.

Does this mean NanoEFX does not work on generators?

No. The eleven-site trial produced a small positive median result of 1.7%, but the result was too small and variable to establish a clear treatment effect for that duty cycle. According to the operator, the generators spent more than 90% of their running time at idle with no load, which is a plausible explanation for the limited result; however, load was not recorded in the trial workbook, so that explanation remains unverified from the dataset itself. Generators carrying real load are a different duty cycle and should be measured separately.

Some individual days show much bigger savings. Why not use those?

Because the same daily data also shows single days far into negative territory. Picking the good days out of noisy data is not measurement, it is selection. The site-level averages are the honest figures, and they are what we publish.

IHS Towers — where this data comes from

Provenance

Operator: IHS Towers, Nigeria
Plant: 11 generator sites, 17 kW to 50 kVA
Document: NanoEFX POC analysis, IHS Towers Nigeria (14 Nov 2025 version)
Trial period: June to October 2025
Open item: Load profiles requested from the operator
Last checked: 15 August 2026

Prove it on your own fleet

The only baseline that matters is yours

Every trial on this site was run on someone else’s vehicles, duty cycle and baseline. A NanoEFX pilot measures yours — and we will help you design it with a control group so the result stands up.

Book a risk-free pilot All trial results → Calculate your ROI → Ask a question