Chapter 7

Colocation Engine

A telecom tower is a fixed-cost asset that gets dramatically better with a second tenant. On the company's own simulation, adding a second lessee to an existing tower roughly doubles revenue at almost no extra cost, lifting unlevered return on investment from about 11% to about 21% and halving the payback period. That operating leverage — not the concrete and steel — is TOWR's moat. The problem this chapter documents: the engine is idling. Tenancy has sat near 1.67x, sector-wide ratios are falling as three carriers deduplicate overlapping sites, and 2025 tower-leasing gross profit grew just 0.81%.

What a second tenant does

The economics of the tower business live in one number: the tenancy ratio, or tenants per tower. Building a tower and signing its first ("anchor") tenant is a mediocre standalone investment. Loading a second tenant onto that same structure is one of the best returns in infrastructure, because — in the company's words — "the operating costs of the towers are largely fixed, and the Company incurs only relatively low costs to add tenants." [1]

TOWR's public simulation for a single tower makes the leverage explicit.

Source: TOWR public expose, "Colocation Boosts ROI" single-tower simulation (ROI defined as EBITDA less 10% final tax over total capex; 10-year ground lease) [2].

The second tenant pays Rp144 million more in annual rent against roughly Rp150 million of extra capex; the incremental EBITDA margin is close to 100%, which is why the blended margin rises from 85.0% to 88.5% and the unlevered ROI nearly doubles. [3] This is the entire reason independent towercos exist: a landlord can earn a return on a shared mast that no single carrier building alone ever could.

Two features protect that stream once it exists. Leases run "10 years for tower and can be longer for fiber," are non-cancellable and renewable; renewal is near-automatic because relocating equipment is costly and disruptive to a carrier's network. [4] The barriers to a new entrant — capital, scale, permitting, local relationships — are high. [5] The asset quality the value case rests on is real, and it is this: contracted, escalating, high-switching-cost cash flows with a dormant high-return upgrade path built in.

The engine is idling

The upgrade path is what has stalled. TOWR ended 2025 with 36,247 towers and 60,540 tenants — a tenancy ratio of 1.67x, against 35,400 towers and 58,035 tenants (about 1.64x) a year earlier. [6] A rising ratio would be the collocation engine at work. This one barely moved, and management was candid that even the small uptick was not new demand: the ratio is "1.67, slightly higher than 2024, because we basically restructured some reseller contracts to become direct lease to our towers… in the past, we did not count reseller as part of tenancy ratios." [7] Strip out the reclassification and organic tenancy was flat.

This is not a TOWR execution problem — it is the market structure the report has been circling. Across Indonesia's listed towercos, tenancy ratios are declining as the carrier count falls. Tower Bersama, the operator that has historically run the sector's highest ratio, has seen it slide every year: 1.87x in 2022, 1.84x in 2023, 1.79x in 2024, and 1.73x in 2025. [8] [9]

Sources: Tower Bersama operational tables (FY2024, FY2025 Annual Reports) [10] [11]; TOWR ratios derived from year-end tower and tenant counts [12]. TOWR's 2025 figure includes reclassified reseller contracts.

The peer that sits closest to TOWR draws the ceiling plainly. Gihon reports its own ratio at 1.67x, notes the 2025 Indonesian industry average ran "1.60x–1.75x," and describes even 1.80x as a "near-term target" it has not reached. [13] The pattern is not local to Indonesia: in India, another market consolidated to roughly three national carriers, Indus Towers runs a tenancy ratio of about 1.65x, while portfolio towers in the fragmented United States market carry 2.4–2.6 tenants apiece. Tenancy ratio is, to a first approximation, a function of how many independent buyers exist — and Indonesia now has three. That is the structural fact the derating side of the thesis leans on: with the buyer base fixed, the collocation lever cannot be pulled at will.

Where the growth went instead

If the highest-return lever is stuck, growth has to come from somewhere lower down the return ladder — and the segment accounts show exactly where. TOWR now reports two segments: tower leasing and everything non-tower (fibre-to-the-tower, FTTH, connectivity). In 2025 the tower-leasing segment's gross profit grew just 0.81% year-to-date, while the non-tower segment's grew 14.05%. [14] Tower leasing brought in Rp8.73 trillion of revenue and non-tower Rp4.60 trillion. [15]

Sources: segment revenue split (FY2025 Annual Report, Directors' Report [16]); segment operating profit and growth (Business Segment Operational Review [17]). Operating margins computed on segment revenue.

The tower line still earns a 62% operating margin against 44% for non-tower [18] — so as the mix tilts toward fibre, blended profitability dilutes, which is the mechanical source of the margin drift earlier chapters traced. And the tower line itself is decelerating: leasing revenue grew 5.0% in 2024, to Rp8.53 trillion [19], then roughly 2.3% in 2025 to Rp8.73 trillion. [20]

Even within towers, the growth that does occur is the low-return kind. TOWR's model is anchor-first: it builds "build-to-suit" towers only once an operator commits as anchor tenant, then hunts for collocation later. [21] Of the 847 towers added in 2025, management noted most were relocations tied to the Indosat-Hutchison integration rather than fresh multi-tenant demand — "short of a couple of hundred towers that we need to conclude for Indosat, Hutchison." [22] A single-tenant new tower earns roughly the 11% ROI in the simulation above, not the 21%. The company is adding towers and fibre — deploying capital at the low-return end of its own opportunity set — precisely because the 21% collocation return is not on offer in a three-carrier market.

What it means

This is the operational spine beneath the valuation gap. The best thing about a tower business — the near-free second tenant — is exactly the thing TOWR cannot currently do much of, and that is a defensible reason the market has stopped paying a growth multiple for it. The tenancy ratio is the single most falsifiable line item in the whole thesis: if it keeps grinding down toward the low-1.6s, the derating is right; if it turns up, the mispricing is.

The counter is that the lever is dormant, not broken. Every tower already standing carries embedded, near-100%-margin capacity that costs almost nothing to fill, and the demand that would fill it — surging data traffic, eventual 5G densification, and the resolution of merger-driven site overlap once carriers finish rationalising — is a question of when, not whether new towers get built. A move from 1.67x back toward the 1.80x that peers still frame as reachable [23] would drop through to profit at incremental margins the fibre business cannot match, and the market is currently paying nothing for that optionality. The honest position is that both readings are live, and the tenancy ratio — reported every quarter — is where a patient investor gets to watch which one is winning.