The Economics
of Availability
What Infrastructure
Actually Pays For
Energy is only one part of the cost equation. For infrastructure operators, the larger challenge is often availability — the cost of having power that is continuous, predictable and dispatchable wherever it is needed.
This page studies the economics of availability across the three largest economic clusters shaping distributed power systems: diesel, intermittent generation, and electrification infrastructure.
The analysis is based on published industry reports, regulatory data, and publicly available market research.
The Cost of Energy Is Not
the Cost of Availability
Most energy decisions are framed around a single number: cost per kilowatt-hour. For infrastructure, that number describes only a fraction of what is actually paid.
A telecom tower, a data center, a water utility or a charging hub does not ultimately purchase electricity.
It purchases uptime.
Electricity is only one of the inputs required to achieve it.
Cost of Energy
€ / kWh — the number most decisions are based on
Cost of Availability
This page studies the economics of availability.
Read this way, diesel systems, intermittent generation assets, and electrification infrastructure are not competing solely on energy price. They are three different ways to purchase availability — each exposing the operator to a different combination of costs.
What does this page measure?
It measures the economics of availability — the full cost of keeping infrastructure powered continuously — rather than the headline cost of energy per kilowatt-hour. The analysis is based on published industry reports, regulatory sources, and market data cited throughout the page.
Diesel: When Energy Cost
Is Really Logistics Cost
In remote and weak-grid infrastructure, diesel often functions as a logistics system as much as an energy system. Its cost is driven by fuel supply chains, service access and price volatility — not by combustion efficiency. The global diesel genset market was estimated at roughly $36.8 billion in 2024 and is projected to roughly double by 2034 (GMInsights). In data centers, diesel still holds about 73–74% of backup generator capacity (Grand View Research, Precedence Research).
Around 500,000 telecom towers operate across Africa, and roughly 70% rely on diesel. At remote sites, energy can reach 30–60% of operating cost (CrossBoundary, GSMA). A single off-grid tower burns on the order of 13,000 litres a year — over $21,000 in energy OPEX and around 35 tonnes of CO₂ (GSMA) — before fuel transport adds a further 15–30% to operating cost. Over two years diesel rose 40–60% across many markets, and as much as 200% in Nigeria after subsidy removal (Ecofin, Africanews).
Asia-Pacific accounted for about 40.1% of global diesel generator revenue in 2024, driven by weak grids in India and China (MarketDataForecast). India's diesel genset market alone reached roughly $928 million, even as CPCB IV+ emission norms tighten the operating envelope (IMARC).
Diesel backup capacity at data centers nearly tripled, from about 20 GW in 2018 to 55 GW in 2024 (Better Data Center Project). European data center demand reached about 96 TWh in 2024 — 19% of national electricity in Ireland and close to 80% in Dublin — and the interconnection queue stood near 30 GW at year-end (Ember). Even as operators move to HVO and hydrogen, diesel remains the dispatchable fallback.
North America holds about 37.5% of the data center generator market. In Virginia alone — the world's largest data center market — more than 10,500 generator units were permitted, totalling 27 GW (Latitude Media). With downtime estimated near $9,000 per minute (Ponemon), backup is treated as non-negotiable, and diesel remains one of the fastest and most widely deployed forms of dispatchable backup.
Installed diesel backup capacity at data centers increased from approximately 20 GW in 2018 to 55 GW in 2024. Source: Better Data Center Project (via Latitude Media).
The growth of diesel backup capacity does not indicate falling diesel dependence. It indicates that the economic value of availability continues to outweigh the cost of maintaining backup infrastructure.
For a detailed engineering and total-cost comparison of diesel against the VENDOR.Max architecture, see the dedicated analysis at /compare/vs-diesel/.
Intermittent Generation:
The Economics of Buying Availability
Intermittent renewable generation is a mature and valuable class of technology. The point here is not whether it generates cheaply — often it does. The point is the cost of converting intermittent generation into continuous availability. That conversion has three structural cost components that headline LCOE numbers tend to omit.
The cost of 24/7 availability
To deliver power around the clock, an off-grid system is sized for the worst day, not the average one. Arrays are commonly oversized by 25–50% (OhmSnap), and battery banks are sized for two to five days of autonomy — on the order of 60–176 kWh for a roughly 20 kW system (IntegrateSun). Usable autonomy in December can be 40–60% shorter than in June (Anern). In many off-grid scenarios a backup generator is additionally used to bridge extended low-generation periods, so full-site resilience often still includes a fuel-burning reserve. A real 20 kW off-grid system frequently lands at $60,000–100,000+, or $4–7+ per watt installed, driven largely by storage (IntegrateSun, OhmSnap).
Indicative, representative 20 kW off-grid system. A battery bank sized for two to five days of autonomy spans roughly 60–176 kWh (IntegrateSun). Required capacity scales with the autonomy target.
The cost that appears above the site level
Availability also has a cost the individual site never sees on its bill. Globally, curtailed renewable output represented over $20 billion in lost revenue in 2024, with IRENA estimating cumulative foregone value near $100 billion by 2030 if unaddressed (Energy Central). Europe curtailed roughly 72 TWh in 2024, at a cost near €8.9 billion (Aurora Energy Research, via Solar Data Atlas). California's CAISO curtailed 3.4 million MWh in 2024, up 29% year-on-year (EIA). Negative prices — a market signal of midday oversupply — reached a record 23.1% of trading intervals in Australia's NEM in Q4 2024 (ScienceDirect), while German congestion management cost €2.77 billion in 2024 (industry analysis).
The cost of the storage lifecycle
Storage carries its own recurring economics. Off-grid battery banks typically require replacement every five to eight years (industry). And the end-of-life cost is now regulated: under the EU Battery Regulation (EU) 2023/1542, extended producer responsibility and take-back obligations apply from 18 August 2025, lithium recovery targets rise to 50% by 2027 and 80% by 2031, and the cost of recycling can no longer be routinely passed to the buyer (EUR-Lex, IEA, energy-storage.news).
For the architecture-fit comparison against solar-plus-battery, see the dedicated analysis at /compare/vs-solar-battery/.
Electrification Infrastructure:
The Cost Behind the Plug
The economics of electrified transport are less about vehicles than about the infrastructure that powers them — charging hubs, grid upgrades, peak demand, and the stationary buffers used to manage it. This is the cost layer that determines whether electrification scales.
Charging as a peak-demand problem
Poorly optimised charging can sharply raise peak power demand, increasing both cost and grid-connection timelines (IEA). On the operating side, demand charges — billed on peak draw — account for 23–85% of DC fast-charging operating cost; a 350 kW charger under a $20/kW demand rate can incur roughly $7,000 per month in demand charges alone (US DOE, Joint Office of Energy and Transportation). A DC fast-charging station typically costs $75,000–150,000 per connector installed, and retrofitting an existing site costs 40–60% more than building it ready from the start (US DOE AFDC).
Demand charges — billed on peak power draw — account for 23–85% of DC fast-charging operating cost. Source: US DOE, Joint Office of Energy and Transportation.
The cost of upgrading the grid
The distribution grid is the deeper constraint. A 2024 PNAS analysis of California's feeders projected that 50% would experience overload by 2035 and 67% by 2045 under EV adoption, requiring around 25 GW of capacity upgrades at $6–20 billion — with official state estimates reaching $26 billion (via Avanza Energy). Extrapolated nationally, infrastructure requirements exceed $300 billion. Residential EV charging can already draw more power than any other single household load (IEA).
Stationary buffers and battery replacement
The IEA itself recommends pairing high-power chargers with on-site storage to shave peak demand — but that adds battery CAPEX and a lifecycle that loops back to the previous cluster. Out-of-warranty traction or buffer battery replacement runs $4,000–20,000, with degradation of roughly 1.8–2.3% per year (Recurrent, MotorWatt).
Safety as an insurance and operational cost
Lithium battery fires are statistically rare compared with internal-combustion vehicle fires. The economic issue is different: when thermal-runaway events occur, they can be high-cost, complex-to-manage events for charging hubs, fleet operators, insurers, and facilities using stationary battery buffers. The Chevrolet Bolt recall, reported at approximately $2 billion, shows the scale of event-level cost exposure. For this page, the relevance is insurance, parking and charging protocols, suppression infrastructure, and operational continuity — not an argument against EV adoption.
On VENDOR's auxiliary role at charging sites, see /solutions/ev-charging-auxiliary-power/.
Seven Hidden Cost Categories
of Availability
Across all three clusters, the same cost categories recur. They are rarely on the price tag, but they determine the real cost of availability over a system's life.
- Fuel — combustion fuel and its price volatility
- Storage — batteries, oversizing, and their lifecycle
- Logistics — delivery, site access, technician visits
- Grid Upgrades — interconnection, reinforcement, demand charges
- Maintenance — servicing, cleaning, monitoring
- Compliance — emissions, noise, reporting obligations
- End-of-Life Disposal — regulated recycling and take-back
Different technologies expose operators to different combinations of these costs.
● Primary cost driver · ◐ Partial exposure · — Typically not exposed. Exposure varies by site, scale and operating conditions.
Why This Architecture
Exists
Every major availability model solves one problem by introducing another.
- Diesel provides dispatchability, but creates fuel and logistics dependence.
- Intermittent generation reduces fuel dependence, but introduces storage, oversizing and curtailment costs.
- Electrification removes combustion, but shifts cost toward grid upgrades, peak-demand charges and battery infrastructure.
The VENDOR architecture was developed around a different question:
Can availability be improved without inheriting the full cost structure of these models?
Its purpose is not to replace diesel, solar or electrification. Its purpose is to reduce exposure to the availability-cost categories that most strongly affect infrastructure operators.
That is the economic problem the architecture was designed to address.
Where the VENDOR Architecture
Changes the Cost of Availability
VENDOR.Max is a validation-stage architecture at TRL 5–6. The statements below are design intent, not certified commercial performance, and contain no pricing. The relevance to this page is narrow and specific: which of the seven cost categories the architecture is intended to affect through reduced exposure.
The architecture is designed to follow demand rather than produce surplus energy during periods of low consumption. It carries no battery bank in its primary architecture — so storage lifecycle and regulated end-of-life disposal are not primary cost drivers, and the thermal-runaway risk profile of large battery banks does not apply to the primary architecture. It removes the continuous fuel-logistics layer from the operating model. At the point of use, by following demand and limiting peak draw, it is designed to ease — not eliminate — pressure on grid upgrades and demand charges. The architecture is intended to support a CAPEX-dominant operating model with reduced exposure to recurring availability-related costs.
Energy is accounted at the complete device boundary, where classical conservation holds at all operational states:
Pin,boundary = Pload + Plosses + dE/dtThe architecture organises energy; it does not create it.
VENDOR.Max is not "cheaper energy per kilowatt-hour," and it does not compete with solar or electrification — both remain the correct choice in many deployments.
What changes is the structure of availability cost — which categories an operator is exposed to, and which are reduced or avoided within the operating model.
Why This Matters
Across the three clusters analysed on this page, the dominant cost categories recur: fuel, storage, logistics, and grid upgrades.
If future validation confirms the current design intent, the VENDOR architecture could reduce exposure to several of the largest availability-cost categories at the same time. Potential effects include:
- reduced dependence on fuel logistics;
- reduced dependence on large battery banks;
- reduced exposure to battery replacement cycles;
- reduced exposure to battery end-of-life obligations;
- reduced exposure to peak-demand costs;
- reduced exposure to certain grid-upgrade requirements.
The magnitude of any such effect remains subject to future validation and deployment conditions.
The potential advantage is not a lower headline cost per kilowatt-hour. It is the possibility of reducing several recurring availability-cost categories at the same time.
Exposure is qualitative and varies by site, scale and operating conditions. The VENDOR.Max values reflect design intent at TRL 5–6; it is not third-party verified, and the magnitude of any effect remains subject to future validation and deployment conditions.
For a detailed engineering and total-cost comparison, see /compare/.
Data Sources
and Limitations
This page is an industry-pain analysis. The figures are drawn from published industry, government and peer-reviewed sources, and are cited inline. It contains no VENDOR pricing, no VENDOR LCOE, and no claim of commercial superiority.
VENDOR statements are design intent at TRL 5–6 and are not independently verified. Market and competitor figures reflect published ranges and vary by geography, scale and site conditions.
Diesel — GSMA, CrossBoundary Energy, GSMA Intelligence, Ecofin Agency, Africanews, GMInsights, MarketDataForecast, Grand View Research, Precedence Research, IMARC, Better Data Center Project (via Latitude Media), Ember.
Intermittent Generation — Energy Central, Solar Data Atlas (Aurora Energy Research), EIA, ScienceDirect, OhmSnap, IntegrateSun, Anern, EUR-Lex, IEA, energy-storage.news.
Electrification Infrastructure — IEA Global EV Outlook, US DOE and the Joint Office of Energy and Transportation, US DOE AFDC, PNAS 2024 (via Avanza Energy), Recurrent, MotorWatt, Reuters, NBC News.
Source classification: INDUSTRY · GOV/REG · PEER-REVIEWED · MODELED · CANONICAL.