Set as preferred source
Insights from The Future of Electricity: From Building Capacity to Shaping Flexibility.
The electricity transition is often framed as a race to build more generation. That framing is becoming incomplete.
Global electricity demand is projected to rise by approximately 40%–55% between 2024 and 2035. Solar and wind could increase from around 15% of total supply to 35%–40% over the same period. On the surface, this still looks like a capacity-expansion story: more data centers, electric vehicles, cooling systems and industrial electrification matched by more renewable generation.
The harder problem sits between supply and demand.
Electricity must be available at the right location, in the right quantity and within increasingly narrow time intervals. A system may have sufficient annual generation while remaining unable to connect a factory, support a data-center cluster or respond to a steep evening demand ramp. The constraint is shifting from how many megawatt-hours can be produced to how quickly, reliably and precisely power can be delivered.
That distinction reframes the next investment cycle. The assets with the highest strategic value may not be those that generate the most electricity, but those that make the wider system operable.
The system problem has changed
Traditional electricity systems were built around a simple operating assumption: demand fluctuated, but supply was controllable. Coal, gas, nuclear and large hydro plants could be dispatched in response to changes in consumption. Networks were designed mainly for one-way power flows from large generators to passive users.
That architecture becomes less effective as variable generation grows and demand becomes more concentrated.
Solar production can create surplus electricity during the middle of the day, followed by a sharp requirement for dispatchable supply when output falls in the evening. Wind variability can extend beyond hours into multiday periods. Electric vehicles, heat pumps and cooling systems add new peaks to distribution networks. Data centers can introduce loads comparable to industrial facilities, but often within regions whose connection capacity was planned years before those projects appeared.
Peak electricity demand may increase by roughly 40% by 2035, while short-term flexibility requirements could rise by two to seven times. This is why generation cost and system cost are beginning to diverge. Renewable electricity can become cheaper to produce while becoming more expensive to integrate.
The evidence is already visible. European congestion-management costs reached €4.2 billion in 2023 and could rise to €26 billion by 2030. US congestion costs were approximately US $11.5 billion in 2023. Negative electricity prices are becoming more frequent in renewable-heavy markets, not because energy has lost value, but because networks and demand cannot absorb it at the moment it is produced.
Curtailment reflects the same structural mismatch. California curtailed 3.4 million MWh of renewable generation in 2024, while Germany’s solar curtailment nearly doubled. These are not isolated market distortions. They are indicators that capacity is being built faster than systems can integrate it.
The transition is therefore moving from a supply-follows-demand model toward a flexible buffer model. Reliability increasingly depends on the combined contribution of storage, responsive demand, firm capacity, stronger networks and digital control.
Grid access is becoming an economic location factor
The grid was once treated mainly as supporting infrastructure: a regulated network that moved electricity between producers and consumers. It is becoming a direct determinant of industrial competitiveness.
Generation investment has risen from approximately US $600 billion in 2015 to more than US $1 trillion annually. Grid investment has increased much more slowly, reaching roughly US $400 billion. A decade ago, about US $0.60 was invested in grids for each dollar invested in generation. Today, the ratio is closer to US $0.40.
The result is a widening gap between nominal supply and usable supply.
Around 2,300 GW of projects were awaiting connection in the US at the end of 2024. In Europe, approximately 1,700 GW of renewable capacity is reportedly caught in connection queues. Transmission projects commonly require 10 years to complete, while some grid permits in Europe can take eight to 10 years. Generation can often be financed and constructed in a fraction of that period.
This turns grid availability into a location-specific asset. Regions with available capacity can attract data centers, advanced manufacturing and electrified industrial processes. Regions with constrained networks may have abundant electricity at the national level while remaining unable to accommodate new local demand.
Grid expenditure therefore carries a broader economic role than conventional utility accounting suggests. It supports industrial expansion, digital infrastructure, energy security and regional productivity. Treated in this way, grid capital is not simply a regulated cost recovered through tariffs. It is productivity infrastructure with long-duration economic returns.
The scale of required investment reflects this shift. The US may need more than US $1 trillion through 2030. Europe may require around US $3.5 trillion through 2035. China plans approximately US $574 billion in grid upgrades between 2026 and 2030, while the ASEAN Power Grid could require US $764–800 billion by the 2040s.
Yet a single capital strategy will not work across all markets.
The five-market archetype framework provides a useful way to distinguish where value is likely to emerge. VRE-Frontier markets such as Denmark and Ireland require storage, stability services and sophisticated short-term balancing. Grid-Bottleneck markets such as the UK, Germany, the Netherlands, Australia and Japan place greater value on interconnection, congestion relief and grid-enhancing technologies.
Transition-Scaling markets—including China, India, Brazil and parts of Southeast Asia—must expand demand, renewable generation, firm capacity and institutional capability at the same time. Adequacy-Stressed systems face a more basic reliability challenge, while Resource-Rich markets can use surplus dispatchable energy to support industrial development, regional trade or new energy-intensive sectors.
The framework matters because technologies do not carry equal value everywhere. A battery may primarily provide frequency response in one market, congestion relief in another and basic backup capacity in a third. Capital efficiency increasingly depends on matching the intervention to the type of imbalance rather than deploying a standardized technology portfolio.
The innovation bottleneck is becoming institutional
Most of the technologies required before 2035 already exist.
Battery storage is commercially mature. Demand response is well established in several markets. Dynamic line rating can increase usable transmission capacity by 20%–40% without waiting for new corridors. Renewable forecasting, predictive maintenance and advanced distribution management systems are proven technologies.
The obstacle is often not technical readiness, but institutional readiness.
High-impact projects stall because of permitting delays, interconnection procedures, fragmented market rules, unclear data ownership and insufficient coordination between transmission operators, distribution networks, regulators and private asset owners. Capital may be available and technology may be bankable, yet delivery remains constrained by the number of approvals, interfaces and organizations involved.
This changes the basis of competitive advantage. The ability to standardize engineering, navigate approval cycles, coordinate supply chains and convert investment plans into operating infrastructure becomes as important as the underlying technology.
Digitalization adds another layer. Future electricity systems will need to coordinate millions of rooftop solar systems, batteries, electric vehicles, heat pumps and flexible industrial loads. Centralized dispatch cannot manage this complexity alone. Virtual power plants, distributed-resource management systems and automated markets will aggregate small assets into resources capable of responding to grid conditions.
But orchestration depends on trust.
Every participating device requires a secure identity. Ownership, operating permissions, performance and settlement rights must be verifiable. Self-sovereign identity and related credential frameworks could provide this layer, allowing machines, companies and individuals to prove their authority to participate without relying entirely on a single platform operator.
This is more than a technical identity question. The entity controlling the aggregation layer may also control access to markets, customer relationships and operational data. A closed model can scale quickly but create platform dependence. A shared identity and registry architecture supports competition between aggregators, but requires stronger standards and governance.
The grid is therefore becoming software-defined, but software alone is not the value layer. The defensible position sits at the intersection of data quality, control authority, cybersecurity, market access and trusted participation.
Geopolitics is moving inside the electricity system
Electrification reduces exposure to some imported fuels, but it introduces new dependencies.
Transformers, switchgear, power electronics, semiconductors, batteries and critical minerals become part of the security equation. Long lead times for high-voltage equipment can delay entire grid programs. Trade fragmentation and local-content requirements may improve domestic resilience while raising project costs and limiting access to specialized suppliers.
Cross-border interconnection introduces similar tension. Regional power trading can reduce costs and diversify exposure to weather conditions. Yet countries remain reluctant to depend on imported electricity during periods of domestic scarcity. The ASEAN Power Grid illustrates the challenge: technology is available, but sovereignty, domestic adequacy and trading rules remain unresolved.
Europe has built the most mature model of market coupling, but physical constraints still limit how far price coordination can compensate for weak networks. Electricity can flow across borders only when interconnection capacity exists and national systems are willing to make it available.
This makes strategic insurance increasingly relevant.
Optional firm capacity, diversified equipment sourcing, long-duration storage and hardened digital infrastructure may appear underutilized during normal conditions. Their value emerges during low-probability, high-impact events: prolonged low-wind periods, heatwaves, fuel interruptions, cyberattacks or simultaneous equipment failures.
These risks are non-linear. A transformer shortage alone may be manageable. A transformer shortage combined with extreme heat, high cooling demand and delayed transmission upgrades can become a constraint on economic activity. Resilience depends less on predicting the precise event than on preserving enough optionality to absorb compound shocks.
From static planning to adaptive governance
Electricity infrastructure has traditionally been governed through long planning cycles. Forecasts are produced, investment plans are approved, and performance is reviewed periodically.
That model becomes less reliable when demand can change rapidly and locally. A single data-center cluster can materially alter a regional load outlook. Electric vehicle adoption can shift distribution peaks. A sequence of extreme weather events can weaken assumptions about seasonal adequacy.
Governance therefore needs measurable stress signals rather than dependence on fixed forecasts alone.
Connection queue length, congestion frequency, curtailment, negative-price hours, capacity margins and project-delivery performance provide early evidence that the system is moving out of balance. These indicators can reveal problems before they appear as shortages or outages.
Adaptive governance does not imply constant policy change. It means linking predefined indicators to structured reviews of market rules, investment sequencing and capacity requirements. The objective is not to eliminate uncertainty, but to shorten the distance between a change in system conditions and the institutional response.
This may become one of the defining differences between electricity markets. Systems able to interpret emerging stress and adjust before failure will retain reliability at lower long-term cost. Systems that wait for visible disruption will depend increasingly on emergency interventions, retained fossil capacity and connection restrictions.
The strategic shift
The Age of Electricity will not be determined solely by how much renewable capacity is installed. It will be shaped by whether grids, flexibility, firm supply, digital infrastructure and institutional capability scale together.
That parallelism is difficult. Generation projects, network upgrades, market reforms and digital systems operate on different timelines and depend on different stakeholders. But sequencing them as separate stages creates its own risk. Building generation before deliverability produces curtailment. Deploying distributed assets before digital coordination leaves flexibility unused. Electrifying demand before adequacy and grid capacity are secured converts power into a constraint on growth.
The next phase therefore favors system-level positioning over isolated asset strategies.
Grid capacity becomes productivity infrastructure. Flexibility becomes a core service rather than an ancillary market. Digital trust becomes part of physical reliability. Institutional delivery becomes a source of operating advantage. Strategic insurance protects against the moments when infrastructure, climate and geopolitics interact.
The central question is no longer whether the electricity system can produce enough energy in aggregate. It is whether power can remain available, affordable and controllable as economic activity becomes more dependent on it.
