The Future Doesn’t Forecast Itself
Rethinking energy foresight in a system that no longer waits
If the future is already here, why are we still trying to predict it with yesterday’s tools?
Introduction
Forecasting has long served as the compass of the National Electricity Market (NEM), helping us navigate an intricate balance of reliability, investment, and system stability. At its best, it has enabled smooth transitions and market confidence. But in a time of accelerating change, are we asking too much of it — or perhaps, asking the wrong questions altogether?
Today, the NEM is being reshaped by digital infrastructure, electrification, decentralised generation, and environmental volatility. These forces operate at speeds and scales that traditional forecasting tools were never designed to anticipate. If the purpose of forecasting is to help us see what’s coming, we must also ask: what happens when the horizon itself keeps moving?
To begin exploring that, we need to understand how forecasting has shifted from mapping steady demand to navigating disruption.
From Demand Curves to Disruption Signals
The future no longer arrives in smooth lines
Energy forecasting once relied on population growth, historical demand cycles, and planned industrial expansion. That world is fading. AI data centres now appear with little warning, consuming more power than small towns. Consumer behaviour around electric vehicles and rooftop solar is variable and sensitive to incentives, costs, and sentiment.
We’re facing a reality where:
Demand is discontinuous: New loads arrive not as gradual increases but as abrupt jumps, often outpacing grid preparation.
Traditional signals no longer correlate: Forecasting based on past trends struggles with decoupled indicators — like demand not matching economic growth.
Forecasts trail rather than lead: Reactive updates become a constant, but rarely ahead of fast-moving change.
As these patterns shift, it’s worth examining not only what forecasts show, but how they shape decision-making behind the scenes.
The Quiet Politics of Forecasting
All models reflect choices — and those choices shape outcomes
Forecasts appear neutral, but they carry implicit decisions. They reflect assumptions about growth, technology, behaviour, and risk appetite. They are used not only to inform but also to frame what is considered possible or permissible.
This dual role means:
Forecasts become filters: They highlight certain trends while marginalising others — especially ones that don't fit established categories.
They guide investment — but may constrain vision: Developers and governments look to forecasts for certainty, even when the future is uncertain by nature.
They defer the hard questions: A forecast that shows no urgency can delay needed infrastructure — until the need becomes a crisis.
These political dimensions gain significance when forecasting becomes the gatekeeper for timely action — which leads us to the challenge of timing and readiness.
Reactivity Is Not Readiness
Waiting for data in a volatile system is a decision — often a costly one
Forecasting in the NEM has become a continual update process — new reports, revised expectations, emerging drivers. But while the cycle gives the appearance of responsiveness, it can foster delay. Infrastructure cannot pivot overnight.
Current practice may underplay:
Timing mismatches: Forecasts are updated yearly, but new loads and technologies move quarterly, or faster.
Investment paralysis: When forecasts shift too frequently or lack clarity, they risk freezing capital, not enabling it.
Systemic risk accumulation: When early signals are dismissed or slow to appear in modelling, systems fall behind — and catch-up is costly.
To meet these challenges, we must expand our view beyond reactive modelling and explore how strategy is being redefined at a system level.
The Strategic Step: AEMO’s Integrated System Plan
Planning long-term doesn’t guarantee foresight — but it’s a necessary foundation
While much of the operational forecasting in the NEM is reactive, AEMO’s Integrated System Plan (ISP) represents a more proactive, systems-level attempt to chart the future of Australia’s electricity grid over two decades. It is designed not just to anticipate change, but to guide coordinated action.
It reflects several important shifts:
Scenario thinking over single forecasts: The ISP models a range of plausible futures, including rapid digitalisation and decarbonisation.
System integration across domains: It links generation, transmission, and storage to physical realities and emissions goals.
Evolving assumptions: Stakeholder feedback has pushed for faster updates and stronger attention to disruptive new loads like AI-driven infrastructure.
The ISP signals an institutional awareness that forecasting must stretch beyond past patterns and into the space of system transformation. Still, strategy alone is not enough if it lacks cultural and operational follow-through.
Seeing the NEM as a System, Not Just a Market
Planning the grid must reflect how people, technology and institutions interact
The NEM is no longer just a price-based marketplace. It is a socio-technical system — a network shaped by human behaviour, corporate ambition, policy choices, and weather. Linear projections are insufficient tools for a system of feedback loops, thresholds, and tipping points.
To forecast a system means asking:
What futures are we preparing for — not just predicting? Probabilities alone may not help us act when low-probability events carry high consequences.
Who benefits from accuracy — and who bears the cost of error? Forecasting can protect some actors while exposing others.
Where do signals come from — and what are we missing? Not all changes begin with data. Some begin with political pressure, social expectations, or a breakthrough invention.
This systems view invites a different kind of mindset — one not just of technical mastery, but of foresight and cultural change.
Foresight as a Cultural Shift
Forecasting tools matter — but forecasting cultures matter more
Technical reforms will help — scenario planning, non-linear modelling, digital twins. But the deeper shift is cultural. What kind of mindset do we bring to uncertainty? Do we treat it as a threat — or as a prompt for adaptation?
We may need a planning culture that embraces:
Plausibility over probability: Preparing for what could happen, not just what’s likely.
Thresholds and tipping points: Using early warning indicators to trigger actions, rather than waiting for certainty.
Open, shared foresight: Bringing more voices into the process, and treating the public not just as users, but as stakeholders in the system’s future.
These cultural shifts don’t replace forecasts — they enable us to use them more wisely, more inclusively, and with a fuller sense of consequence.
Conclusion
The tools we’ve used to forecast Australia’s energy future have served us well — until they didn’t. Now, as the system is reshaped by forces that move faster than our reports and defy our assumptions, the role of forecasting must evolve.
What we face is not just a technical challenge. It is a governance challenge — a question of how we think, plan, and act in the face of uncertainty. If the NEM is to remain reliable, low-carbon, and responsive, our forecasting must do more than follow the data. It must help us imagine what kind of future we are willing — and ready — to build.
The future is not a point on a chart — it is a choice we make with what we’ve learned, and what we choose to notice.