Powered for Change: Why 90% of Energy Projects Repeat the Same Mistake

The energy transition is being delivered as a series of one-offs

The structural finding is blunt: 90% of projects follow a one-off pattern, with only 10% benefiting from repeatable teams or supply chains. That is a delivery model problem rather than an engineering one — each project assembles its own team, negotiates its own supply arrangements and rediscovers its own lessons, then disbands before any of it can be reused.

The cost is quantified, and it compounds. Accumulated project learnings can drive up to 20% cost savings in sequential projects within five to six years, and the research's inverse S-curve modelling shows those savings sharpen after an initial phase of modest gains, through a tipping point where experience and scale drive significant reductions, into a sustained phase where capital expenditure keeps declining. Applied to green hydrogen specifically, a multigenerational approach could reach cost parity with grey hydrogen nearly a decade earlier than a project-by-project approach, and generate up to $60 billion in net present value by 2050 from capturing just 5% of global demand.

Shipbuilding is the research's proof that this compounding is real, not theoretical: once shipbuilders adopted standardised, modular components and sequenced production to capture the learning curve, the fifth vessel in a series cost 50% less to produce than the first. Energy and heavy industry are earlier in that curve, which is precisely the opportunity — the compounding hasn't been captured yet.

Four levers, and why supply chains are the one most Australian executives underrate

The research identifies four levers as essential to making a multigenerational approach work — resilient supply chains, community and customer demand, workforce reinvention, and a strong digital core — and treats them as interdependent rather than sequential. Supply chains carry outsized weight for Australian readers because of where this country sits in the value chain: 74% of heavy industry executives expect supply chain volatility to negatively impact large capital projects by 2028, and separate research on procurement visibility found only 50% of companies have insight into more than half of their Tier-1 suppliers — a figure that drops to 20% for Tier-2 and 15% for Tier-3, which is exactly where Scope 3 emissions exposure and shortage risk live.

The fix the research prescribes is not more supplier auditing but structural: build long-term supply chain partnerships instead of transactional procurement, standardise and modularise designs the way offshore wind and grid-scale solar already have, and strengthen regional supply through local-for-local manufacturing. That last point is already moving in Australia's favour — 89% of surveyed chemicals, mining and metals executives expect to produce goods regionally within three years, up from 39% in 2023 — and the research ties unit cost reductions of 30-50% directly to the economies of scale that follow.

Investor sentiment is the pressure making this urgent rather than optional. The share of companies using predominantly negative tone when discussing new capital investments or projects rose from 30-35% in 2024 to more than 50% in 2025 — capital is getting more nervous about project viability at exactly the moment infrastructure spend needs to accelerate, which raises the cost of continuing to plan and procure project-by-project.

Community consent is now a capital-project risk, not a communications task

The second lever reframes community engagement as demand-side infrastructure rather than a late-stage courtesy. Fifty-two percent of heavy industry executives expect a lack of stakeholder engagement to negatively impact capital projects by 2028, and the research is specific about the cost of getting this wrong: unaddressed disconnect between decarbonisation projects and communities' perceived benefit can delay projects by three to five years — often longer than the engineering itself takes.

A live case study makes the mechanism concrete. Facing public reluctance to discuss nuclear power, a US campaign combining broadcast media, digital channels, influencer content and interactive tools tackled misconceptions directly — and drove a 50% increase in public support for nuclear energy expansion in nine weeks. For an Australian audience navigating its own contested conversations around transmission corridors, offshore wind and critical minerals extraction, the transferable lesson isn't the topic, it's the timing: early, transparent, fact-based engagement measurably shortens the approval runway that late engagement extends.

The research's own numbers on planning horizons reinforce this. Most companies report timelines of one to three years specifically for building community support, which means engagement now sits inside the critical path of project delivery, not alongside it — a line item that competes directly with the engineering schedule rather than trailing behind it.

The workforce gap AI spending is currently making worse, not better

The third lever is the one the research is most pointed about: industrial decarbonisation is equally a challenge of people and processes, yet three times more generative AI-related spending goes to technology than to the people who need to use it. More than 75% of companies analysed recognise the urgency of workforce reinvention, but only 30% expressed confidence in their ability to manage the change required — a large gap between recognising the problem and being ready to execute against it.

TenneT's approach to its German grid expansion is the research's model for closing that gap deliberately. Rather than treating safety training as a compliance checkbox, the European transmission operator built an actors-based, immersive, behaviour-driven training programme with 360-degree projection rooms and scenario-based simulations, mandatory for every person on site for more than three days. The results — a learner satisfaction rate above 90% across more than 3,000 sessions, with a target of training over 10,000 people over five years — show what happens when workforce investment is treated as infrastructure rather than overhead.

The research's prescribed sequence is to codify and share learnings across projects through centralised knowledge hubs, build interactive work-learn fusion so employees build fluency in new tools while deploying them, decentralise decision-making toward local teams with real-time authority, and embed a talent engine into the decarbonisation strategy itself rather than bolting workforce planning on afterward.

Hinkley Point C and the case for a digital core before scaling further

The fourth lever — a strong digital core — is what makes the other three compound rather than reset with each project. More than 75% of heavy industry and energy companies discussing digital transformation are prioritising it within the next three years, and the research found that companies deploying AI as part of an integrated digital core saw up to a 40% increase in project success rates alongside stronger regulatory and ESG compliance.

Hinkley Point C, the 3,260-megawatt UK nuclear project designed to generate low-carbon electricity for 60-80 years, illustrates why this matters beyond a single build. Facing data fragmentation and execution inefficiencies typical of first-of-a-kind projects, the operator established a centralised secure cloud platform for real-time data sharing, digital twins for 3D modelling that reduced engineering errors and enabled predictive maintenance, and AI-driven insights from mobile and IoT data for safety monitoring and automated compliance tracking — reducing costs, mitigating risk and accelerating timelines in the process.

The pattern across both case studies is the same one the shipbuilding data shows at the industry level: a digital core doesn't just make one project run better, it is the mechanism by which one project's learning becomes the next project's starting point — which is the entire argument for treating decarbonisation as a programme rather than a portfolio of one-offs.

Reading the economics honestly

The financial modelling in the research uses a 7% weighted average cost of capital discounted from 2025 to 2050, with net present value calculated on the difference in levelised cost between base and optimised scenarios. Those assumptions matter for Australian readers, because the discount rate is what determines whether long-horizon optimisation survives an investment committee: at 7%, value arriving after 2040 is heavily discounted, which is precisely why the near-term 20% saving from repeatable delivery is the more persuasive argument to bring to that committee than the 2050 figure.

The demand picture reinforces the point about staging. Green hydrogen at its projected tipping point of direct industry control still represents just 5% of global demand, and global electricity demand overall is forecast by the International Energy Agency to grow 80% by 2050 — nearly twice the rate of overall energy consumption — so build-out sequencing matters at least as much as build-out speed.

What Australian energy and infrastructure leaders should do next

Convert projects into programmes. The 90%/10% split on repeatable teams and supply chains is the single largest recoverable inefficiency in the research, and it is a structural choice, not a resourcing constraint.

Move stakeholder engagement into the critical path. With 52% of executives expecting engagement gaps to hurt capital projects and delays running three to five years when trust breaks down, community consent belongs on the same schedule as procurement and construction, not after it.

Rebalance the AI-to-people ratio. Three times more spend on technology than on the workforce that operates it is a specific, correctable imbalance — not a rounding error — and it is the difference between a digital core that compounds and one that sits unused.

Check your planning horizon against your asset life. Up to 75% of plans oriented to immediate results is what keeps the one-off delivery pattern locked in, on assets designed to run for decades. Explore Factor's research library or join a Factor energy and infrastructure event to compare delivery models with Australian peers.