Insights · Edition 01 · October 2026

Welcome to our first edition

Why we started ClarityAxis

ClarityAxis began, fittingly for two people who have spent much of their careers between programs on different continents, over coffee in airport lounges. Somewhere between delayed boarding calls and one flat white too many, our two founding partners kept returning to the same question.

How do we improve productivity on major projects?

Between them, they had carried accountability for cost, schedule, risk and reporting on programs including Inland Rail, AUKUS, MTR Hong Kong, Sydney Metro and the Schiphol Airport expansion. On each, they had watched capable controls teams lose days to manual data work and disconnected systems, and seen reporting arrive too late to change a decision.

They founded ClarityAxis to close that gap, pairing practitioner experience with AI built for project controls, so that controls professionals can spend their time on judgement rather than reconciliation. Today we work in three ways: ClarityAxis Consult, ClarityAxis AI and ClarityAxis Academy.

How we work

One group, three ways to accelerate your growth.

Advise, automate, uplift. Each stream stands on its own, and capability built in one carries straight into the others.

Expert-led delivery

ClarityAxis Consult

Integrated project delivery and controls: planning, cost, risk and reporting.

AI-human ensembles

ClarityAxis AI

Project controls apps that run across Microsoft 365.

Practitioner training

ClarityAxis Academy

Hands-on training across project controls, Contruent and applied AI.

ClarityAxis AI

Inside Microsoft 365 · Copilot · Teams · Outlook

Scope in. A linked, checked schedule out.

Schedule Generator in Microsoft Teams

Live

Build the 48 MW data centre schedule. Two halls, staged energisation. Export to P6.

Read scope

Build WBS

Link logic

Run CPM

DCMA check

4D view

Beta preview

Complete

In progress

Planned

1,180

activities

wks critical

13/14

DCMA score

.xer

P6 export

Minutes, not days. You review. You decide.

Also risk registers, management plans, QCRA/QSRA, DCMA checks and contract reviews, all from the Excel, P6 and Microsoft Project files you already keep. People stay in the loop on every output.

Assess

Create

Manage

Assure

Try it yourself

Each edition of ClarityAxis Insights will examine one question facing Australia's major projects and what the evidence says. We start with the one reshaping our own work.

Insight

Major projects

Building the AI factory: why Australia's next capital cycle will be won or lost in project controls

Data centres, transmission and a record public pipeline are competing for the same grid, supply chains and people. The binding risks now sit at interfaces that baselines treat as a single milestone or a contingency line.

The ClarityAxis Group team6 min read

Key takeaways

Demand is converging. The five-year Major Public Infrastructure Pipeline has risen 14% to $242 billion,1 while AEMO forecasts data centre electricity use in the NEM to grow almost sevenfold by 2035-36.3

Delivery capacity is not keeping pace. Infrastructure Australia projects a peak shortfall of 300,000 infrastructure workers by 2027, including around 59,000 project management professionals.1

Risk has moved to the interfaces. Grid connection, transmission cost escalation and labour availability now drive outcomes more than any single work package.

The response is a controls response. Treat energy as scope, re-price estimates, set objective triggers, and use AI to relieve scarce people.

Two booms, one delivery system

The next era of capital projects has an unusual circularity: artificial intelligence is now both an end product and a tool of construction. Developers of large data centres are, in effect, building AI factories. They are doing so while governments deliver record investment in transport, housing, energy and social infrastructure, with pipelines in Queensland firming around the Brisbane 2032 Games.1

Australia's Major Public Infrastructure Pipeline stands at $242 billion over the five years to 2028-29, up 14% on the previous outlook. Utilities investment more than doubled to $36 billion, largely because transmission projects were added.1 History counsels caution: across a database of more than 16,000 projects, only 8.5% met both their cost and schedule targets.2

Exhibit 1

The build-out in numbers

$242b

five-year Major Public Infrastructure Pipeline, up 14% 1

300k

projected peak infrastructure workforce shortfall, 2027 1

8.5%

of 16,000+ large projects met both cost and schedule targets 2

34 TWh

forecast NEM data centre use in 2035-36, from ~5 TWh 3

40%+

of data centre projects since 2025 dropped out or regressed in connection status 4

25–55%

real increase in overhead transmission line costs vs 2024 ISP estimates 6

Sources: Infrastructure Australia; Flyvbjerg and Gardner; AEMO. See references below.

What the data shows

Power is the new critical path for data centres

AEMO forecasts NEM data centre consumption to rise from about 5 TWh in 2025-26 to 34 TWh by 2035-36, from around 3% to 13% of grid-supplied electricity.3 Its forecast considered 225 known projects, yet more than 40% of data centre projects since 2025 have either dropped out or regressed in connection status.4 Large connections are targeting about two years from application to energisation.5 The energisation date, not building completion, now governs when revenue starts.

Transmission estimates are ageing faster than baselines

AEMO's updated cost data shows overhead line costs up about 25 to 55% in real terms against equivalent 2024 ISP estimates, and substation costs up 10 to 35%.6 The 2026 ISP still calls for about 6,000 km of new transmission and new gas supply chain assets to support 17 GW of flexible gas generation.7 Schedule confidence is low: 58% of 298 energy projects tracked by Infrastructure Partnerships Australia were rated as having a low to moderate likelihood of meeting their schedule.1 In transport, Grattan found the third of projects announced before formal commitment caused more than three quarters of cost overruns.8

People, not capital, are the scarcest input

Infrastructure Australia projects the infrastructure workforce shortage could peak at 300,000 in 2027, with regional shortages forecast to quadruple between 2025 and 2027, driven by energy projects.1 59% of firms surveyed cite labour and skills shortages as a substantial threat to delivery.1 In the data centre sector, 81% of industry leaders surveyed by Turner & Townsend say skilled labour in their region is constrained.9

The energisation date, not building completion, now governs when revenue starts.

What leading project teams are doing

In our work across major programs, four practices stand out among the teams keeping pace.

Practice 01

Treating energy as scope

Grid connection, network works and firming contracts are modelled as owned, logic-linked activities in the integrated master schedule, not one "power on" milestone, and energy obligations sit in base capex, not contingency.

Practice 02

Re-pricing estimates against the market

Open estimates are refreshed against current market data, with the estimate class stated, and contingency set using reference-class data and QCRA/QSRA at P50 and P90.

Practice 03

Setting objective triggers

Agreed Rules of Progress Measurement (RPM) and earned value thresholds (CPI, SPI, TCPI) force a forecast review and change control before a variance becomes a headline.

Practice 04

Augmenting before automating

AI is first used to remove administrative work, such as tender schedule builds, DCMA checks and monthly reporting packs, with people accountable for every output. Only then are the newly standardised workflows automated.

Implications for project controls leaders

Data centres

Anchor the program on the energisation date and report MW energised against MW contracted alongside physical progress.

Energy and transmission

Test whether current estimates reflect post-2024 cost data, and track contingency drawdown against risks actually retired.

Major programs

Resource-load schedules against realistic regional labour, track productivity rather than hours, and decide which recurring controls tasks AI should take first.

Questions to put to your team

Is our grid connection in the schedule as owned, logic-linked activities, or as one milestone?

When were our open estimates last re-priced, and what class are they?

Would our forecast change if we could not fill our next controls vacancy?

Which recurring controls tasks could AI take on this quarter, and who remains accountable for the output?

AI is now helping to build the infrastructure it runs on. The teams that keep pace will not be those with the most tools, but those that put connection, cost and capacity risk into their baselines early, and free their people to manage it.