Australian grid-connection intelligence

Realistic flexible capacity, priced in dollars — not just whether a line runs nearby.

GridFlex calculates realistic curtailable capacity at specific Australian connection points, so developers can negotiate faster power and networks can assess flexible-demand proposals with a defensible method — not a guess.

SUBSTATION Firm capacity 180 MW · Level 5 + Flexible capacity 120 MW · Level 2 · modelled Candidate site B · Level 1

Illustrative diagram. Solid lines are confirmed. Dashed lines are modelled or indicative — and labelled as such, every time.

The problem

Australia's grid is fielding more connection requests than it can physically serve — with no agreed method for assessing the ones willing to be flexible.

Proximity to a transmission line has never meant deliverable capacity. Now the gap between the two is the single biggest obstacle standing between a data-centre project and an energisation date.

20 GW
of large-load connection enquiries received by Transgrid, against capacity it describes as largely exhausted in Western Sydney beyond 2033
Transgrid, public correspondence, 2026
67 GW
of proposed data-centre connection capacity considered in AEMO's 2026 Electricity Statement of Opportunities, up from 38 GW the year before
AEMO ESOO, 2026
0.25–5%
of the year that most networks actually spend under genuine peak stress — the rest of the time, real headroom exists but goes unrecognised
CSIRO, flexibility pilot announcement, 2026
"There is insufficient capacity in Sydney's transmission network today to connect any of the 20 GW of connection enquiries we have received for new large electrical loads." — Transgrid, EGM Network, public statement, August 2026

What we do

One calculation, two audiences.

GridFlex answers a single question — how much flexible capacity genuinely exists at a given connection point, and what would a data centre have to agree to in order to use it — for the two parties who need that answer to actually negotiate a connection.

For developers & operators

Flexible-connection studies

A commissioned study of one or more real connection points: estimated firm capacity, estimated flexible capacity on top, expected curtailment frequency and depth, and what that pathway is worth against waiting for a conventional firm connection.

  • Built from public AEMO dispatch and constraint data
  • Values firm, staged, non-firm and flexible-only pathways in dollars
  • Every output labelled with its evidence level
For network operators

A method for assessing flexible demand

Regulators are moving toward assessing large loads on their flexible profile rather than nameplate capacity — but no standard method exists yet for doing that. GridFlex's methodology is built to become that assessment engine as flexible connection agreements are formalised.

  • Transparent, source-referenced methodology, not a black box
  • Every output shows its evidence level, not just a number
  • Designed to be checked against a network's own operational data

A worked example

Same 50 MW site. Four ways to connect.

Net present value in A$ millions for a 50 MW data centre, grid-only configuration. Three pathways that start earlier come out at roughly double the value of waiting for a firm connection.

Firmwait about 4.5 years
106.3
Stagedflexible now, rest at the firm date+113.2 vs firm
219.5
Non-firmon the N rating+113.7 vs firm
220.0
Flexible onlycurtailable capacity+111.4 vs firm
217.7
NPV, A$mGrid only
159 A$/MWh
Grid + battery
131 A$/MWh
Grid + solar + battery
129 A$/MWh
Gas primary
154 A$/MWh
Gas island first
162 A$/MWh
Firm106.3106.3106.3208.6182.7
Staged219.5+113.2219.5+113.2219.5+113.2––
Non-firm (N rating)220.0+113.7220.0+113.7220.0+113.7––
Flexible only217.7+111.4217.7+111.4217.7+111.4––

Firm = 62.5 MW at 4.5 years on grid, 2 years on gas. – = not modelled. Illustrative model output from public ratings and stated assumptions; revenue and connection-cost inputs are assumptions. Not a network commitment.

The confidence model

Every number says where it came from.

A capacity figure is only useful if you know how much to trust it. GridFlex grades every output on a seven-level evidence scale — from a public data point to a signed connection agreement — and never presents one as the other.

0UnknownInsufficient evidence to make any assessment.
1Public-data indicationDerived from public infrastructure or regulatory information.
2ModelledCalculated from defined assumptions and a stated methodology.
3Engineering reviewedChecked against relevant technical expertise before release.
4Network indicativeSupported by preliminary engagement with the network operator.
5ConfirmedFormally confirmed through the relevant network process.
6ContractedSupported by an executed connection agreement.

Who it's for

Built for the two sides of a real connection negotiation.

Owner-operators, not landlords

A flexibility commitment only works if you control the load. That fits AI-factory operators and neoclouds running their own compute — not colocation providers whose tenants hold the uptime guarantees.

e.g. sovereign AI factories, neocloud GPU operators

Transmission-connected sites

The gap we work in sits at transmission scale — the large connection points serving 100 MW+ campuses — where no standard flexible-connection product exists yet, unlike some distribution-level offerings already in market.

e.g. NSW, Victorian and South Australian transmission-connected sites

Why now

The rules are moving — not there yet, but moving.

Aug 2026
AEMC advice to Energy MinistersRecommends a National Electricity Rules framework letting data centres voluntarily enter flexible demand agreements with networks.
Sep 2026
NSW proposes flexibility-weighted grid accessA proposal to allocate constrained transmission access partly on how flexible a project commits to being.
Oct 2026
AEMC Package 2 final determination dueTechnical access standards for large inverter-based loads, including data centres.
2027–2029
Formal connection-agreement reforms (indicative)Estimated window for flexible demand agreements to be written into the National Electricity Rules.

Where we are

Early, deliberately.

GridFlex is a new Australian company building this methodology from public AEMO data, starting with real, named connection points rather than a broad national platform. We are not a network operator, an engineering consultancy, a demand-response operator, or a guarantee of available capacity — and we say so before anyone has to ask.

Working model built; now validating it against published Australian data and seeking network input on real ratings.
Outputs are illustrative until confirmed by the network, and are labelled that way.

Get in touch

Have a connection point you need a real answer on?

Whether you're negotiating a connection, assessing a flexible-demand proposal, or working on the research side of this problem, we'd like to hear from you.

Email us directly

contact@grid-flex.com.au

Based in Adelaide, working across the NEM. You can also message the GridFlex page on LinkedIn.