​Ko changes how users interact with Modo Energy's market intelligence inside the Terminal. It answers questions using the same research, proprietary datasets, and market data that underpin Modo's analysis.

This article walks through three examples of how users can use Ko’s analysis with layered prompts. The examples cover market analysts, developers and financial analysts. In each case study, we show the prompt we gave Ko, charts built from the data Ko returned, and the insights Ko generated.

Access Ko for free in the Terminal here.

1. Market analysts: from constraint events to price formation

Market analysts can use Ko to investigate recent NEM activity. Ko combines live AEMO data with Modo's asset library and market datasets, linking regional price outcomes to the assets and constraints behind them.

In New South Wales, this is particularly useful for understanding how local constraints affect battery revenues. The prompts below use Ko to compare constraint losses across the fleet, examine Limondale's worst local pricing event, and assess the growing role of batteries in setting the energy price.

Output: Local price constraints in NSW from November 2025 to May 2026 were a very targeted problem. Only five batteries recorded any revenue loss, and the impact was heavily concentrated in one asset.

• ​Broken Hill dominates at $225,962 — almost 5x the next worst asset. 148 constrained intervals, at an average RRP of $339/MWh, meant that constraints were binding during genuine high-price events.
• The Riverina corridor (Riverina 1, Riverina 2, Darlington Point) accounts for ~$100k combined — near-identical interval counts and RRP averages confirm they’re being constrained off simultaneously on the same binding path.
• Limondale BESS is 5th at $8,116 — a more modest figure due to its short 7 weeks of operation.
• Waratah, Smithfield, Eraring 1, and Wallgrove had zero qualifying intervals — their locations look largely insulated from the constraints hitting western and far-west NSW.

Output: The constraint window ran roughly 08:15 to 18:00 AEST, around 9.75 hours. The pattern is striking and unambiguous.

• 00:00-08:10: No constraint. Local price tracks the NSW RRP exactly, both sitting in the $58 to $80 per MWh range typical of overnight and morning periods.
• 08:15: Constraints activate abruptly. The local price instantly diverges from the RRP and collapses into deeply negative territory, driven by surging solar generation creating congestion on the local network serving Limondale's Riverina location.
• 10:15-14:45: Peak divergence. Local prices repeatedly hit the market floor of -$1,000 per MWh, while the NSW RRP remained a modest $22 to $105 per MWh. The worst single interval was 10:45, with a local price adjustment of -$1,009.89 per MWh.
• 16:55-18:05: Constraint eases. Solar generation ramps down in the late afternoon, and the local/RRP divergence narrows progressively.
• 18:10 onward: Constraints fully clear. Local price snaps back to tracking the RRP at around $75 to $80 per MWh.

Output: The step-change from January 2026 is stark.

• April to December 2025. BESS held a fairly stable share of around 3 to 6%, peaking at 6.3% in December.
• January 2026 onward. The share tripled almost month-on-month, jumping from 6.3% to 17.3% in a single month, and has continued climbing to 22.4% in March 2026.

This inflection almost certainly reflects significant new BESS capacity coming online in NSW around late 2025 and early 2026, increasing competition at the margin and giving batteries much greater influence over spot price formation. By March 2026, roughly 1 in 4 to 5 dispatch intervals in NSW had BESS as the marginal unit.

2. Developers: where to build, what duration, and at what cost

Developers can use Ko to test early-stage investment assumptions against Modo's research catalogue. This includes regional outlooks, substate analysis, capex data, contracting research and the asset register.

The prompts below use Ko to compare state-level returns for a 2029 build, show how location changes the result within New South Wales, and benchmark build costs against disclosed project data.

Output: South Australia leads the Central scenario at a 12.7% IRR, followed by Victoria at 12.4%, New South Wales at 11.8%, and Queensland at 11.5%.

• South Australia leads because it has the most pronounced intraday price volatility, creating deep energy spreads that a 4-hour battery can capture across both morning and evening peaks.
• Victoria is the standout for cycling upside. Yallourn West closure creates near-term volatility that lifts revenues, and the state's morning peak structure creates a second profitable dispatch window.
• New South Wales is flat through the early 2030s. Eraring's retirement provides a smaller volatility boost than the southern state coal exits.
• Queensland ranks last among mainland states. The Gladstone retirements provide a mid-decade revenue boost, but revenues decline more sharply through the 2040s.

Output: NSW has the largest intrastate IRR gap of any mainland state. In the Central scenario, a 4-hour battery's unlevered IRR ranges from 11.8% at Central NSW (the Regional Reference Node), to 11.3% at Southern NSW, to 10.0% at Northern NSW. That 1.8 percentage-point spread is the largest location penalty in any NEM state, enough to materially shift an investment decision, and larger than the equivalent North/South split seen in Queensland.

Two compounding mechanisms drive the gap:

• MLF drag. MLFs scale energy revenues directly across the asset life. Northern NSW has the lowest MLFs of any mainland region, meaning every MWh dispatched is worth proportionally less at the meter.
• Network constraints. Constrained subregions limit how often a battery can dispatch into high-price intervals, precisely the periods a storage asset most needs to access to generate returns.

The rule of thumb is that proximity to the Regional Reference Node and a strong MLF are as important as state selection.

Output: Cost per MWh has broadly declined, from around $990 per kWh at Hornsdale to $560 per kWh in 2025, a roughly 27% reduction. This reflects global cell cost deflation and increasing project scale.

Cost per kW is less clear-cut. It dipped sharply in 2023 to $770 per kW before rising again. This largely reflects the duration mix shifting longer: newer projects are predominantly 2-to-4 hour systems, meaning more kWh per kW of power capacity, which inflates the per-kW figure even as per-kWh falls.

3. Financial analysts: deal terms, revenue history, and market depth

​Financial analysts can use Ko to pull the data and analysis that inform due diligence reports. This includes offtake terms, revenue history, asset benchmarks and research on deal structures, M&A and risk allocation.

​This helps teams benchmark proposed terms against the market. Contract tenor, structure and signed volume can inform how an offtake is positioned, what counterparties are accepting, and where negotiation ranges may sit.

The prompts below use Ko to show how contract tenor has changed, how revenues have shifted from FCAS to energy, and how much offtake volume has been signed since 2020.

Output: Across all contract types, average tenors have generally lengthened since 2020, reflecting growing offtaker confidence in long-dated BESS commitments.

• Physical tolls. The dominant structure throughout the period. Tenors started at 12.5 years in 2020, dipped to 10 years in 2021 to 2022, then climbed sharply, averaging 16.7 years in 2023 and holding at 14 to 15 years in 2024 to 2025. The 2023 spike was driven by several 15-to-20 year deals.
• Virtual tolls. Emerged in 2022 at 7 years, but have since extended significantly, reaching 10.5 years in 2024 and 13.3 years in 2025. 6 virtual tolls were confirmed in 2025 alone, making this the most active structure that year.
• PPAs. Only appeared from 2024 onward but at long tenors straight away. 12.5 years in 2024 and 17.5 years in 2025, suggesting they are being used for assets seeking the longest-dated revenue certainty.
• Revenue swaps. A thinner market, seen in 2021, 2023, and 2025 to 2026. Tenors have ranged from 7 to 12 years with no clear directional trend.

The 2025 cohort across all structures shows a clear preference for 10-to-20-year commitments.

Output: Energy arbitrage now drives almost all NEM battery revenue. FCAS has compressed from the main source of revenue in 2022 to a small supporting role in 2026.

• 2022. FCAS dominance. FCAS, particularly raise and lower regulation and raise and lower 6-sec, made up a substantial share of total revenue. In November 2022, total revenue hit around $360k per MW per year annualised, with energy contributing only around $34k or 9%, while FCAS accounted for the bulk.
• 2023. A more balanced mix. FCAS remained elevated but energy grew its share, representing roughly 40 to 50% of total revenue most months.
• 2024. Energy takes over. A decisive shift occurred from early 2024. Energy arbitrage surged, peaking at around $244k per MW per year in mid-2024, while individual FCAS services largely fell below $10k per MW per year.
• 2025 to 2026. Energy dominant, FCAS near-negligible. By 2025, FCAS revenues had compressed to near-zero for most services. Energy now regularly accounts for 90 to 95% or more of total revenue.

• 2023. Market inflection. Total signed MW jumped from 160 MW in 2022 to nearly 1 GW, driven by large physical toll and revenue swap deals as project developers raced to secure financing.
• 2024. Breakout year for virtual tolls. 640 MW signed via virtual toll structures, emerging as a genuine alternative to the physical toll, likely reflecting growing appetite from energy retailers wanting financial exposure without dispatch obligations.
• 2025. Biggest year on record at 2.2 GW. All four structures were active simultaneously for the first time, and PPAs surged to 640 MW, their largest year.
• 2026. Partial year data, 244 MW so far through early May.

Physical tolls remain the backbone, present every year and consistently the largest single structure, but their share of total MW is gradually declining as virtual tolls and PPAs grow.

​Key takeaways

Across the three examples, Ko helps users move from a question to the relevant data and analysis faster than the manual alternative.

The value is clearest when a question cuts across multiple sources. Ko can identify the relevant data, return the result, and explain what it means in context. This gives users a faster starting point for analysis, whether they are reviewing a market event, testing a project assumption, or preparing commercial material.

Ko does not replace analyst judgement. It helps users test the first answer, understand the supporting data, and decide where deeper analysis is needed.

If you want to understand how Ko can support your role, please get in touch with the Modo Energy team.