OptiGrid says BESS optimisers need six months of onboarding before go-live, and is building a portfolio-level product to stop assets cannibalising each other

Speaking to ESN Premium following OptiGrid's selection to optimise Vena Energy's 408 MW Bellambi Heights BESS in New South Wales, CEO and co-founder Sahand Karimi has set out what he considers the minimum viable onboarding runway for a battery storage optimisation engagement: six months^[1].

What the six months covers

Karimi describes the window as covering SCADA integration work in parallel with setting up the trading strategy, two workstreams that cannot be meaningfully compressed^[1]. The SCADA link determines how quickly the optimiser can read asset state and push dispatch instructions; the trading strategy configuration determines what the optimiser is actually trying to achieve within the constraints the asset owner brings to the table.

Those constraints matter more than they might appear. Karimi says battery warranties and long-term service agreements can distort trading behaviour if an optimisation platform fails to account for them properly^[1]. A warranty that caps annual cycle count, for instance, changes the risk calculus on every dispatch decision. He notes that OptiGrid has observed this directly within its own portfolio, with optimal bids varying across assets on the same platform depending on how each asset's constraints were configured during onboarding^[1].

The point extends to constraint management more broadly. "Sometimes those constraints can actually change the trading behaviour in a way that would result in weird outcomes," Karimi says^[1], pointing to South Australia's price cap event on 21 June 2026 - when the NEM's AU$20,300/MWh ceiling was hit twice in a single evening - as an illustration of what poor constraint configuration can produce in a live market.

The human-algorithm boundary

Karimi also addresses the question of when human traders should override automated optimisation, and his answer is: rarely, and never by switching the algorithm off. His preferred model has human operators input preferences, constraints, and objectives, with the optimiser then working within those bounds rather than ceding control entirely^[1]. OptiGrid's OptiBidder platform is designed to keep the algorithm running continuously even when traders adjust its parameters^[1].

The practical implication is that the quality of the onboarding conversation - where asset owners articulate their commercial objectives, offtake obligations, and risk appetite - determines the ceiling on what the optimiser can deliver. A six-month runway is, in Karimi's framing, the minimum time needed to have that conversation properly and translate it into a working configuration.

A second product for portfolio owners

Karimi discloses that OptiGrid is developing a second product, OptiTrader, designed to sit above OptiBidder for asset owners managing multiple batteries in a single market^[1]. The problem it targets is intra-portfolio cannibalisation: when two assets under common ownership bid against each other and suppress each other's revenue^[1].

The distinction Karimi draws is between cross-owner correlation - where different owners' assets happen to behave similarly - and same-owner conflict, where a single portfolio's assets are structurally competing. OptiTrader is intended to address the latter^[1].

The Bellambi Heights project illustrates why the problem is becoming more acute. The 408 MW facility comprises two adjacent 204 MW units, each carrying its own commercial arrangements^[1], and OptiGrid's total live or contracted portfolio has now crossed 1 GW. As Australian battery fleets grow and individual owners accumulate multiple assets, the portfolio coordination problem will only become more common.

The Bellambi Heights BESS is expected to reach commercial operation in 2027. Whether OptiTrader is live by then has not been disclosed.