Post: Demand response in 2026: from strategic advantage to grid imperative | an interview with GridBeyond’s Chief Commercial Officer, Mark Davis

Demand response in 2026: from strategic advantage to grid imperative | an interview with GridBeyond’s Chief Commercial Officer, Mark Davis

As global electricity demand accelerates and AI-driven load growth reshapes power systems, demand response has moved out of the margins and into the centre of strategy. Here we speak to GridBeyond’s Chief Commercial Officer, Mark Davis to get his read on the developments shaping the demand response sector in 2026.

Q: Let’s start with the big picture. Where does demand response sit in the energy conversation right now, compared to even two or three years ago?

It’s a fundamentally different conversation. Two or three years ago, demand response was often framed as a cost-saving measure, something a finance director got interested in because it trimmed the energy bill. That framing hasn’t disappeared, but it’s been overshadowed by a new reality. Grids are under structural pressure, and demand response is one of the few tools that can be deployed at scale, quickly, without waiting for new transmission lines or new generation capacity to be built.

Global electricity demand is now growing faster than at any point in decades, driven by electrification of heat and transport, but above all by the data centre and AI buildout. It’s easy to understand why grid operators are thinking about demand-side flexibility in a completely different way. It’s no longer a nice-to-have. It’s load management at a scale we’ve never had to think about before.

Q: Data centres keep coming up in any conversation about the future of power. Are they part of the problem or part of the solution?

They’re both, and the industry is figuring out how to tip the balance. Right now, the dominant narrative is that they’re a source of enormous, concentrated, and fast-growing load. Data centre energy consumption globally was around 415TWh in 2024 and it’s been growing at roughly 12% per year. With AI workloads scaling rapidly, forecasts suggest that figure could more than double by 2030.

But here’s what’s changed in 2026; the more sophisticated operators have stopped thinking of themselves as passive consumers of electricity and started positioning themselves as grid participants. The UPS battery systems, backup generators, and cooling infrastructure inside a hyperscale data centre represent real flexibility. The ability to shift non-urgent AI training workloads, pre-cool or reduce cooling during periods of system stress, or discharge on-site storage are genuine grid services. We’re seeing some of the largest operators co-investing in grid infrastructure and actively exploring demand response programmes, not just because regulators are asking them to, but because it makes commercial sense and helps them secure the grid connections they need.

The challenge is unlocking that flexibility systematically, at scale, in a way that doesn’t compromise the uptime guarantees that underpin the whole business model. That’s where smart aggregation and AI-powered optimisation come in and where companies like GridBeyond see a significant opportunity.

Q: On the regulatory side, what are the most significant developments in the UK that flexibility providers need to understand right now?

In the UK, two things stand out. First, the Demand Flexibility Service has evolved from an emergency winter tool into a year-round balancing mechanism. NESO data shows it has engaged over 4 million participants delivering 7GWh of flexibility across two winters. That’s a meaningful proof point, and recent moves set out a direction that broadens access further, including reducing the minimum unit size to 0.1 MW, which opens the door to aggregated portfolios of smaller assets that previously couldn’t participate. That’s a significant shift.

Second, the Capacity Market is going through reform. Demand response’s share of awarded capacity in the most recent T-4 auction surged, reaching near 1,800MW. The government’s consultation response in late 2025 set out reforms to improve how consumer-led flexibility is treated, better reflecting the value of demand response during system stress events and creating cleaner technology classifications. There’s still work to do on the de-rating methodology, but the direction of travel is clear: demand response is being taken seriously as a structural part of the capacity mix, not just a marginal contributor.

Q: GridBeyond operates across the UK, Ireland, Australia, Japan, and the US. Are the market dynamics quite different in each region, or is there a common thread running through all of them?

There’s a powerful common thread, which is that every grid we operate in is going through the same fundamental transition: more variable renewable generation, more electrification-driven load growth, and infrastructure that wasn’t built with this combination in mind. That creates a consistent underlying demand for flexibility services.

But the market structures differ significantly, and that matters for how you participate. In Great Britain, you have a well-developed set of services that provide multiple revenue stacking opportunities for aggregated assets. In Australia, price volatility can be extraordinary, and the combination of rooftop solar penetration and ageing baseload retirements creates genuine scarcity dynamics. Japan is earlier on the journey but moving quickly, with the energy transition creating real urgency around flexibility in a market that has traditionally been very supply-side focused.

The US is more complex because you have multiple ISOs and RTOs with quite different rules, and the political backdrop at the federal level has created some uncertainty. But the load growth driving change is structurally real regardless of federal energy policy.

Q: How is AI changing your own platform and what GridBeyond can do for customers?

We’ve always been an AI-first business; our origins are in using machine learning to predict asset behaviour and optimise dispatch decisions in real time. But the capability has stepped up considerably. The step-change in the past 12 to 18 months has been in what I’d call predictive commercial intelligence; not just optimising individual assets, but building a dynamic picture of how a portfolio of assets can behave across multiple markets simultaneously, accounting for forecast uncertainty, asset constraints, market price signals, and customer operational parameters all at once.

For a manufacturing business, that means the platform can identify that, in four hours’ time, balancing prices are likely to spike, that the site has a production window that can be shifted by 45 minutes without operational penalty, and that doing so would generate £X in revenue while reducing the site’s peak demand charge. That decision-making used to require a specialist energy trader with deep market knowledge. Now it happens automatically, continuously, and at a granularity no human could replicate across a large portfolio.

Our FlexPilot service creates a virtual replica of a site’s energy consumption, identifies the interplay between production schedules and energy costs, and generates real-time recommendations or automated actions. In cement manufacturing, for instance, we’ve modelled savings in excess of £3.5M annually for a single site. That kind of value proposition moves the conversation well beyond traditional demand response.

Q: What sectors do you see as the biggest untapped opportunity for demand response right now?

Beyond data centres, which we’ve discussed, I’d highlight three. Manufacturing is underappreciated. The combination of energy-intensive processes, on-site generation and storage assets, and the increasing pressure to reduce operational costs makes industrial C&I a natural fit for advanced flexibility. FlexPilot is a direct response to that opportunity; the ability to align production scheduling with energy market signals is genuinely transformative for energy-intensive manufacturers.

Cold storage and food distribution is another. These facilities have significant thermal inertia, they can pre-cool within tight temperature bands and then reduce consumption during peak periods without any impact on the product. It’s an almost ideal demand response asset, but many operators haven’t fully monetised it yet.

And increasingly, commercial real estate, retail portfolios, logistics hubs. The proliferation of smart building management systems and sub-metering technology means that assets which were invisible five years ago are now connectable and dispatchable. The reduction in minimum unit size is directly relevant here as it means aggregators can now build viable portfolios from assets that wouldn’t previously have qualified individually.

Q: For a business that’s new to demand response, what’s the single most important thing they need to understand going into 2026?

That the barrier to entry is lower than they think, and the cost of delay is higher than they think. I still speak with businesses that have significant flexible load capacity and haven’t participated in any flexibility markets because they assume it will be operationally disruptive, or that their site isn’t large enough, or that the administration is too complex. In 2019, some of those concerns had merit. In 2026, they don’t; or at least, a good aggregator should be able to manage that.

The cost of delay is that these markets are maturing and the easiest revenue stacking opportunities go to those who prove assets and build track records first. Capacity Market agreements, for instance, reward reliability over time. A business that participates now is building something that has value beyond the immediate payment; it’s establishing a position in markets that are only going to grow in importance.

The wider energy transition is also pushing network charges and wholesale volatility in one direction. Businesses that understand how to work with markets rather than just absorb higher costs are going to have a structural cost advantage over those that don’t. Demand response is, at its best, a business strategy not just an energy product.