Data centre energy pledge unlikely to benefit solar PV

Trump had previously mentioned the Bring Your Own Power (BYOP) requirement, more particularly during his State of the Union speech on 24 February 2026, which would require companies to pay for the energy needed for their own data centres and avoid an increase in electricity costs for consumers.

Seven tech companies have already signed the pledge, including Meta, Microsoft, Oracle, xAI, and the aforementioned Amazon, Google, and OpenAI.

Under the Ratepayer Protection Pledge, the seven companies will negotiate separate rate structures with utilities and State governments.

Furthermore, they will commit to paying the rates for the power and related infrastructure brought online to service their data centres, whether they use the electricity or not.

On top of that, the companies that have signed the pledge will also coordinate with grid operators to make backup generation resources available, allowing for a more resilient grid and preventing power shortages or blackouts in cases of emergency.

One of the companies that signed the pledge, Amazon, welcomed the administration’s ‘leadership’ on the issue and the pledge’s commitment to establish “an important baseline that will protect ratepayers and enable responsible, long-term energy partnerships that strengthen the grid and communities where data centres operate,” said Matt Garman, CEO of Amazon Web Services.

Solar PV unlikely to benefit from the pledge

Solar PV is unlikely to benefit from this pledge, as shown by previous initiatives in the PJM market, explained Oliver Kerr, North America managing director at Aurora Energy Research.

“The biggest risk is that policy interventions distort prices and undermine competitive market dynamics – we’ve already seen how initiatives to fast-track new capacity in markets like in PJM can be unfavourable for solar and storage, for instance,” added Kerr.

The role of energy storage might, on the other hand, be more favourable than solar PV, given its applications.

“To the degree that data centres are willing and able to pay for a combination of ‘speed-to-power’ and 24/7 reliability, regulations that push for data centres to pay for their own energy infrastructure are especially bullish for storage, which provides both. The picture is more mixed for solar, which is faster to develop than large gas plants but not necessarily faster than on-site gas, and has more dispatchability challenges even with batteries than gas,” said Flakoll.

However, Garman added that more needs to be done to modernise the grid, highlighting “outdated permitting processes and bureaucratic delays” that are slowing critical energy and infrastructure progress, which, according to him, leads to “rising utility bills and placing growth and US competitiveness at risk.”

“Regulators at the state, regional, and increasingly federal levels are converging on making data centres pay their ‘fair share’ of power infrastructure costs with the aim of preventing cost shifts to ordinary consumers. Trump’s ratepayer protection pledge is non-binding and largely confirms a consensus that has emerged among policymakers independent of Trump, but the publicity may maintain pressure on hyperscalers to commit to paying,” said Derrick Flakoll, senior policy associate, BloombergNEF.

Flakoll added that most hyperscalers had already shown interest in paying the power plants required to operate their data centres in order to secure faster grid connections.

Can the pledge actually help prevent bills from rising?

“The intention behind the pledge is a good one; how effective at protecting ratepayers from rising bills will depend on how it’s implemented,” Kerr said. “We see particular promise in new tariff structures for large loads.”

Kerr added that the new tariff structure is to ensure that cost-allocation better reflects cost-causation.

“This can include additional ‘margins’ for large load that can be used to cross-subsidise low income households (e.g. Georgia Power), ‘take-or-pay’ tariffs that require large load to cover the costs of requested capacity even if they don’t end up using it (e.g. in Ohio), or structures that pass on price spikes to large load to incentivise demand response.”

The effectiveness of the pledge to prevent customers from seeing their bills increase, as mentioned by Kerr, is something that is yet to be demonstrated.

Flakoll mentioned that cost shifts from industrial customers to residential consumers can be extremely subtle. “And some observers like former FERC Chair Mark Christie have criticised early policies by FERC and various PUCs as insufficient to prevent those cost shifts.”