Amazon, the global leader in online commerce and cloud services, professes to be an eco-friendly enterprise. Astonishingly, it claimed to achieve its 100% renewable energy objective seven years prior to its self-set deadline. However, a contrasting view depicts Amazon as a substantial polluter, producing extensive greenhouse gases through electricity consumption, surpassing other cloud computing competitors, especially in the USA where it has the majority of its operations and where about 60% of its power generation in 2023 was reliant on fossil fuels.
This dual portrayal of the company – as an environmental saviour or a significant contributor to climate change – stems from the regulations for calculating greenhouse gas emissions. These rules enable organisations to balance their actual energy-related emissions with investments in clean energy initiatives.
This complexity is demonstrated by the social media company Meta, who claims to have already reached ‘net-zero’ emissions for its energy usage. But when the Financial Times analysed its 2023 sustainability report, it was revealed that its real-life CO₂ emissions from power use the previous year were 3.9 million tonnes, a drastic contrast to the 273 net tonnes cited in the report.
As technology companies like these are expected to be significant energy consumers in the future, thanks to the increasing demand for power-intensive artificial intelligence, they are individually exerting influence over the redefinition of the rules that regulate the disclosure of power usage pollution. Companies such as Amazon, Meta and Google have executed campaigns to influence the Greenhouse Gas Protocol, the regulatory body that audits carbon, and have funded research to endorse their positions, as per documents retrieved by the FT.
Despite this collective effort, there isn’t a consensus within the tech world about how these rules should be framed. A conglomeration, which includes companies like Amazon and Meta, proposes a methodology that observers believe may lead to emission reports that may inaccurately represent the entity’s real-life pollution and may not fully recompense for these emissions.
A person closely associated with reform talks described the proposal as attempting to manipulate regulations in order to obscure their undertakings from the public eye. The coalition notes the emphasis on precise emission data and clarity in its approach. Google’s competing proposal, which necessitates corporations to balance out their emissions through power sourced from comparable means, has been condemned by the Amazon coalition and others as too pricey and challenging.
“Every enterprise has different requirements,” Amazon claimed in a statement. “Considering the vast differences between Amazon and Google, our strategies are also divergent.” Meta revealed that its policy involving “market-based emissions” enabled it to balance all its power consumption with renewable sources since 2020, citing a lengthy track record of supporting new green energy projects.
While tech firms invest in green energy, they are incapable of thoroughly regulating how contaminated the energy from the local grid consumed by their data centres is. Currently, energy used by a data centre during the night in heavily coal and gas-dependent locations like Virginia can be neutralised by purchasing a certificate linked to daylight solar energy produced in cleaner grid areas, such as Nevada.
The prevailing structure for declaring greenhouse gas release was established in the 1990s when non-profit organisations, such as the World Resources Institute, set up the Greenhouse Gas Protocol. These Protocol’s directives on carbon reporting are referenced in EU and the upcoming SEC reporting rules for large firms, as well as the Science Based Targets initiative, a company climate goals voluntary monitoring body.
A wind, solar, or hydroelectric plant can issue an energy attribute certificate, typically called a renewable energy certificate or REC in the US, each time it produces a unit of clean energy. These can be incorporated into a clean power contract or purchased individually from a generator or marketplace middlemen.
“Firms can buy RECs to lessen their environmental footprint,” according to the US Department of Energy’s National Renewable Energy Laboratory report. Effective action in supporting clean power finance and guiding investment toward green energy growth is thereby demonstrated.
Firms argue that they aren’t fully in control of the energy grids they’re connected to, and since energy from renewable and non-renewable sources can’t be separated once it’s in the system, certificates seem to be a fair way to compromise and create an incentive for further investment. However, Matthew Brander, a professor at the University of Edinburgh, suggests this is akin to buying the right from a fitter peer to claim you have cycled to work, when in reality, you have travelled using petrol-fuelled transportation.
Several experts voice concerns about the usage of these Renewable Energy Certificates (RECs) in offsetting actual emissions. The rules currently state that these certificates must originate from the same broad geographic region as the pollution they’re offsetting, like Europe or North America, but not the same energy grid and not synchronously. Therefore, the clean energy countering the emissions could be generated in a different nation, at a different moment, or even historically.
For Killian Daly, Executive Director of Energy Tag, a non-profit organisation, it’s preposterous that current accounting standards would allow for someone to claim they are using solar power round the clock. He argues the timing and location of generated energy are both vital in realistically assessing emissions. For instance, a potential buyer connected to a coal-dependent grid might purchase the same certificate as someone on a cleaner grid to offset one megawatt-hour of power use. However, the emissions level from that usage will vary on both grids.
The affordability of these certificates also raises concerns. According to commodity trader STX Group, the average forward price for one US renewable energy certificate for the upcoming calendar year has yet to exceed $5 since as early as 2022. This has led some experts to question if this cost is really enough to encourage the establishment of new renewable energy projects.
Studies conducted by experts at Princeton, Harvard, and the Greenhouse Gas Management Institute indicate that investing in such certificates typically doesn’t promote a new influx of renewables and hasn’t led to a decrease in emissions. Matthew Brander is sceptical of clean energy claims that heavily rely on purchasing certificates associated with power produced at different times and places from where it’s consumed, labelling these as instances of “bad practice”.
However, the system is due for its first review in nearly ten years – an event which could provide the opportunity to rectify these issues. On the other hand, it could also serve to further the interest of large energy consumers to manipulate the system in their favour.
Google has put forth a proposal suggesting that energy consumption should be matched exclusively with renewable energy and green energy certificates from the grids where electricity is consumed. This should be done by taking into account the specific time the electricity is used. Google argues that utilising certificates from a location different to where a company is operating could misrepresent the company’s actual dependence on fossil fuel electricity, without truly addressing the emissions they are directly accountable for.
In its March 2023 submission, Google contends that this approach is beneficial as it encourages collaboration with local decision-makers, promoting discussions on how to transition their electricity grid to more sustainable alternatives. The approach would also support a variety of solutions, including battery technology. Michael Terrell, Google’s senior director for energy and climate, highlights the need to view climate change not merely as an accounting problem, but as a market and technology challenge.
Google’s 24/7 localised approach has captured the interest of Microsoft, the US federal government, and other primary energy consumers, who approve of the shift towards more temporally and geographically distinctive electricity use. However, they have yet to publicly endorse either proposal.
In contrast, the Emissions First Partnership lobby group, comprising Amazon, Meta and others, advocates for more flexibility in the usage of green certificates without geographic restrictions. These companies believe that the certificates should “emulate actual emission reductions”, allowing companies to value the certificates based on the degree of their emissions influence.
Google’s plan is to counterbalance its emissions from energy usage with certificates corresponding to the same area’s power production during the same hour. In opposition to this, the Emissions First Partnership lobby group posits that this power could be offset with certificates tied to renewable energy produced somewhere like Norway during the day.
Thus, this group advocates for companies procuring clean energy certificates from grids like India- more polluted than cleaner ones like Norway’s- to reflect the potential carbon emissions offset by renewable energy use like wind power.
A vital component not taken into account in either of the proposed approaches, according to academics, is “additionality”. There should be effective checks to ascertain that the renewable power would not have been generated regardless of the additional profit from the green certificate sales. This would inspire companies to fund clean power especially in locations where the grid is predominantly polluted.
Advocates contend that this methodology would be less costly, enabling corporations to fund green energy even in nations where foreign investment in energy ventures is restricted. Lee Taylor, CEO of REsurety, a company that markets the data utilised in the Amazon and Meta approach, characterises the Google methodology as “idealistic” and not always “feasible from a financial standpoint”. For instance, if the purchases were limited to solutions on the same grid from where they draw power, it might necessitate investing in a “large-scale battery” for storing energy when a wind farm is inactive. Taylor insists that energy consumers need more choices. “Where can my $10 best reduce carbon emissions?” he questions.
Climate technology investor Jimmy Jia views the competing propositions as two distinct “paradigms of change”, however, he is apprehensive that the Amazon-backed plan might inadvertently “unleash emissions manipulation”. Amazon refutes this claim, asserting that the programme would result in a “more cost-efficient, expedited grid decarbonisation trajectory and enhanced energy fairness”. Meta asserts that the proposal is “the best solution to completely address the repercussions of [its] environmental impact”.
The existing system has been reprimanded for not effectively leading to real-world emission reduction. The Emissions First Partnership was conceived as a potential response to these worries, says Amazon. Skeptics, however, worry that this strategy would make energy certificates resemble carbon credits — disputed devices intended to signify a tonne of CO₂ prevented or eliminated from the environment. One challenge with the carbon credit market is the ability of companies to use theoretical “prevented” pollution to counterbalance actual pollution. Applying that to the REC market would be akin to claiming a cost that was never incurred in financial accounts, according to Brander, essentially creating a “Trojan horse in your greenhouse gas accounts”.
In response, the Emissions First Partnership argued that this accusation was more about shock value than substance, and asserted it did not promote the use of carbon credits to neutralise emissions from power consumption. Max van Meer, US managing director at STX, points out that the stakes are high as large tech groups are already the largest corporate buyers of RECs by a significant margin.
Analysts at Rystad Energy have identified Microsoft and asset manager Brookfield as significant contributors in renewable energy ventures worldwide. A collaboration between the two aims to produce 10.5 gigawatts of power, approximately equivalent to the energy needs of 1.8 million households. The estimated expenditure to create 1GW of fresh capacity is supposedly close to $1 billion.
Online retailing giant Amazon, recognised as the most extensive corporate purchaser of renewable energy, is investing heavily in green energy initiatives, including wind and solar projects in various countries such as India. In pursuit of its 100 per cent renewable energy objective, the company asserted that it achieved the bulk of its target in 2023 through investing in sustainable energy ventures. To make up for the shortfall until fresh renewable projects are operational, it utilises unbundled certificates, albeit predicting a reduction in their use over time.
Social media powerhouse Meta confirmed that it balanaces much of its power consumption with investments in green energy on the same grids as its data centres. The firm has poured investment into over 8GW of operational renewable energy.
Even so, the surge in financial commitment to data centres and computing infrastructure necessary for generative AI could potentially increase power usage, technology companies are banking on this for future sales and profit expansion. The International Energy Agency estimates that by 2026, power usage by data centres will have inflated beyond twofold, comparable to Japan’s current yearly power consumption.
This escalation brings into question the viability of Big Tech’s net-zero targets. Between 2020 and 2023, Microsoft’s emissions saw a steep 30 per cent rise, while Google’s surged by nearly 50 per cent between 2019 and 2023. Both corporations attributed these increases partially to the demand for new data centres. A bulk of this growth is predicted to be in the US, where many grids primarily run on fossil fuels.
In the protocol’s most recent renegotiation, Amazon, Meta, Salesforce, Microsoft and Google, along with Ikea, Cargill, and a number of philanthropic foundations, have revealed to be financial contributors. Some of the funding from these entities was distributed before the beginning of the reform process. Amazon also sponsored research, one of which argued for the flexibility of energy consumers to procure certificates from other countries when operating in challenging markets.
The company has stated that its funding goes towards independent research for expert critique, involving a variety of collaborators, and endorsing diverse perspectives; with the latest funding happening in 2022. Last year, Jeff Bezos’ charitable division, the Bezos Earth Fund, contributed $9.25 million to the protocol, making it one of the primary funders of WRI, the non-profit which shares management responsibilities with the accounting oversight organisation.
In a statement to the Financial Times, the company indicated it had no intention of swaying the results, adding that they support the protocol as it employs the most widely recognised international standard for calculating carbon footprints. High-tech firms were also part of gatherings with the secretariat of the protocol, with the gathering in May confirming representation from corporations such as Amazon, Meta, and Chevron among others, according to the agenda documented by the FT.
During a closed meeting in June, Amazon and Meta sent representatives to NREL to stress on the need for adaptability in certificate purchase rules as reported by an attendee. A report sponsored by Meta and presented by energy consultancy E3 supported this sentiment, with E3 asserting that it was in line with their former stance.
The manager of WRI, Craig Hanson, stated that the protocol has conducted numerous conferences with representatives from public, private and the third sector as part of an extensive reform process. A protocol spokesperson informed the FT that their decision-making process was globally inclusive involving stakeholders from businesses, NGOs, academia and global governments.
The upcoming amendments will be supervised by an autonomous standards board and funders won’t have any privileged access, added the spokesperson. They also mentioned that their legal advisor was contemplating the introduction of a “cooling-off period” for previous benefactors. Though the finalisation of the rules isn’t expected till 2026, the consequences could be wide-ranging as the protocol is also contemplating the extent of permissible offsetting by companies when they calculate other direct and indirect emissions.
While these discussions are ongoing, activist groups are stressing on the pivotal role of the protocol in tackling climate change. Laura Kelly, who has scrutinised the carbon accounts of Big Tech for an Australian pressure group Action Speaks Louder, expressed her apprehension about the significance of the protocol.
The achievement of Paris [climate] objectives fundamentally relies on tackling industry emissions, given their significant contribution to global emissions – as reported by The Financial Times.