Connect with us

Accounting

Carbon accounting obscure AI’s footprint

Published

on

An Amazon Web Services data center in Ashburn, Virginia
An Amazon Web Services data center in Ashburn, Virginia, in 2024

Nathan Howard/Bloomberg

Tech companies’ relentless push into artificial intelligence is coming at an undisclosed cost to the planet. Amazon, Microsoft and Meta are concealing their actual carbon footprints, buying credits tied to electricity use that inaccurately erase millions of tons of planet-warming emissions from their carbon accounts, a Bloomberg Green analysis finds.

Recently Microsoft reported that its emissions are 30% higher today than in 2020, when it set a goal to become carbon negative. Other tech companies’ emissions are rising, too. However, Microsoft and other AI leaders insist that the increase is because of the carbon-intensive materials used to build data centers — cement, steel and microchips — and not because of the massive amount of energy AI requires. That’s because they have said the power is mostly or all from zero-carbon sources, such as solar and wind.

Is AI being powered exclusively by clean energy? “There is no physical reality for that claim,” said Michael Gillenwater, executive director of the Greenhouse Gas Management Institute.

Companies are buying credits — called unbundled renewable energy certificates — that can make it seem that power consumed from a coal plant came from a solar farm instead. Amazon, Microsoft and Meta rely on millions of unbundled RECs each year to claim emission reductions when making voluntary disclosures to CDP, a nonprofit that runs a global environmental reporting system.

The current carbon accounting rules allow for the use of these credits for calculating a company’s carbon footprint. However, work that many academics have done shows the accounting rules need to be updated in order to accurately reflect greenhouse-gas emissions.

That’s because these carbon savings on paper are not actual emissions reductions in the atmosphere. If companies didn’t count unbundled RECs, Amazon could be forced to admit that its 2022 emissions are 8.5 million metric tons of CO2 higher than reported — that’s three times what the company disclosed and matches Mozambique’s annual impact. Microsoft’s sum could be 3.3 million tons higher than the reported tally of 288,000 tons. And Meta’s reported footprint could grow by 740,000 tons from near zero. (See below for methodological details.)

“Companies shouldn’t be allowed to use unbundled RECs to claim emissions reductions,” said Silke Mooldijk, who focuses on corporate climate responsibility at the nonprofit NewClimate Institute. “It’s misleading to consumers and investors.”

Not all tech companies have gobbled up unbundled RECs to obscure the rising emissions that have resulted from the hotly-contested AI race. Alphabet Inc.’s Google phased out its use of unbundled RECs several years ago after acknowledging that it doesn’t amount to real emissions reductions. “Studies have raised legitimate questions about whether [these credits] displace fossil-powered generation,” said Michael Terrell, senior director of energy and climate at Google.

Amazon relied on unbundled RECs for 52% of its renewable energy in 2022, making it the most dependent of the four on the instruments. A spokesperson for Amazon said the number of unbundled RECs the company uses is expected “to decrease over time” as more of its directly contracted renewable energy projects come online. Microsoft, which relied on unbundled RECs for 51% of its renewable energy, also plans “to phase out the use of unbundled RECs in future years,” a company spokesperson said.

A spokesperson for Meta, which relied on unbundled RECs and power from utilities labeled “green” for 18% of its renewable energy, said the company takes “a thoughtful approach” and that the “majority” of the company’s “renewable energy efforts” are focused on projects that “would not have otherwise been built.” 

The thousands of companies using Amazon-powered AI for their customer chat bots, Microsoft’s AI Copilot for summarizing meetings, or Meta’s Llama for generating images may assume there are few or no energy emissions from relying on these models. It’s a powerful marketing tool for these big tech companies, helping to allay concerns of potential customers who are themselves likely under pressure from users and investors to lower their own carbon footprints. In reality, it’s creating a cascading impact of misreported emissions and growing demand for energy-intensive AI products.

“If consumers do not understand what the climate impact of AI is, because tech companies do not transparently report on it, then there’s no incentive for consumers to change their behavior and change to a different AI model,” said Mooldijk.

It’s a concern across finance, too. Banks and investors which tend to stuff big tech in sustainable funds too often take emissions claims at face value. “At the moment, there’s just not a sophisticated understanding of this issue,” said Gerard Pieters, a director at Tierra Underwriting that helps banks on clean-energy deals. “We’re still in a period where people make claims quite easily and they’re just copied and accepted as fact.”

Tech companies are the largest buyers of unbundled RECs in the world. Whether or not they continue buying these credits to make climate claims matters a great deal as more corporations look to cut their carbon footprint and green their credentials.

Back to the source

To understand how the companies’ use of RECs works, consider the origins of the power generated on a grid. Usually it comes from a mix of sources: from coal and gas to wind and solar. Climate-conscious companies are increasingly looking to secure power exclusively from sources that generate the least planet-warming emissions.

One way to do this is to sign a contract for clean power directly with the supplier through a power-purchase agreement, where a tech company is signing a long-term contract and thus taking on some of the risk for a period of 10 or 15 years. That, in turn, makes it easier for the developer to acquire the financing to build the solar or wind farm.

To help tech companies trace the source of that power, renewable-energy producers also issue energy attribute certificates, or RECs, which are a type of tracking instrument. However, RECs can also be bought on their own, separate from an electricity purchase. The idea behind these so-called ‘unbundled’ RECs is that there’s value in renewable energy generation beyond simply the electrons produced and sold — its lack of emissions also has a value. So since renewable energy generators produce two things of value — energy and, specifically, low-emissions energy — they should be able to get paid not just for producing electricity but also for being green.

This idea — and the calculation that sprang from it — was developed when renewable energy was expensive to produce and not price-competitive with fossil fuels. The thinking was that the extra money renewable energy developers would receive in the form of a REC might work as an incentive to produce more wind and solar development than would have been otherwise and thus be “additional.”

Studies as far back as 2010 showed that unbundled RECs weren’t delivering on that theory of stimulating the production of renewables. But that inconvenient fact was mostly ignored, and the enthusiasm for RECs led to a quirk in emissions reporting rules that allows companies to buy unbundled RECs and then deduct the emissions from their CO2 accounts. This means companies can report reduced emissions from their electricity use even if their actual use has not changed in any way (and may still come from a coal power plant).

Solar and wind power have now become cheaper than the fossil-fuel alternative, and a growing body of evidence shows that most unbundled RECs aren’t what those who count emissions call “additional.” That is, they don’t spur new wind or solar farms and thus there is no second value producers should be paid for, and certainly no emissions reductions for the buyer.

“The widespread use of RECs … allows companies to report on emissions reductions that are not real,” Anders Bjorn, assistant professor at the Technical University of Denmark, and a team of researchers, wrote in a paper published in the scientific journal Nature in June 2022. After adjusting for companies’ use of RECs, they found 40% no longer showed alignment of their activities with the Paris Agreement goal of keeping global warming to within 1.5C.

Last month, Amazon claimed that it had reached 100% renewable energy use in 2023 using its own accounting methodology and thus will have no emissions from electricity use. The company has not yet reported the details underpinning its 2023 renewable energy consumption, but Bloomberg Green’s analysis suggests that the claim likely relies on the use of unbundled RECs. In response, an Amazon spokesperson said, “It can take several years for the projects we invest in to come online, so we sometimes utilize unbundled RECs — a fundamental part of the global renewable energy market — to temporarily bridge the gap to a project’s operational date.”

Like Amazon, Google claims to be 100% renewable powered on an annual global basis. In lieu of using unbundled RECs, Google purchases more clean energy than it consumes in some places, like Europe, and less in others, like Asia-Pacific, depending on the availability in those locations. Google, however, makes clear that it does not consume carbon-free energy on an hourly and location-specific basis. That’s now “our ultimate goal,” said Terrell.

Amazon, Microsoft, Meta and Google are following the accounting rules set out under the Greenhouse Gas Protocol that was first developed in 2001. Those disclosures underpin the analyses that investors rely on to make decisions about what counts as a green company. While the protocol has received small updates over the years, it’s due for a big update and experts are working to propose changes. All the big tech companies are now involved in lobbying on those changes.

“Standards need to evolve, because measuring carbon emissions isn’t an exact science,” said Google’s Terrell. “It’s continuing to improve and we’re committed to helping improve it.”

Continue Reading

Accounting

Continuous Auditing Transforms Corporate ERPs

Published

on

continuous auditing transforms corporate erps

As corporate accounting departments cross the threshold into late July 2026, the adoption of continuous, automated auditing systems has reached a definitive turning point. Driven by advances in artificial intelligence and deep integration with modern Enterprise Resource Planning (ERP) platforms, leading finance organizations are moving away from traditional, periodic post-hoc audits in favor of real-time, 100% transactional verification. This technological transition is redefining internal control environments, reducing compliance costs, and eliminating the structural delays inherent in legacy quarterly closing processes.

Unlike traditional auditing frameworks that rely on statistical sampling—a process that inevitably leaves operational blind spots—continuous auditing software monitors operational data feeds continuously. Every purchase order, electronic invoice, payroll disbursement, and cross-border wire transfer is automatically cross-referenced against established corporate governance parameters, regulatory tax schedules, and anti-fraud algorithms in real time. Anomalies or unauthorized ledger entries are flagged instantly, allowing internal audit teams to investigate and remediate compliance gaps immediately rather than months after the close of a financial period.

The implications for executive financial management are far-reaching. By embedding continuous verification directly into daily transaction workflows, chief financial officers gain uninterrupted visibility into the organization’s true financial standing. Real-time balance sheet auditing eliminates the severe operational bottlenecks associated with month-end and quarter-end financial reconciliations, freeing accounting professionals to focus on strategic financial modeling, tax planning, and capital allocation rather than manual data entry and spreadsheet consolidation.

However, implementing continuous auditing requires accounting leadership to invest heavily in data governance and technical upskilling. Internal audit teams must evolve from manual ledger reviewers into system architects capable of auditing complex algorithms and validating automated data pipelines. Accounting firms and corporate controllers that master continuous auditing will establish a resilient compliance framework capable of meeting stringent international regulatory standards with total transparency.

Continue Reading

Accounting

U.S. Imposes New 50% Tariffs on Canadian Imports Under Rare Legal Provision

Published

on

U.S. Imposes New 50% Tariffs on Canadian Imports Under Rare Legal Provision

WASHINGTON — In a major escalation of cross-border trade friction, U.S. President Donald Trump has signed executive orders imposing new 50% tariffs on a wide selection of Canadian exports, citing discriminatory practices by Ottawa targeting American auto, dairy, and beverage industries.

The new duties, announced Monday, will take effect in 30 days. They target a broad spectrum of consumer and industrial goods—ranging from wine, liquor, and milk products to commercial cement, furniture, clothing, and hockey equipment.

Untested Legal Mechanism

To enact the sweeping measures, the administration invoked Section 338 of the Tariff Act of 1930—a rarely used legal provision allowing the executive branch to levy additional tariffs of up to 50% on foreign nations deemed to discriminate against U.S. commerce.

White House officials noted that Section 338 addresses trade discrimination rather than national security or economic emergencies. The move comes months after prior global emergency tariffs faced legal challenges in domestic courts, signaling Washington’s pivot toward alternate statutory authorities to maintain import duties.

Senior administration officials briefed reporters that the measure directly responds to Canadian provincial bans on U.S. alcohol, restrictions on American vehicle exports, and import quota disparities affecting U.S. dairy and cheese producers relative to third-party trading partners.

“While the administration continues to secure reciprocal trade agreements globally, Canada retaliated against efforts to protect domestic industry,” U.S. Trade Representative Jamieson Greer stated.

USMCA Impact and Carve-Outs

Significantly, the newly ordered 50% duties will apply to designated items even if they otherwise comply with the United States-Mexico-Canada Agreement (USMCA).

However, the administration confirmed key targeted exemptions:

  • Energy products (including oil and natural gas)
  • Potash and critical minerals
  • Fish and seafood
  • Goods already governed by sector-specific duties (such as existing steel and aluminum tariffs)

Administration representatives emphasized that the tariffs do not stem from recent disputes concerning drifting Canadian wildfire smoke, noting that policy options regarding environmental spillover remain under separate review.

Canadian Response and Market Reaction

Following the White House announcement, the Canadian dollar experienced a sharp decline against the U.S. dollar, falling approximately 0.4% during evening trading.

Canadian Prime Minister Mark Carney issued a statement emphasizing that Canada’s earlier counter-duties had merely matched previous U.S. trade actions. “Canada stands ready to engage intensively to address outstanding issues with the U.S. to the mutual benefit of our citizens,” Carney stated, pointing to detailed proposals Ottawa submitted to modernize the USMCA framework.

Ontario Premier Doug Ford took a firmer stance, urging a “dollar-for-dollar” reciprocal response if the measures go into effect on August 19.

With a 30-day implementation window before the duties officially lock in, industry associations and trade groups on both sides of the border are calling for urgent bilateral negotiations to avert further supply chain disruption across North America.

Continue Reading

Accounting

Automated Continuous Auditing: Transforming Compliance and Real-Time Financial Oversight

Published

on

Transforming Compliance and Real-Time Financial Oversight

The traditional accounting paradigm—defined by periodic monthly closures and post-hoc annual audits—is rapidly giving way to continuous, automated financial oversight. As of July 2026, forward-thinking accounting practices and multinational corporate finance departments are leveraging continuous auditing systems powered by advanced machine learning models. These systems monitor operational transactions in real time, shifting audit methodologies from sample-based post-analysis to absolute, 100% transaction-level verification.

The operational advantages of continuous auditing are transformative. Standard auditing procedures historically relied on statistical sampling, which, despite rigorous methodology, inherently left gaps where anomalies or fraudulent transactions could go undetected for months. Modern continuous auditing platforms integrate directly with enterprise resource planning (ERP) databases, instantly cross-referencing purchase orders, invoices, bank feeds, and tax records. Any deviation from established control parameters or unusual transaction behavior triggers immediate flags for internal audit teams, dramatically reducing detection lag from quarters to seconds.

Beyond fraud prevention, continuous auditing fundamentally alters internal reporting and decision-making. Executive leadership no longer has to wait weeks after the close of a quarter to evaluate precise financial standing; real-time verified ledger data provides an uninterrupted view of operating margins, tax liabilities, and cash flow dynamics. This real-time visibility enables corporate controllers to adjust capital allocation strategies dynamically, mitigating liquidity constraints and capitalizing on emerging commercial opportunities far more efficiently than competitors bound to legacy reporting cycles.

However, implementing continuous auditing requires accounting professionals to acquire new analytical capabilities. The role of the auditor is evolving from manual data reconciliation toward system validation, algorithmic model governance, and strategic risk interpretation. Accounting firms and corporate finance departments must invest in continuous technical education, ensuring that audit staff possess the data engineering skills necessary to design, maintain, and evaluate complex automated compliance systems.

Continue Reading

Trending