For many years, Intel was an American success story: the most valuable chipmaker in the world and the undisputed leader in the manufacture of microprocessors, the brains inside personal computers and internet servers. 

Its sustainability ambitions led the industry, as well.   

For nearly a decade, Intel bought more clean power than any other U.S. company. It was also an early leader in tackling the powerful greenhouse gasses released in the chipmaking process and known for setting ambitious emissions targets. As a percent of revenue, the company’s emissions peaked in 2006 and remain far below other major manufacturers. 

Then Intel’s business faltered, and so did its climate action. 

Over the past 15 years, the semiconductor behemoth has foundered as rivals captured new markets for chips used in smartphones and artificial intelligence data centers. A strategic misstep left its chips a generation behind the state of the art. Intel’s sales and share price plummeted, forcing it to lay off a quarter of its workforce this year. 

This profile — the latest installment in Chasing Net Zero, our company-by-company series that has probed the climate strategies of Nestlé, Salesforce and others — reveals how business storms have eroded Intel’s once-lauded climate strategy. A critical goal on supply-chain emissions was quietly pushed back 20 years under its new CEO, for instance, and the company no longer ties executive compensation to climate performance. 

Intel declined to make executives available for this article, but said in a statement that it is “committed to achieving a more sustainable future by advancing bold, measurable goals.” Analysts, however, say its progress towards those commitments has slowed as it fights for survival.

“I wouldn’t say that Intel has stopped caring about sustainability, but they have more pressing concerns right now,” said Stephen Russell, a senior technical fellow for sustainability at Techinsights, an Ottawa semiconductor consulting firm. “They’ve fallen behind — and they are desperately trying to catch up.”

When Intel led on climate

Intel’s climate efforts began in the late 1990s when, under pressure from the Environmental Protection Agency, the company joined an industry group to scale back its emissions of fluorinated gases. These “F-gases,” used to carve microscopic circuits onto the surface of silicon wafers, have thousands of times the warming power of carbon dioxide when leaked into the atmosphere.

In the following decade, the company also turned its attention to the electricity consumed by the machinery used to squeeze millions of transistors onto chips the size of postage stamps. Using renewable energy certificates (RECs), which allow companies to claim credit for using green energy by funneling money to renewables projects, Intel became the nation’s largest buyer of clean power between 2008 and 2016, according to the EPA.

These initiatives helped Intel’s operational emissions — Scope 1 (mainly F-gases) and Scope 2 (the electricity it buys) — peak at 4 million metric tons of carbon dioxide equivalent in 2006. By 2019, emissions had declined to 2.8 million tons, even as the company’s revenue doubled. That same year, it stepped up its ambition with goals to use 100 percent renewables and shave 10 percent more off operational emissions by 2030.

As with most companies, however, these sources are dwarfed by emissions from Intel’s value chain. 

Intel’s emissions in its 2019 baseline year

In 2019, the electricity used to operate chips sold by Intel (downstream Scope 3) accounted for two-thirds of the company’s footprint.

Chipmakers tackle these emissions by designing less power-hungry chips — Intel’s target is a tenfold efficiency improvement by 2030. But making chips more efficient encourages greater use, which drives up emissions. The trend is compounded by demand for chips to power AI applications. And the ultimate solution to Intel’s downstream Scope 3 emissions is the decarbonization of global grids — not something within the company’s control.   

One area of Scope 3 where the company has more influence is emissions generated by its suppliers, which represent a quarter of Intel’s footprint. In a target announced in 2022, the company said it would reduce supplier emissions by 30 percent by 2030 “from what they would be in the absence of action.” 

The upgraded near-term commitments were capped off with a long-term goal to reduce Scope 1 and 2 emissions to zero in 2040. A few years later, after activist shareholders asked for more, Intel published a detailed climate transition action plan and committed to a net-zero upstream supply chain by 2050.

These are meaningful goals, particularly as the semiconductor industry has become a major contributor to climate change. Each new generation of chips requires more electricity and F-gases to manufacture, and by 2030 chipmaking is expected to account for 0.5 percent of global emissions.

These goals compare favorably to Intel’s two main competitors in processor manufacturing, fast-growing Taiwan Semiconductor Manufacturing Company (TSMC) and the semiconductor division of consumer electronics giant Samsung. Both target net-zero value chains by 2050, a decade later than Intel’s operational emissions goal. Samsung lacks a near-term target, and TSMC only added one this year: a commitment to return its Scope 1 and 2 emissions to 2020 levels by 2030. 

Most other semiconductor companies, such as Nvidia, the high-flying AI chip maker, are “fabless,” meaning they outsource manufacturing and can claim relatively low Scope 1 and 2 emissions as a result.

How Intel stalled on emissions

With operational emissions falling and its commitments gaining in ambition, the Intel of five years ago had one of the most ambitious climate programs in the semiconductor industry. But the company was also grappling with the consequences of a bad bet made the previous decade.

As they considered how to etch ever-smaller circuits onto the next generation of chips, Intel’s engineers had deemed one new technology — using extreme ultraviolet light to draw finer lines on silicon — too difficult to work with. They opted instead to adapt existing methods, an approach that proved even more challenging to implement. Intel’s next-gen chips, due in 2015, were delayed for years. Meanwhile, TSMC successfully brought its next-gen chip to market using ultraviolet technology.

Intel’s delay meant it didn’t build new fabs with the latest approaches to capturing F-gases as rapidly as its rivals. Eventually, Intel bought its own ultraviolet machines, but other problems at its fabs caused a relatively high number of its chips to fail, raising its emissions per chip.  

“If you are throwing away 50 percent of your chips, you still have all the emissions but you are throwing away half of your profits,” said Techinsights’ Russell.

These factors help explain why rivals, which have been building new fabs, have seen their Scope 1 intensity — tons of emissions per dollar of revenue — fall: Samsung’s has dropped by a third since 2019, and TSMC’s by two-thirds. Intel’s intensity, by contrast, is up 56 percent.

Chipmakers’ Scope 1 emissions intensities

An Intel spokesman said that due to the complexities of semiconductor manufacturing technology, emissions intensity by revenue doesn’t accurately represent its performance.

On close examination, Intel’s most recent Scope 1 intensity may be even worse because it does not manufacture all the chips it sells. Two years ago, the company realized it couldn’t manufacture its most advanced designs and outsourced some fabrication to TSMC. Intel reports the emissions associated with making those chips the same way fabless companies do, as purchased goods and services in Scope 3. 

Intel has shared limited details about the deal, but has said it outsources the fabrication of about 30 percent of its chips. If what Intel reported for Scope 1 in 2024 then represents only 70 percent o