Google’s environmental report hints at enterprise cost-saving tactics

Google touted some of its energy efficiency accomplishments in its annual report on its sustainability efforts published Tuesday, but also highlighted the many pros and cons to be weighed in deciding actions, especially around generative AI. 

The high-end systems required to run the latest generative AI algorithms require a vast amount of energy and water for cooling, but they could potentially be used to make discoveries that reduce their environmental footprint. For now, there are so many variables that it’s unclear whether generative AI will ultimately help or hurt the planet.

[ Learn why AI is a wild card in data center energy demand and how AI networking adds to data center sustainability challenges | Sign up for Network World newsletters.]

There’s also the matter of hard cash: The cost of running those AI applications may exceed the value they generate.

“We know that scaling AI and using it to accelerate climate action is just as crucial as addressing the environmental impact associated with it. In spite of the progress we are making, we face significant challenges that we’re actively working through. In 2023, our total GHG (global greenhouse gas) emissions increased 13% year-over-year, primarily driven by increased data center energy consumption and supply chain emissions,” Google said in a blog post summarizing the report.

Google advanced the use of clean energy in many of the areas where it operates, it said, but “there are still some hard-to-decarbonize regions like Asia-Pacific where CFE (carbon-free energy) isn’t readily available. In addition, we often see longer lead times between initial investments and construction of clean energy projects and the resulting GHG reductions from them.”

Aging equipment

There are other matters to be weighed here. Consider equipment age and when systems should be replaced. Google said in its 86-page report that its efforts to reduce operational waste and lessen its burden on landfills prompted it to “maintain servers for as long as possible by refurbishing, reusing, or reselling components, and we work to ensure device longevity.”

The flip side of this approach is that newer equipment can often be far more energy efficient, depending of course on the age and specs of the equipment being replaced. Therefore, is keeping old systems longer good or bad for the environment?

Google said it continued to match its global energy consumption with renewable energy production in 2023, even as its data center electricity consumption grew 17%, and its total greenhouse gas emissions increased by 13%.

“We see our growing infrastructure as an opportunity to drive the innovations and investments needed to power a low-carbon economy,” it said.

Google listed a variety of ways it claims to be improving sustainability and power/water efficiency, but provided insufficient detail to enough to help other enterprises replicate its tactics. Many of the measures read as sales pitches for its own products and services, but there were a few ideas that companies could consider. 

Training AI models for less

“We’ve identified tested practices that our research shows can, when used together, reduce the energy required to train an AI model by up to 100 times and reduce associated emissions by up to 1,000 times, which are all used at Google today,” the report noted. “We’ve sped up AI model training through techniques like quantization, boosting large-language model training efficiency by 39%.”

As for the custom tensor processing unit (TPU) chips it has designed for applications like this, “Our TPU v4 was 2.7 times more energy efficient than our TPU v3 and we’ll soon offer Nvidia’s next-generation Blackwell GPU to Cloud customers, which Nvidia estimates will train large models using 75% less power than older GPUs to complete the same task,” Google said. “Additionally, our new Google Axion Processors are up to 60% more energy efficient than comparable current-generation x86-based instances. These advancements, including AI-powered optimizations like AlphaZero, show how we’re constantly improving hardware efficiency.”

Water usage is another area where enterprises can try to realize savings.

“The expansion of AI products and services is leading to an increase in data center workloads and the associated water footprint required to cool them efficiently,” Google said. “In 2023, our data centers consumed 6.1 billion gallons of water — 17% more water than the previous year, mirroring similar growth in electricity use.”

The report noted fuel efficiencies IT can deliver outside the data center — such as by leveraging the latest mapping technologies suggesting fuel-efficient routes. Google suggested Google Maps, but other mapping products could deliver similar benefits. 

“By building AI models on the emissions profile of different vehicle types, fuel-efficient routing in Google Maps analyzes traffic, terrain, and the vehicle’s engine (gas/petrol, diesel, hybrid, or electric) to find the most efficient route. This may mean fewer stops for gas engines, routes favoring highway speeds for diesel vehicles, and maximizing downhill stretches for electric cars to boost regenerative braking—all while providing the same or similar ETA,” the Google report said. “In 2023 alone, we estimate that fuel-efficient routing enabled more than 1.7 million metric tons of GHG emissions reductions — equivalent to taking approximately 380,000 fuel-based cars off the road for a year.”

Scoping out emissions

In comparison with those savings, Google’s total GHG emissions in 2023 were approximately 14.3 million metric tons of CO2 equivalent, including its Scope 1, 2, and 3 emissions. (Scope 1 emissions of 79,400 million metric tons came predominantly fuel to power data center generators and from refrigerant leakage, it said; electricity to power its data centers and offices accounted for the majority of the 3.4 million metric tons of Scope 2 emissions, while the biggest contributors to the 10.8 million metric tons of Scope 3 emissions were purchased goods and services and capital goods.)