Every day, a deluge of data gets created and stored on our phones, the cloud, or data centres. As the world goes digital, finding sustainable ways to store data becomes critical.
Traditional data centres consume massive amounts of energy and resources while powering everything from social media to AI.
But what if we told you that plant DNA might hold the key to this problem? We wouldn’t then need to rely on silicon chips and data centres.
American-Hawaiian researchers Keolu Fox and Cliff Kapono suggest that the genetic material of plants like sugarcane could serve as eco-friendly data storage units. This idea combines traditional ecological knowledge with data science.
Since their TED talk in 2024, the duo has been actively working on projects that investigate the idea of biological data centres. A major focus area has been ‘Earth-friendly computation’, a concept they co-developed with fellow genome scientist Eric Dawson, Fox told AIM.
This vision encapsulates the establishment of decentralised, climate-resilient data centres that utilise renewable energy sources like solar, wind, saltwater, and hydroelectric power. It focuses on implementing these systems on sovereign Indigenous lands.
Fox highlighted two key projects to AIM: the first – Breathing Life Back into Zombie GPUs – repurposes discarded but functional GPUs to reduce electronic waste and extend hardware lifespans, cutting carbon emissions from AI and machine learning.
The second project – Green Rush Prospector – is a GIS-based app that helps plan sustainable data centre locations. It overlays geographic data on indigenous reservations, factoring in renewable energy, biodiversity, and topography, to ensure tech development aligns with environmental preservation.
When asked if they have conducted any physical tests to determine the success of plant DNAs, Fox said, “We’ve run initial tests using the GIS prospecting app to map out optimal locations for small, self-sustaining data-processing units, which we refer to as ‘data terrariums’.”
Fox also commented on the reduction in the environmental footprint of biological data centres as compared to traditional ones.
“I believe, biological data centres, coupled with innovations like repurposing ‘zombie GPUs’ and using natural ecosystems as part of the data infrastructure, have the potential to significantly reduce the environmental footprint of the growing data centre industry.”
In another development, Microsoft, in collaboration with the University of Washington, has demonstrated the first fully automated DNA data storage system.
Climate Crisis and AI Data Centres
As AI becomes a key player in everything from personalised ads to medical diagnostics, the data centres powering this intelligence are contributing massively to global emissions.
As Fox said, over 50 billion IoT devices will be online by 2025, producing a staggering 80 billion terabytes of data.
A new iPhone can hold about one terabyte. That gives us data worth 80 billion iPhones. Storing such data is not just a matter of free space—it’s a matter of energy.
But that is not all the data there is. We also have tons of data stored digitally on our clouds. Kapono explains that every gigabyte of data stored in the cloud needs up to 7 kWh of energy.
This could mean that streaming a 4K video for just two minutes uses the same amount of energy as your phone uses in a year.
“To keep pace with the amount of data we’re creating, the entire surface of the planet would need to be covered with data storage centres by 2060,” according to Emily Leproust, the CEO and co-founder of Twist Bioscience, in her article published by Real Clear Science.
AI data centres are constantly fighting heat. Just like air conditioners, they rely on energy-consuming cooling systems to prevent overheating. Cooling alone accounts for about 40% of their energy usage, which is why their environmental impact is already bigger than that of the entire airline industry.
As previously reported by AIM, in America, the average water withdrawal for electricity generation is about 43.8 litres per kWh. AI is becoming a major player in climate change, and the stats are concerning.
The data industry currently consumes about 200 terawatt-hours of energy every year, responsible for 2% of global CO2 emissions. But with the rise of AI, this number is expected to explode. In fact, the International Energy Agency (IEA) predicts that by 2030, AI data centres will gobble up 4.5% of the world’s energy production.
In their Stated Policies Scenario, global electricity demand is set to rise by 6,750 terawatt-hours (TWh) by 2030, which is more than the current combined usage of the United States and the European Union.
“The sector has already surpassed 10% of electricity consumption in at least five US states. In Ireland, it now accounts for over 20% of all electricity consumption,” as predicted by the International Energy Agency (IEA)
Tech giants are already feeling the heat. Google, for example, has seen its greenhouse gas emissions rise by 48% since 2019, thanks to its growing reliance on AI-powered data centres. A single query to an AI model requires a lot more electricity than your average search engine query.
If we multiply that by millions of users and trillions of chats, we’re looking at a giant environmental footprint.
The Solution of Evolution
Recent studies highlight DNA’s remarkable potential for data storage. Studies show that one gram of DNA can store up to 215 petabytes of data, which is much more than the current silicon-based storage technologies.
This efficiency is due to DNA’s ability to encode information in a compact and durable format, making it ideal for the growing demands of digital data storage.
DNA can store up to 215 petabytes (215 million gigabytes) of data in a single molecule.
Here's how much DNA is inside of a human body. We are gigantic storage of data: pic.twitter.com/DEzbvTsS4L— Slava Bobrov (@slava__bobrov) December 23, 2021
“A DNA hard drive the size of a teaspoon could store all the world’s data,” says The Global 50 report of 2024.
The report also mentions that advances in DNA technology could make data storage faster and cheaper, turning it into a practical option for everyday use. Scientists are working on gentler ways to read and write DNA data, making it last longer.
BioCompute, an Indian biotech research startup (based out of Bengaluru), is working towards building the future of storage and computation architecture by transitioning to biomolecule-powered storage and processing infrastructure.
In an exclusive interview with AIM, Anagha Rajesh, the co-founder of BioCompute, funded by WTFund, said, “We haven’t gotten to using DNA storage in data centres yet. I mean, so far, it’s largely been scientific demonstrations around the possibility of doing this versus actually getting to commercial outcomes.”
They build end-to-end integrated microfluidic chips that enable them to write, read and store data in bacterial DNA. Currently, they are “developing proof-of-concept by leveraging CRISPR-based DNA editing to determine whether we can store and retrieve digital data that way”.
Another biotech company, Grow Your Own Cloud (GYOC), based in Paris, is investigating DNA data storage in plants. They believe that “the best way to alleviate the threat of ‘Data Warming’ is to work with nature”.
They reimagine systems for data storage that can absorb CO2 in a world where a data centre could grow data in your backyard and you can ‘literally Grow Your Own Cloud’.
The post Is the Race to AI Driving Data Warming — A New Global Warning? appeared first on AIM.