Mass adoption of generative artificial intelligence aside, data creation and replication is growing.
Researchers expect its current compound annual growth rate of 23% to continue through 2025. However, because AI is both computing and energy intensive, its presence makes big data even bigger.Â
All of this could throw a wrench in business leaders' ability to meet their sustainability goals if left unaddressed. Fortunately, companies across the IT spectrum are looking to take on that challenge.
"From our perspective, AI looks like a really big data problem," said Ben Golub, CEO of decentralized cloud storage company Storj. Through a shared network, Storj helps organizations rent out their spare capacity. This startup is joined by legacy enterprises like Dell and IBM in tackling data storage efficiency for economic and environmental sustainability.
Arthur Lewis, president and COO of Dell Technologies' Infrastructure Solutions Group, is helping customers transition from a traditional three-tier data center model to a software-defined (or decentralized) architecture. This transition helps customers optimize their workload in terms of cost, performance, ease of use, and efficiency.
Lewis noted that with a software-defined architecture, "You have the ability to buy what you need as you need it and scale out with your capacity requirements."
The market is roughly split in half, Lewis says, with those adopting decentralized data storage versus those sticking with traditional, consolidated architecture. With a single data center consuming 50,000 homes' worth of electricity (and the cloud at large now exceeding the airline industry's carbon footprint), industry stalwarts and startups are laser focused on solutions.
For 42% of CEOs, environmental sustainability is the top challenge over the next three years, IBM's 2023 CEO study says. Meanwhile, CEOs report facing pressure to adopt generative AI while weighing data management needs to make that adoption successful. For many, cloud storage is part of that equation, influenced by explosive data growth and sustainability goals.
Where the disk drives diverge
There are numerous solution pathways for the big data problem, but they all strive for the same destination: a more efficient and effective cloud.
Storj and other startups optimizing spare capacity offer one way for companies to tackle their data footprint, which can consume upwards of 40% of a data-intensive company's overall carbon footprint, Golub posits. "It can be really economical to use spare capacity," he said. "That turns out to also be the way you avoid most of the carbon impact."
Meanwhile, Dell seeks to guide customers to a decentralized, software-defined cloud architecture, but is also making the traditional model more efficient in the meantime. They have more EnergyStar certifications than any data storage vendor on the market, Lewis says. "We have built what we call 'design for sustainability' into our offer lifecycle process," he said.
For those that do go decentralized, Lewis says it's still important to maintain what he calls "centralized management of a decentralized architecture." Of course, Dell has the benefit of owning the entire stack, including the underlying compute, allowing them to drive significant efficiencies across the model. Even for companies without that advantage, prioritizing a holistic management model is key.
Another option centers around repurposing unused office buildings that have fallen out of style since the work-from-home revolution. Ermengarde Jabir, Moody's Analytics senior economist, specializes in commercial real estate. In today's growing data economy, that includes data centers.
Jabir has a front-row seat to the increased investment interest in repurposing unused office buildings as data centers. As co-location or shared space data centers are more accessible, these repurposed buildings not only provide more opportunities for decentralized data storage, but also improve physical building security. "Decentralization helps companies secure data because there isn't one specific location," said Jabir.
Whether a company seeks to optimize a traditional three-tier data center model or move towards a decentralized architecture, efficiency in the hardware used also plays a role in  sustainability. To help with this, the Open Compute Project developed guidelines for data center circularity, keeping physical equipment in use for longer and reducing waste. The raw materials, assembly, transportation, distribution, product use, disposal and recycling involved in manufacturing storage drives are no small matter.
According to Alan Peacock, general manager of IBM Cloud, organizations face increasing pressure from investors, regulators and clients to reduce their carbon emissions. "As part of any AI transformation roadmap, businesses must consider how to manage the growth of data across cloud and on-premise environments," Peacock said.
IBM recently launched its Cloud Carbon Calculator, an AI-informed dashboard designed to give clients access to standards-based greenhouse gas emissions data and help manage their cloud carbon footprint. This is yet another way for leaders to redirect their data focus to inherently include sustainability goals.
'Breaking down those walled gardens of technology'
Lewis predicts decentralized data storage will have its moment in the next few years, but the transition is already underway. AI, with its heavy data and computing power, is pushing it further.
"For customers that are looking to reduce their product carbon footprint, they really need to start thinking about how they are moving to more hybrid and connected environments and breaking down those walled gardens of technology," Lewis said.
Golub recommends performing a data audit to help address a company's carbon footprint as well as economic and performance efficiency. Some questions to ask include: What are your workloads? Which workloads are best suited for a decentralized data storage model?
Golub added, "People need to start by saying where am I creating data, what are my goals with that data, and now what is the carbon impact of that data?"
Big data sets like photo and video, medical images, scientific research, large language models and much more all make up a data-intensive environment. As Golub said, "The greenest drive out there is one that never has to be built and the greenest data center out there is one that never has to be created."
