AI is power-hungry. In line with estimates, by 2030, knowledge centres powering synthetic intelligence might be consuming an additional 200 terawatt-hours of electrical energy yearly. That’s a large 160% bounce from immediately’s ranges, as per Goldman Sachs analysis.
That’s not all—AI additionally calls for quite a lot of water, placing our restricted sources below immense pressure. Even a small 1 MW knowledge centre can use as much as 25.5 million litres of water yearly, whereas large-scale knowledge centres, equivalent to these operated by Google, eat greater than a billion litres annually.
To fight a few of these points, Microsoft launched a brand new knowledge centre design in August 2024, aimed toward decreasing water utilization and selling a extra sustainable future. This revolutionary design incorporates chip-level cooling know-how, successfully eliminating the necessity for water evaporation in cooling processes.
Microsoft isn’t alone. Pushed by these sustainability issues, Mysore-based software program agency Vigyanlabs Improvements Non-public Restricted took a proactive method by growing its personal micro inexperienced knowledge centres.
The Genesis
Vigyanlabs has undertaken the complicated activity of creating knowledge centres extra energy-efficient and decreasing their carbon footprint. Srinivas Varadarajan, CEO and the visionary behind this innovation-driven firm, shared with AIM the journey that led to the event of their energy-saving options.
As the previous principal architect of Aadhaar, Varadarajan witnessed first-hand the large vitality consumption and carbon footprint related to constructing a large-scale platform. It was throughout this time that he totally realised the numerous vitality calls for of working hundreds of servers, resulting in a larger consciousness of their environmental affect.
In India, the ₹85,580 crore ($10 billion) knowledge centre trade is primarily concentrated in Mumbai (49%) and Chennai, with different key hubs in Noida, Bengaluru, and Pune, as per studies. Moreover, the demand for hyperscale knowledge centres is rising quickly, with projections indicating a 35% annual enhance till 2040.
Prior to now 5 years alone, Hyderabad, Navi Mumbai, and Pune have collectively dedicated 440 acres to hyperscale growth.
By means of intensive analysis, Vigyanlabs’ crew found that the vitality consumption of information centres was similar to that of the aviation trade. Designing an energy-efficient knowledge centre, nevertheless, isn’t any small feat.
“The method of constructing knowledge centres is extremely refined. It entails civil, electrical, cooling methods, community and IT infrastructure, software program —all of which add to its complexity,” he defined.
With a number of distributors contributing to totally different elements of the infrastructure, addressing this subject required a scientific and complete method.
Vigyanlabs’ journey started in a startup incubator at SJCE-STEP, a government-backed enterprise incubator in Mysore, the place a bunch of engineers had entry to 1,000 computer systems to check their algorithms. This analysis led to the event of their flagship software, which offered roughly 42% vitality financial savings on desktop variations.
“We constructed the product and received the Nasscom Innovation Award for it. That recognition gave us visibility and opened doorways to collaborations with main establishments like SBI and different massive firms,” Varadarajan recalled. Since then, Vigyanlabs’ know-how has saved almost 100 gigawatt hours of vitality.
Innovation that Makes an Influence
On the coronary heart of their innovation is Vigyanlabs’ sensor-based Enterprise Energy Administration Software program, that includes a singular gentle vitality meter (SoftEM)—an software that measures vitality consumption throughout laptops, desktops, and servers. Its flagship product, Clever Energy Administration (IPM+), helps Home windows, Linux, servers, UPS methods, vitality meters, and printers.
Increasing past desktop computer systems, Vigyanlabs turned its focus to tackling broader vitality inefficiencies in knowledge centres. It discovered that almost 50% of the vitality consumed by knowledge facilities was both wasted or used inefficiently.
Recognising that totally different customers have various computing wants, the corporate developed adaptive algorithms that be taught and regulate primarily based on particular person utilization patterns, guaranteeing extra environment friendly vitality administration.
Right this moment, Vigyanlabs’ vitality optimisation options are deployed on hundreds of thousands of machines worldwide, considerably decreasing firms’ carbon footprints. Their algorithm actively runs on between 1.5 and 6 million machines globally, making a considerable environmental affect.
Vigyanlabs Inexperienced AI Chatbot
Moreover, the corporate has a community of micro knowledge facilities, backed by 7+ US patents. It operates on renewable vitality and consumes 50% much less vitality in comparison with hyperscalers. Vigyanlabs additionally claims that its inexperienced micro knowledge centre community operates with zero water for cooling. By optimising chiller operations, it has efficiently diminished PUE (energy utilization effectiveness) from 2.0 to 1.20.
This knowledge centre depends on an open supply massive language mannequin (LLM), which has been additional optimised, known as Mysore.ai—a inexperienced AI chatbot.
What makes Mysore.ai particular? In line with Vigyanlabs, its uniqueness lies in its considerably decrease value and vitality footprint, claimed to be the order of magnitude lower than the main AI-supported chatbots equivalent to ChatGPT (OpenAI), Perplexity.AI, Gemini (Google), and CoPilot (Microsoft).
Not like these fashions, Mysore.ai can run on a typical laptop with no GPU or a low-powered GPU and makes use of open supply instruments, permitting free utilization. The corporate shared its ardour for constructing its personal AI stack utilizing fewer sources, very similar to DeepSeek, quite than counting on business AI know-how.
Not like conventional AI fashions designed for general-purpose duties, Vigyanlabs’ AI system is particularly constructed to optimise vitality use throughout all operations. It manages your complete knowledge centre, together with servers, storage, cooling, electrical, and safety methods, automating processes to minimise errors and guarantee optimum efficiency.
Recognising the issue of retrofitting present buildings, the corporate selected to construct its personal facility with sustainability as a key focus. To minimise environmental affect, it designed particular bricks that diminished cement utilization by 70%, permitting the constructing to remain cool naturally and cut back air-con wants.
This design achieved a 30% vitality saving of the cooling system, decreasing the constructing temperature by 6 levels Celsius. “We designed transformers with solely 5% vitality losses in comparison with the standard 10-15%.”
Moreover, Varadarajan talked about {that a} real-time vitality auditing system was launched to make sure environment friendly energy administration. Cooling stays one of many largest vitality customers in knowledge centres, with conventional chillers relying closely on water.
The initiative extends past their very own operations. “We are literally being known as world wide to assist construct nationwide vitality insurance policies. We labored within the Philippines, Togo, and a number of other different nations as know-how advisors to sort out related challenges,” Varadarajan added.
The submit Mysore Has a Information Centre that Doesn’t Drink Your Water! appeared first on Analytics India Journal.
