AI boom to redefine data centres, power and security: Cisco executive

Artificial Intelligence (AI) is triggering what global tech leader Cisco calls the most profound platform shift in human history, reshaping jobs, workflows, infrastructure and national competitiveness and pushing the industry to rethink even where data centres should be located.

Speaking to PTI on the sidelines of the World Economic Forum Annual Meeting here, Cisco President and Chief Product Officer Jeetu Patel said the scale of AI’s growth is “non-trivial”, with compute, power and connectivity needs growing far faster than current infrastructure can support.

Infrastructure, trust and data gaps are binding constraints, Patel said , but each also creates investment opportunities. “There’s not enough power, computing and network bandwidth in the world to fulfil the needs of AI,” he said, adding that trust and a growing “data gap” as publicly available human-generated data runs thin are limiting adoption and performance.

Patel’s remarks come as the industry begins to explore data centres beyond Earth’s surface, a futuristic solution being pursued by several companies and space-industry players.

Space data centres: long-term vision, early pilots

Tech titans including Google and Amazon have publicly floated plans to put computing infrastructure in space. Google’s “Project Suncatcher” initiative aims to launch small solar-powered computing racks into orbit as soon as 2027 , with an eye toward scaling over the next decade.

Amazon founder Jeff Bezos has predicted that gigawatt-scale space data centres could be built within 10-20 years , powered by abundant solar energy and free from Earth-based power and cooling limits.

Meanwhile, space tech companies such as Starcloud, Inc. and Crusoe are working on orbital systems to deploy GPU-powered servers as early as 2026-2027 , marking a shift from concept toward experimental pilots.

Costs and economic debate :

Current launch costs, often thousands of dollars per kilogram, remain a major barrier, though reusable rockets like SpaceX’s Starship aim to drastically reduce these in the coming decade. Early models suggest that once orbital deployments scale and launch costs fall, space data centres could approach cost parity with Earth facilities over a multi-year lifecycle, as high broadband and power costs are offset by near-free solar energy over time.

However, not all experts are convinced. Some engineers argue that thermal management, radiation shielding and maintenance costs could make space data centres economically inferior to terrestrial ones for decades, if not permanently, without major technological breakthroughs.

Technical and operational risks :

Space-based data centres face a range of unique engineering challenges. Radiation exposure can rapidly degrade hardware unless costly shielding or specially hardened components are used. Heat rejection in the vacuum of space must rely on radiative cooling, which is far less efficient than the convective cooling systems available on Earth. Maintenance and repairs are exponentially more complex and expensive, while latency and data transfer between orbiting centres and ground users remain higher than terrestrial fibre links, potentially limiting real-time applications. These factors together make designing, operating, and securing space data centres a significant technical and economic challenge.

Experts also warn that orbital debris, space weather and the sheer physical distance introduce reliability risks that could slow adoption well into the 2030s.

Legal and regulatory hurdles :

Orbital data centres raise questions of space law, spectrum allocation and orbital traffic management. International treaties like the Outer Space Treaty govern peaceful use and liability but do not yet clearly assign jurisdiction over commercial computing infrastructure in orbit, creating a regulatory grey zone that global policymakers will need to address.

Cybersecurity dimension :

Patel stressed that as AI infrastructure scales whether on Earth or in space cyber defence must operate at machine speed. “If you don’t use AI for your cyber defences, you will be outsmarted,” he said. He emphasized that securing AI systems themselves with runtime guardrails is as critical as defending networks, underscoring that advanced infrastructure introduces new attack surfaces.

India remains strategic :

Against this backdrop, Patel reaffirmed Cisco’s commitment to India, describing it as one of the company’s most strategic growth markets. India hosts Cisco’s largest employee base outside the United States, and the company is expanding manufacturing and partnerships, including a Chennai facility launched last year.

“Our goal is to continue serving the best technologies so that India can generate capabilities for AI,” Patel said, highlighting strong public-private engagement and broad sectoral deployment.

As the AI wave accelerates, Cisco’s dual focus investing in current ground infrastructure while embracing high-risk future bets like space data centres reflects the vast transformation underway in global computing.