The New AI Infrastructure Race Is Moving Into Space

For decades, the technology industry’s infrastructure strategy has been remarkably straightforward: Build bigger data centers, add more fiber and deploy more compute capacity closer to users. But what if the next major leap in AI infrastructure happens above the planet rather than on it?

That question is gaining attention as SpaceX continues expanding its Starlink satellite network and explores ways its orbital infrastructure could support AI-related computing and global data movement. While the concept of space-based AI infrastructure remains in its early stages, it represents a potentially significant shift in how organizations think about compute, connectivity and data distribution. Instead of relying exclusively on terrestrial networks, future AI systems could leverage orbital infrastructure to extend services into remote regions, improve resilience and create entirely new competitive dynamics.

The idea is gaining traction at a time when demand for AI infrastructure is accelerating rapidly. According to a Goldman Sachs analysis, AI-related data center power demand is expected to increase dramatically through the end of the decade as organizations race to secure the compute capacity needed to support next-generation AI applications. As those investments accelerate, executives are increasingly asking whether future infrastructure strategies will be limited to Earth—or whether space will become a critical extension of the global AI stack.

To better understand the opportunities and risks, members of the Senior Executive AI Think Tank shared their perspectives on how space-based AI infrastructure could reshape cloud providers, telecommunications companies and AI platform vendors over the next decade. Their insights reveal both extraordinary possibilities and significant challenges, from global connectivity and distributed computing to governance, economics and the growing concentration of infrastructure power.

“Cloud giants spent decades pouring concrete while orbital compute skips that entirely.”

– Pon Murugesh Devendren, SAP Enterprise AI Architect at Deloitte

SHARE IT

The Battle Will Be Won at the Orbital-to-Ground Handoff

Pon Murugesh Devendren, SAP Enterprise AI Architect at Deloitte and a technology leader with more than 20 years of experience designing enterprise AI platforms and cloud-native solutions, believes the most important shift is not simply where computing occurs but who controls the connection between space and Earth.

“Space-based AI infrastructure rewrites where infrastructure lives,” Devendren says. “Cloud giants spent decades pouring concrete while orbital compute skips that entirely.”

He argues that telecommunications companies could face particular disruption if satellite providers begin delivering both connectivity and compute services through a unified platform.

“For telcos, the threat is serious,” he says. “When a satellite delivers compute and connectivity together, leasing terrestrial backbone starts to collapse.”

At the same time, Devendren cautions that launching hardware into orbit does not eliminate practical constraints. Heat management, system maintenance and economic viability remain substantial hurdles.

“The winners will not be whoever launches first but whoever owns the orbital-to-ground handoff,” he says. “Space will not replace the data center. It will reach the workloads where the data center was never coming.”

Engineering Reality May Slow Adoption

Marc Massar, Founder of AURA Labs and a longtime technology executive who has held senior leadership roles at JPMorgan Chase, HSBC and Worldpay, believes enthusiasm for orbital AI infrastructure often overlooks significant engineering realities.

“There are so many issues with space-based compute infrastructure,” Massar says. “Heat dissipation, radiation shielding and node-to-mesh latency dynamics.”

He notes that deploying compute infrastructure in orbit involves challenges far beyond simply launching servers.

“You can’t simply load a satellite up with GPUs and launch it into space,” he says.

While Massar credits satellite connectivity services for expanding access in difficult geographies and politically sensitive regions, he sees a meaningful distinction between communications infrastructure and large-scale AI computing.

“It’s been a great boon to remote locations as a network access method,” he says. “But there’s a big leap between access and scaled compute.”

For executives evaluating future infrastructure strategies, his comments serve as a reminder that technological possibility does not always translate into commercial viability.

A New Orbital Edge Layer Emerges

David Obasiolu, AI Security, Governance and Systems Consultant at Vliso AI, sees space infrastructure as an extension of existing technology ecosystems rather than a replacement for them.

“Space-based AI infrastructure won’t replace AWS, Azure or telecom networks,” Obasiolu says. “It becomes a new orbital edge layer.”

That edge layer could create new competitive dynamics across industries.

“It could reshape competition by giving cloud providers global low-latency reach, forcing telecoms to defend backhaul and remote connectivity, and giving AI vendors another compute and data-movement marketplace,” he says.

However, he warns that resilience and global reach must be weighed against substantial risks.

“The benefits are resilience, global coverage, faster access in remote regions and reduced dependence on terrestrial grids,” he says. “The risks are regulation, orbital debris, cyber and physical security, thermal limits, repairability, unclear SLAs and vendor concentration.”

Ultimately, Obasiolu believes organizations should avoid viewing space infrastructure as a standalone strategy.

“The winners will combine terrestrial cloud, telecom networks and orbital infrastructure rather than betting on space alone.”

“I see space-based AI infrastructure as a strong extension of existing Earth infrastructure, not a replacement.”

– Manpinder Singh Panesar, Senior Solutions Architect at Amazon Web Services

SHARE IT

Distributed Infrastructure Expands Global Reach

Manpinder Singh Panesar, Senior Solutions Architect at Amazon Web Services, focuses on the networking implications of orbital infrastructure.

“I see space-based AI infrastructure as a strong extension of existing Earth infrastructure, not a replacement,” Panesar says.

He believes space networks could dramatically expand edge computing capabilities.

“A network in space can extend edge networks here on Earth, especially in remote areas, and make global reach wider with a consistent user experience,” he says.

Panesar also points to the possibility that orbital networks could democratize advanced infrastructure capabilities once available only to a handful of technology giants.

“If space networks make these primitives more available, they can unlock a lot of potential,” he says.

Still, he identifies familiar concerns around “cost, regulation, security, reliability and centralization risk.”

Governance Could Become the Ultimate Competitive Advantage

Will Conaway, President of Tuxedo Cat Consulting and an award-winning technology and healthcare executive, believes regulation may prove more decisive than technology.

“Regulatory complexity may be the biggest factor shaping space-based AI infrastructure,” Conaway says.

According to Conaway, future competition will extend far beyond speed and coverage metrics.

“Cloud, telecom and AI vendors will not just compete on speed or coverage,” he says. “They will compete on who can satisfy spectrum rules, orbital licensing, export controls, data sovereignty, privacy requirements and cross-border cybersecurity obligations.”

The upside remains compelling.

“The upside is a more resilient global network that can support underserved regions, disaster response and distributed AI workloads,” he says.

Yet compliance burdens could increase dramatically.

“Every satellite link may touch multiple jurisdictions, making compliance, liability, auditability and incident response harder.”

He predicts organizations that establish transparent governance frameworks early will enjoy a long-term advantage.

Capital and Integration Will Define Winners

Rajasekhar Chitta, Enterprise Transformation Leader at Cox Enterprises, expects a hybrid future rather than wholesale disruption.

“Space-based infrastructure could shift AI from centralized clouds to globally distributed, latency-aware networks,” Chitta says.

He acknowledges the limitations facing terrestrial infrastructure while cautioning against overestimating the speed of adoption.

“This may not be a gold rush,” he says. “These are capital-intensive, highly regulated and carry security and orbital risks.”

Even as space capabilities advance, Chitta expects Earth-based infrastructure to continue improving rapidly.

“The likely outcome would be a hybrid, strategically competitive ecosystem,” he says, “where leadership is defined by capital, policy alignment and integration capabilities.”

Decentralization Creates Opportunity and Risk

Chandrakanth Lekkala, Principal Data Engineer at Narwal.ai, believes space-based infrastructure has the potential to challenge traditional geographic constraints.

“AI compute could be spread throughout an infrastructure that is inherently decentralized,” Lekkala says.

That capability could enable inference and AI services in locations that are currently underserved.

“This poses a direct threat to traditional cloud providers by breaking geographic monopolies,” he says.

Yet Lekkala sees substantial concerns around governance and ownership.

“Space infrastructure democratically enables access to space,” he says, “but it also allows ownership by a few companies and has not yet been appropriately governed at the international level to safeguard the global commons.”

New Markets Could Emerge Above the Planet

Fabio Danze Montini, Investor and Owner of FDM Industrial Sales & Marketing SL, sees orbital infrastructure creating entirely new market opportunities.

“Space-based AI infrastructure will not replace cloud, but it could create a new edge layer above the planet,” Montini says.

He expects benefits across industries including logistics, climate monitoring, mobility and industrial operations.

“For telecoms, it is both a threat and a partnership opportunity,” he says.

At the same time, he notes that familiar concerns remain.

“The positives are resilience, coverage and faster access where terrestrial infrastructure is weak,” he says. “The negatives are cost, regulation, space debris, security risk and concentration of power.”

“The risk is that we may simply move concentration from a few cloud providers on the ground to a smaller number of infrastructure owners in orbit.”

– Andre Shojaie, Founder of HumanLearn

SHARE IT

The Real Shift Is Strategic Convergence

Andre Shojaie, Founder of HumanLearn and a specialist in AI governance and leadership transformation, believes executives should focus less on infrastructure categories and more on convergence.

“I would be careful not to look at space-based AI infrastructure only as a faster version of cloud or telecom,” Shojaie says.

Instead, he sees compute, connectivity and data movement merging into a single strategic layer.

“The upside is clear: more resilience, broader coverage, lower dependency on local infrastructure and potentially new AI capabilities,” he says.

But governance concerns loom large.

“The risk is that we may simply move concentration from a few cloud providers on the ground to a smaller number of infrastructure owners in orbit.”

Shojaie argues that issues of sovereignty, pricing, security and accountability may become even more complex as infrastructure moves beyond national borders.

Space Could Create New Infrastructure Economics

Rodney Mason, Chief Marketing Officer at Minty, believes orbital AI infrastructure could unlock entirely new economic models.

“Space-based AI infrastructure could include AI data centers, eliminating the need for massive energy and cooling sources required for Earth-based AI data centers,” Mason says.

He also sees significant benefits in global connectivity.

“It could open new markets and enable distributed AI services closer to users,” he says.

For telecom providers, the development creates both opportunity and risk.

“It introduces both partnership opportunities and competitive pressure,” Mason says.

However, he cautions that challenges including “cost, spectrum constraints, regulatory complexity and cybersecurity risks” remain significant barriers to adoption.

Navigating the Rise of Orbital Infrastructure

• Focus on integration, not replacement. The most successful organizations will connect terrestrial and orbital infrastructure rather than betting exclusively on either model.
• Separate connectivity from compute assumptions. Satellite networking success does not automatically translate into economically viable space-based AI processing.
• Prepare for an orbital edge layer. Space infrastructure is likely to augment cloud ecosystems by extending reach, resilience and low-latency access.
• Build distributed architectures now. Organizations that design globally distributed systems today will be better positioned for future orbital networks.
• Treat governance as a strategic capability. Regulatory readiness may become as important as technical innovation.
• Evaluate capital intensity carefully. Hybrid ecosystems will reward organizations that balance investment discipline with innovation.
• Monitor concentration risks. Decentralized technology can still result in centralized ownership and control.
• Look for industry-specific opportunities. Logistics, defense, climate, agriculture and industrial monitoring may benefit first.
• Plan for convergence. Future infrastructure strategies should account for compute, connectivity and data movement operating as a unified layer.
• Watch infrastructure economics closely. New approaches to energy, cooling and global distribution could reshape competitive advantages.

The Expanding Definition of AI Infrastructure

Space-based AI infrastructure is still in its early stages, but the direction of travel is becoming harder to ignore. As demand for AI compute accelerates and terrestrial systems face increasing strain, organizations are being forced to think more broadly about where and how infrastructure is built.

What the Senior Executive AI Think Tank agrees on is that this is not a simple replacement story. Space is unlikely to displace the cloud or telecommunications networks, but it may extend them in meaningful ways—creating new layers of connectivity, resilience and global reach while also introducing new challenges around cost, regulation and control.

For executives, the takeaway is not to bet on orbit or ignore it, but to recognize that the definition of infrastructure is expanding—and that they’ll need to adapt to that broader, more integrated model as it takes shape.