Building infrastructure where it's viable, not just visible
On June 29, 2026, in Phnom Penh, Comin Asia and Nokia announced a partnership that quietly challenges the prevailing logic of AI infrastructure — that bigger markets, built faster, always win. Instead, these two companies are turning toward Cambodia, Laos, and other underserved corners of Southeast Asia, where surplus power, sovereign ambition, and proximity to real data create conditions that saturated hyperscaler markets can no longer offer. It is a reminder that the geography of technological power is not fixed, and that the next layer of the digital world may be built not where capital concentrates, but where necessity and readiness quietly converge.
- Southeast Asian governments are growing uneasy with routing sensitive national data through foreign hyperscaler ecosystems, and that unease has become a market force.
- Cambodia and Laos — long overlooked by large-scale infrastructure investment — are emerging as viable deployment sites precisely because of their surplus power and policy openness, not despite their underdevelopment.
- The partnership deliberately rejects 'land banking,' the industry habit of grand announcements without execution, with feasibility work and early deployments already underway.
- Nokia's networking and automation technology combined with Comin Asia's hard-won regional engineering experience creates a pairing designed for environments where logistics, politics, and geography routinely defeat outside operators.
- Modular, edge-ready facility designs allow incremental deployment rather than single massive builds, lowering the threshold for markets that cannot yet absorb hyperscale investment.
- The initiative is landing as a structural bet that power availability, regulatory alignment, and data proximity will matter as much as capital in determining where AI infrastructure actually gets built next.
When Comin Asia and Nokia announced their partnership in Phnom Penh on June 29, 2026, they were making an argument as much as an announcement. The conventional logic of AI infrastructure favors concentration — massive facilities in established markets where power is reliable, regulation is settled, and capital moves freely. This partnership is built on the opposite premise.
Cambodia and Laos sit at the center of the initiative not as afterthoughts but as primary targets. Cambodia offers genuine early-stage deployment potential. Laos holds surplus electrical capacity and growing cross-border connectivity that could make it a regional hub. Thailand, by contrast, faces grid pressure and regulatory complexity that constrains large-scale expansion. The companies are designing modular, edge-ready facilities — built incrementally, tuned for local power conditions, and sized for markets that cannot absorb a hyperscale footprint all at once.
The division of labor is purposeful. Comin Asia contributes decades of mechanical and electrical deployment experience across some of the region's most logistically demanding environments. Nokia provides the networking fabric, automation platforms, and secure connectivity that AI workloads require. Together, they are targeting not just generic computing capacity but specific sectors — energy, telecommunications, finance, and government — where the need for localized infrastructure is most acute.
Central to the partnership's identity is an explicit rejection of what the industry calls land banking: the practice of announcing ambitious plans that never materialize. CEO Ivan Keogh drew the line clearly, distinguishing between markets that are viable — where power, regulation, and demand are real today — and those that are merely visible, generating headlines without execution.
The timing is not incidental. Across Southeast Asia, data sovereignty has shifted from aspiration to policy priority. Governments want sensitive information kept within national borders. Enterprises are reassessing their dependency on a handful of distant American platforms. This partnership addresses those concerns directly by enabling infrastructure that operates within local legal and geographic boundaries.
What this moment suggests is that the decisive variables in AI infrastructure are changing. Capital still matters, but power availability, user proximity, and policy alignment are becoming equally determinative. The companies that can read both the technology and the local context — and build where conditions are genuinely ready — may well define where the next generation of AI computing takes root.
Two companies announced a partnership in Phnom Penh on June 29, 2026, that signals a different way of thinking about where artificial intelligence infrastructure gets built. Comin Asia, a regional engineering firm with decades of experience deploying complex systems across Southeast Asia, and Nokia, the global connectivity company, are joining forces to construct data centers designed specifically for AI workloads in some of the region's least developed markets—Cambodia and Laos chief among them.
The conventional story of AI infrastructure is one of concentration. Hyperscale operators build massive facilities in already-saturated markets where power grids are robust, regulatory frameworks are established, and capital flows easily. But this partnership is built on a different premise: that the real opportunity lies in places where conditions are different, where power is available but underutilized, where governments are hungry for digital sovereignty, and where enterprises need computing capacity closer to where data is actually generated. Cambodia represents an early-stage market with genuine deployment potential. Laos has surplus electrical capacity and growing cross-border network connectivity that could position it as a regional hub. Thailand, by contrast, faces grid pressure and regulatory complexity that makes large-scale data center expansion difficult.
The division of labor is clear. Comin Asia brings what it knows best: the ability to execute infrastructure projects in environments where large-scale deployment is often constrained by geography, politics, or logistics. The company has spent years building mechanical and electrical systems across the region, learning how to navigate the particular challenges of each market. Nokia contributes the technological backbone—the networking fabric, automation platforms, and secure connectivity that modern AI workloads demand. Together, they're designing modular, edge-ready facilities that can be deployed incrementally rather than as single massive builds, with infrastructure that prioritizes energy efficiency and can adapt to local power conditions.
What makes this noteworthy is the explicit rejection of what the industry calls "land banking"—the practice of announcing grand plans with little immediate execution. Ivan Keogh, Comin Asia's CEO, framed it plainly: "This partnership is about building that infrastructure in the markets where it is most viable, not just most visible." The distinction matters. Viable means deployable now, with real power supply, real regulatory alignment, real customer demand. Visible means the kind of announcement that generates headlines but may never materialize.
The timing reflects a genuine shift in how governments and enterprises across Southeast Asia view their relationship to global technology infrastructure. Data sovereignty has become a priority—the desire to keep sensitive information within national borders rather than routing it through foreign hyperscaler ecosystems. Privacy regulations are tightening. The dependency on a handful of American tech giants is increasingly seen as a vulnerability. This partnership directly addresses those concerns by enabling infrastructure that can be deployed within national or regional boundaries, aligned with local law and local control.
Initial work is already underway in Cambodia and Laos, with feasibility assessments and early deployments in motion. The partnership expects to expand across additional Southeast Asian markets as conditions align. The infrastructure will support not just generic AI computing but specific applications—energy companies, telecommunications operators, financial institutions, and government agencies all have potential use cases. Ajay Sharma, Nokia's country manager for Thailand and Cambodia, described the collaboration as enabling "a new class of AI infrastructure that is distributed, secure, and aligned with real-world deployment conditions."
What emerges from this is a portrait of the AI infrastructure landscape in transition. Capital alone is no longer the decisive factor. Power availability, geographic proximity to users, and policy alignment are becoming equally important. The companies that can navigate those variables—that understand both the technology and the local context—will shape where the next generation of AI computing actually happens. For Southeast Asia, that could mean the difference between remaining dependent on distant data centers or building the infrastructure to process and control their own data.
Citas Notables
This partnership is about building that infrastructure in the markets where it is most viable, not just most visible.— Ivan Keogh, CEO, Comin Asia
We are enabling a new class of AI infrastructure that is distributed, secure, and aligned with real-world deployment conditions.— Ajay Sharma, Country Manager of Nokia Thailand and Cambodia
La Conversación del Hearth Otra perspectiva de la historia
Why does it matter that this partnership is focused on Cambodia and Laos rather than, say, Singapore or Bangkok?
Because those cities already have data centers. The real constraint in Southeast Asia isn't capital or technology—it's power and sovereignty. Cambodia and Laos have surplus electricity and governments that want to keep data local. That's where the actual demand is.
But aren't hyperscalers already building in Southeast Asia?
They are, but in the obvious places. This partnership is betting that the less obvious places are actually more viable right now. Laos has cheap power and wants regional connectivity. Cambodia is early-stage but deployable. The hyperscalers tend to concentrate where infrastructure already exists.
What does "sovereign AI infrastructure" actually mean in practice?
It means a government or enterprise can process sensitive data without sending it to an American data center. It stays within the country, under local control, subject to local law. For governments especially, that's become non-negotiable.
How is this different from just building another data center?
The modular approach. Instead of one massive facility, they're designing smaller, distributed deployments that can adapt to local power conditions and regulatory requirements. It's infrastructure designed for the place, not infrastructure imposed on the place.
Who actually benefits from this?
Governments get digital sovereignty. Enterprises get lower latency and local data control. Nokia and Comin Asia get to build in markets where there's real demand but less competition. It's aligned incentives.
Is this actually happening or is it just an announcement?
They say feasibility assessments are underway in Cambodia and Laos now. That's more concrete than most infrastructure announcements. But the real test is whether they actually deploy and whether it scales beyond the initial markets.