The supply chain behind the AI boom

We sat down with Amanda Rasmussen, Chief Commercial Officer of DHL Global Forwarding and Executive Sponsor of Data Center Logistics at DHL Group, to find out what's really driving the surge, and what it takes to keep up.

What we're seeing right now is that demand has really taken off, and AI is a major driver of that. It's created a gold-rush moment in data center logistics (DCL). And that logistics side is changing and growing quite significantly. In 2025, the DCL market was estimated at roughly $23 billion, and we expect that to rise to at least $35 billion by 2030.

We're not new to serving the data-center sector or handling these kinds of complex technology moves. What’s changed is the level of coordination now required. Companies building AI infrastructure increasingly need end-to-end logistics partners that can handle everything from factory pickup to secure installation inside the data center itself.

For companies scaling AI infrastructure, “end-to-end” means having a single logistics partner manage the full lifecycle of data center equipment – from factory pickup and deployment to decommissioning, refurbishment, and redeployment.

The objective is consistent: fewer handoffs, faster deployment, and greater visibility across every stage. In fact, around 85% of hyperscalers say they would prefer to work with a single end-to-end logistics partner, yet only 43% feel they currently have one. That gap — 85% wanting it, only 43% getting it — defines the commercial opportunity DHL is positioned to capture given our scale, footprint and capability.

That journey typically begins with first-mile pickup at the factory and continues through international transport, staging and warehousing, and high-security, white-glove delivery into the data hall – the secure area within a data center where racks and servers are installed and commissioned.

But the work does not end once equipment is live. As infrastructure is upgraded, servers and racks must be securely removed, refurbished, recycled, or redeployed to other locations. Managing this full lifecycle has become increasingly important as AI hardware grows more valuable and refresh cycles accelerate.

For customers, this approach delivers greater speed, predictability, and accountability through one integrated partner. With capabilities spanning multiple business units, DHL Group is well positioned to support the entire data center lifecycle.

The data center logistics supply chain begins with supplier coordination at the factory and extends through secure handling and linehaul – often using charter aircraft for time-critical shipments – to high-security final-mile delivery into the data center.

From there, white-glove teams take over inside the facility. This includes positioning and levelling racks for precise installation, grounding equipment for electrical safety, and connecting critical cabling. It is highly specialized work that demands extreme care and technical expertise.

The process continues beyond deployment. As technology is refreshed, equipment must be decommissioned, securely removed, and either refurbished, recycled, or redeployed elsewhere.

Throughout the entire lifecycle, protecting the integrity of high-value equipment is essential. Theft prevention, chain of custody, and careful handling are critical at every stage.

The implications reach far beyond the logistics industry. Delays in data center deployment can hinder everything from AI adoption and cloud expansion to the digital services that have become essential to daily business operations. In other words, the speed of infrastructure deployment is starting to shape the speed of innovation itself.

Server values have surged. Pre-AI servers cost around $100,000 to $250,000. In today's AI environment, one to three million dollars per server is common – and the value is expected to continue rising as AI workloads demand increased power. Today's AI servers pack exponentially more processing power, custom silicon, and advanced cooling requirements into every unit — making them both more valuable and more complex to handle safely.

At the same time, servers can't simply keep getting larger. Beyond a certain size, they're no longer practical to move by air, so they become heavier and more difficult to handle, with even higher requirements for specialized, secure logistics.

On critical ad hoc moves, companies operating major infrastructure projects often expect pricing responses within two to four hours for air freight and within 24 hours for ocean freight. In standard freight markets, equivalent quotes typically take 24 to 72 hours — making this level of responsiveness a genuine operational differentiator. That gives you a sense of how fast-paced and competitive this market is. The industry increasingly demands both speed and agility, which means decisions have to move quickly alongside the freight itself.

For predictable, pre-planned flows, price matters more. For sudden ad hoc demand, the cost of missing a delivery slot outweighs logistics costs. Ultimately, it's a balance between cost and reliability. Hyperscalers operate at such a scale that even short downtime is unacceptable, so certainty and consistency become critical.

Near-term demand is strongest in the US – DHL Supply Chain recently announced ten dedicated sites and more than seven million square feet of new warehousing capacity in North America. But growth is happening everywhere. We're active across Europe, India, Australia, Indonesia, and the Middle East, which reflects how global the expansion of AI infrastructure has become.

To support that, we've built a dedicated cross-group setup with regional teams across Asia Pacific, Europe, the US, Latin America, and the Middle East and Africa, complemented by a global Center of Excellence focused on operational standards, partner ecosystems, and solution design. What makes our setup distinctive is that we drive end-to-end data center logistics as one integrated partner – not a patchwork of providers. We ensure mission-critical reliability and uptime across every site and stage. We execute with precision and dedicated expertise in high-stakes environments. And we do all of this at the pace and scale up a rapidly growing global industry demand. I don't see any other provider that can bring that full set of capabilities together.

I think the question of hype is primarily about the market valuations of AI companies. We, on the other hand, see real volume and demand. As long as AI demand remains strong, infrastructure must follow, and that keeps data center logistics growing. Over time, growth will normalize, but don't forget that pulling servers out, refurbishing, and redeploying also means the work continues, just with a different mix.

There's also the factor of what's next in AI and what it will take to power that in terms of server speed, capability, and value. This may continue to fuel demand well beyond our current expectations.

With hyperscalers, we've seen significant revenue growth in 2025. And we're doubling down on our 2026 expectations, scaling our capabilities to capture what's coming. Recent wins include hyperscaler warehousing projects in the US and Asia. We also have a very strong pipeline across all business units. The AI race may look digital from the outside. But behind every breakthrough sits a physical infrastructure that still must be built, moved, installed, secured, upgraded, and eventually replaced. As demand accelerates, the real challenge is no longer just developing more powerful technology – it’s building the global infrastructure fast enough to support it.