Jacobs' Digital Twin Could Reshape AI Data Center Economics as $3 Trillion Build-Out Nears Inflection Point

We are witnessing the start of a fundamental infrastructure supercycle. The demand for AI data centers is not just growing; it is following an exponential S-curve. The global market is projected to expand at a 25.8% compound annual rate, surging from $17.73 billion in 2025 to a staggering $133.51 billion by 2034. This isn't a niche trend. It represents the physical backbone for the next technological paradigm, requiring a massive capital outlay. Industry analysts estimate the sector will need up to $3 trillion in investment by 2030 to keep pace.

The scale of this build-out is immense. Between 2026 and 2030, roughly 100 gigawatts of new capacity are expected to come online, effectively doubling global data center capacity in just five years. This isn't about incremental upgrades; it's about constructing the fundamental rails for an AI-driven economy. The complexity is staggering, with facilities scaling to the gigawatt level, demanding unprecedented planning precision and operational certainty before committing such colossal capital.

Jacobs' new Data Center Digital Twin solution is positioned directly at this inflection point. Built on the NVIDIA Omniverse DSX blueprint, it aims to accelerate the planning and optimization of these colossal facilities. By creating a hyper-realistic virtual twin, the tool allows developers and owners to simulate compute, power, cooling, and site specifics long before a single brick is laid. This isn't a minor efficiency gain. It's a critical tool for reducing the risk and uncertainty inherent in building the next generation of infrastructure. In a market racing toward a 14% annual growth rate through 2030, the ability to plan faster, optimize energy performance, and ensure resilient operations from day one could be the decisive factor between success and costly delay.

The Digital Twin as a First-Principles Efficiency Engine

The core challenge in this infrastructure supercycle is one of brutal economics. As the scale of AI data centers explodes, so do the costs of getting them right. Construction costs have been rising at a 7% annual rate, with 2026 forecasts pegging them at $11.3 million per megawatt. For a gigawatt-scale facility, that translates to over $11 billion in hard costs alone. Any design flaw or operational inefficiency discovered late in the build is a direct hit to the bottom line and a major delay to revenue. This capability is critical. In a market where owners are committing capital to facilities that will operate for decades, the ability to achieve greater certainty upfront is a game-changer. It directly addresses the speed-to-power imperative, where months of delay can mean millions in lost revenue. By compressing the planning and validation phase, the digital twin acts as a powerful efficiency engine, potentially shaving months off the timeline and billions from the cost curve. In the race to deploy the next generation of infrastructure, this kind of precision is no longer a luxury-it's the fundamental requirement for winning.

Financial Impact and Market Positioning

Jacobs' digital twin solution is not a play for the dominant hardware pie. The AI data center market is currently dominated by hardware, which contributed 70.14% of the market share in 2025. Instead, Jacobs is targeting the services layer, a segment projected to grow at a high compound annual rate. This is a strategic bet on the infrastructure stack's most complex and capital-intensive phase: engineering and construction. By aiming to improve speed-to-revenue and long-term operations, the company is positioning itself at a critical choke point in the build-out.

The financial opportunity here is tied to the sheer scale of the coming build. With the market expected to expand at a CAGR of 27.48% from 2026 to 2035, the demand for planning and optimization services will surge in parallel. Jacobs' tool, built on the NVIDIA Omniverse DSX blueprint, directly addresses the brutal economics of this phase. It promises to compress timelines and reduce costly errors, translating into tangible value for clients who are committing billions to gigawatt-scale facilities. The solution's modular design also hints at a scalable revenue model, starting with the largest facilities and potentially expanding to smaller deployments.

This move is not without a track record. Jacobs has already been testing and enhancing the underlying technology through a collaboration with NVIDIA on AI factory digital twins. That prior work demonstrates a proven capability in applying digital twin solutions to the complex, high-stakes world of advanced manufacturing and data center design. This existing expertise provides a credible foundation for the new offering, suggesting the company isn't entering a completely uncharted space.

The competitive moat, however, will be built on execution and integration. The real advantage lies in Jacobs' deep industry knowledge and global workforce, which allow it to combine the digital twin's simulation power with practical engineering insight. The company's ability to deliver the physical facilities that the digital twin plans will be a key differentiator. In a market racing toward a 14% annual growth rate, the ability to plan faster and build with greater certainty could become a standard requirement, turning this service into a critical, sticky part of the infrastructure development workflow.

Catalysts, Risks, and What to Watch

The investment thesis for Jacobs' digital twin hinges on its ability to become the standard planning tool for the coming infrastructure supercycle. The near-term milestones are clear. Watch for customer adoption announcements, particularly integration with Jacobs' existing data center engineering contracts. The company is already involved in delivering high-performance computing environments for major players like NVIDIA, Hut 8 and PsiQuantum. Success will be measured by whether these clients adopt the digital twin for their next gigawatt-scale projects, turning a promising solution into a revenue-generating service.

The key risk is demonstrating tangible cost and time savings. The solution must prove it can directly reduce the $11.3 million per megawatt construction cost or accelerate time-to-power for these colossal facilities. In a market racing toward a 14% annual growth rate, any efficiency gain is critical. The digital twin's value is not in simulation for its own sake, but in its ability to compress timelines and eliminate costly physical errors. Without hard evidence of these savings, the tool risks becoming a sophisticated but non-essential add-on.

A broader structural shift will also test the solution's relevance. The market is poised for a paradigm shift in workload demand. While AI training drove the initial build-out, inference workloads could overtake training as the dominant requirement by 2027. Inference demands different design priorities-geographical distribution for low latency, continuous operation, and efficient power delivery. The digital twin must be able to model and optimize for this new, sustained demand pattern. If it remains optimized for the training-heavy, bursty workloads of today, its long-term utility could be limited. The coming years will show whether Jacobs' tool is a static planning aid or a dynamic platform that evolves with the AI infrastructure stack itself.