Converging Labs and Data: How AI Is Shaping the Future of Life Sciences Infrastructure
Neville Willsmore, principal and regional practice director at HKS
As artificial intelligence continues to redefine the boundaries of innovation across the life sciences sector, a new type of facility is emerging at the intersection of discovery and data. Neville Willsmore, principal and regional practice director at global design firm HKS, has been at the forefront of this shift—closely observing a growing trend: the convergence of laboratory and data center infrastructure. With AI transforming everything from drug discovery timelines to the security demands of sensitive health data, the architectural requirements for these spaces are evolving rapidly.
In this Q&A, Willsmore explores the rise of hybrid lab–data center facilities, particularly in regions like Phoenix, Denver, and Salt Lake City, where underutilized industrial land offers ideal conditions for repurposing. He discusses the strategic advantages of converting vacant high-rises and warehouse sites into dual-purpose research hubs, and how these spaces are being designed to handle the unique demands of AI-powered drug development. From structural adaptability to the critical need for secure, scalable data processing, Willsmore offers insight into what’s next for the physical infrastructure supporting life sciences innovation.
Q: How are architecture and engineering teams approaching the design and renovation of lab facilities to accommodate the unique infrastructure requirements of integrated data centers—especially in adaptive reuse projects like industrial conversions or high-rise retrofits?
A: Architecture and engineering (A/E) teams are adopting a systems-based approach when retrofitting labs to integrate data center capabilities, especially in adaptive reuse scenarios such as industrial conversions and high-rise retrofits. This includes reinforcing structural capacity for heavier equipment, integrating raised-access flooring for cable management and cooling distribution, and leveraging modular mechanical, electrical, and plumbing (MEP) systems. In multi-story buildings, core shaft space has to be optimized for fiber, power, and chilled water infrastructure. Increasingly, prefabricated IT modules are used to reduce retrofit timelines and cost.
Q: What are the biggest design or construction challenges when repurposing older buildings—like factories or vacant office towers—into dual-use spaces for both scientific labs and data processing, and how can these be addressed without major downtime or massive budgets?
A: Repurposing older buildings for new uses as life-science labs or data processing is advantageous in terms of “speed to market” due to shortened construction timelines. Where ground-up construction can often take 3-5 years, conversion projects typically take only 1-2 years to complete. However, older facilities, including factories and vacant office towers, present several limitations: Low floor-to-floor heights, limited load-bearing capacity, outdated HVAC systems, and insufficient electrical capacity. These constraints can be addressed through targeted upgrades such as:
Localized in-rack cooling systems (including liquid cooling close to the heat source to reduce energy waste)
Structural retrofitting (adding selective steelwork to improve floor loading and reduce vibration impacts)
Modular electrical distribution panels and UPS systems
Prefabricated infrastructure pods
Digital modeling using BIM for efficient preconstruction coordination of HVAC system requirements & spaces
These strategies help reduce downtime and avoid large-scale capital expenditures.
Q: With the rise of AI in drug discovery and personalized medicine, how is the need for real-time, secure data processing shaping the way lab spaces are planned or retrofitted to include adjacent or co-located data center capabilities?
A: Life sciences companies are embracing artificial intelligence to improve workflows, accelerate novel drug development and streamline clinical trials. The integration of AI in drug discovery and personalized medicine has made real-time, secure data processing a core requirement. This is driving the need for co-located or embedded micro data centers within lab environments.
New designs prioritize:
Edge computing capabilities
Shielded, climate-controlled server rooms
Redundant fiber pathways
Physical and cybersecurity convergence through smart building systems
Proximity of compute power to lab benches enables faster model training, data analysis, and iterative discovery.
Access to (and control of) an on-site data server is a true difference maker for companies in the biotech, medtech, and fintech industries. This “digital amenity” can provide tenants with greater processing speed, security and customization than their competitors located elsewhere. For these companies, latency and lag associated with cloud computing can have a real impact on their products and services. This is especially true as increasing wireless services and the emergence of large language models (and other artificial intelligence) are placing ever-greater demands on cloud computing.
At the SEMI 2025 Technology Workshop I attended in Phoenix on April 22-24, Intel demonstrated some software and hardware configurations that are “self-contained” solutions, not requiring large cloud-based storage systems—the quest is to “miniaturize” the data center components, similar to the way that other technologies have evolved over time.
As advances in AI and other technologies affect the way scientific research is done, the spatial needs of a facility change—less floor area is needed for the lab and office environments (as equipment sizes reduce and hybrid work models have become the norm—so more shared spaces are common), which then allows more area for on-site co-located data facilities. With the incorporation of direct liquid cooling systems for servers—even the data center component can be reduced in size, since DLC systems only require a third of the space that a traditional air-cooled data center occupies.
Q: In regions like Phoenix, Denver, and Salt Lake City—where there's a lot of industrial land being repurposed—what makes these locations especially suited for hybrid lab/data center facilities from a design and operational perspective?
A: These cities have emerged as prime candidates for hybrid lab/data center development due to:
Abundant, affordable industrial real estate with adaptable footprints
Favorable climate conditions, particularly for free air (economizer mode) or evaporative cooling
Expanding utility infrastructure and grid resilience investments
Access to renewable energy sources such as solar, wind and hydropower
Growing life sciences ecosystems supported by universities, health systems, and local incentives
Increasing pool of industry-specific skilled talent (graduating from local universities)
Competitive tax environment and availability of state and municipal subsidies and tax credits
Limited environmental risks—low likelihood of natural disasters affecting sites or operations
Lower energy costs than major markets
Designers benefit from existing high-clearance industrial building shells (if repurposing) and minimal municipal zoning barriers that enable fast-track development. New facilities offer the potential to be designed as buildings with renewable energy sources and heat reuse technologies, use water-efficient cooling, and thus align with global carbon reduction capabilities.
Q: For lab operators or developers with limited budgets, what are some high-impact, low-disruption upgrades you recommend to prepare existing lab infrastructure for future data integration or AI-powered workflows?
A: For operators with limited budgets, a phased approach to digital-readiness can deliver significant ROI. Recommended upgrades include:
Installing high-capacity fiber backbone and structured cabling
Designating edge compute zones in underutilized areas (to bring the storage/processing of data closer to the user)
Implementing IoT-enabled HVAC zoning
Deploying cloud-based workflows to reduce immediate hardware investment
These updates provide flexibility and prepare labs for AI workflows without extensive renovation.
With regard to IT hardware, integrating AI can be achieved by either utilizing existing servers and their supporting equipment to perform new AI functions, or augmenting hardware deployed with new AI-specific equipment to perform new AI functions. For example, taking an existing rack of CPU-based servers and adding two new GPU-based servers to provide more parallel computing power to launch a chatbot to a company’s internal users. While this may seem easier and less costly than accommodating a new AI high-density deployment, it comes with three sets of challenges:
Adding GPU-based servers to an otherwise low rack density aisle may create hot spots that the building’s cooling system was not originally designed to handle
It may create uneven power loads across the facility and lead to the need to re-allocate backup power resources
It may lead to network congestion as the new equipment multiplies the data transferred per rack
Q: How are HVAC, power, and security systems evolving in lab/data center hybrid facilities to meet the demands of both environments—especially when repurposing older structures?
A: Hybrid environments require converged building systems capable of meeting dual demands. Key developments include:
Dual-loop HVAC systems for managing both lab precision and server heat loads
Redundant electrical configurations with dynamic power management
Unified security platforms combining biometric access, video analytics, and network monitoring
Raised floor systems to allow easy access for systems reconfiguration
For older buildings, the integration of rooftop or mezzanine systems allows for expanded capacity without intrusive retrofits. New developments in fuel-cell technologies, microgrids, direct liquid cooling, chip-level cold plate cooling, liquid immersion cooling utilizing dielectric fluid and closed-loop zero-evaporation cooling methods are gaining traction to provide more efficient power and cooling, with the potential to require less space than traditional HVAC systems in retrofit situations. Nuclear energy and mini-reactors may also play a role as organizations look to meet their growing compute demands without compromising their sustainability initiatives.
Q: From a scalability standpoint, how do you future-proof lab spaces during renovation or design phases to ensure they can support increased AI processing power, data storage, and secure connectivity needs without recurring major overhauls?
A: To ensure long-term adaptability, lab spaces should be designed for scalability from the outset. Best practices include:
Designing for higher structural and power densities
Including spare conduit and cooling paths in infrastructure planning
Adopting modular, rack-ready zones for rapid AI hardware deployment
Preparing select zones for future liquid cooling
Leveraging software-defined networking and infrastructure for flexibility in resource allocation
Flexible and modular design (typically a bay module of 11’ works well for lab spaces)
This minimizes the need for disruptive, large-scale overhauls as compute and storage demands grow.
Future lab expansion should prioritize flexibility and modularity in design to accommodate changing research needs and technological advancements. This approach allows for easy reconfiguration and scalability.
With the growing amount of data generated in labs, proper integration and analysis tools are essential. Lab expansion should consider data management systems that can handle large volumes of information effectively.
Q: What key factors should developers and life sciences companies consider when identifying promising markets or cities for new biotech and research hubs—particularly when evaluating locations for hybrid lab and data center facilities?
A: When evaluating new markets, developers and life sciences companies should consider:
Access to research institutions, healthcare providers and highly skilled technical talent
Flexible zoning and building codes supportive of dual-use facilities
Fiber connectivity and clean utility infrastructure
Inventory of convertible commercial or industrial space
Public incentives, grants, and favorable tax environments
Climate stability and infrastructure resilience for high-uptime operations
Markets that combine life sciences density with digital infrastructure maturity offer the greatest long-term value.
Over the last five years, life sciences employment growth has been robust in all markets—rising 31 percent in emerging markets such as Phoenix and Salt Lake City, and 16 percent in secondary markets like Denver. Cities that prioritize science and innovation ecosystems cultivate an environment ripe for entrepreneurial ventures. By establishing innovation hubs, research centers and technology parks, they encourage collaborations between academia, industry, and startups. These ecosystems foster knowledge exchange, resource sharing and mentorship, giving birth to groundbreaking startups that have the potential to disrupt existing markets and create new ones. Creating space for life sciences presents an opportunity for up-and-coming markets. As the best-known research cities become saturated, costs of living continue to rise and with remote connectivity as the new normal, there is an opportunity for smaller cities to attract companies and talent with the promise of overall affordability.
Data centers are “power hogs”—which is why they’re measured in terms of electrical size (megawatts) vs physical size (square feet), but they are also “ground hogs”—with the typical site for a data center growing from 155 acres in 2022 to 224 acres in 2024, a 144 percent increase (and with the onset of AI—hyperscalers are looking at sites in the 500-800ac range and power demand in the order of 1GW vs 100MW for more traditional data centers). Power availability has become an increasingly pressing challenge in established data center markets such as Phoenix, limiting the market’s availability to meet demand—this provides opportunity for emerging and tertiary markets such as Iowa and Indiana where both land and power are available and affordable.
