The global data center industry is undergoing a profound transformation, propelled by artificial intelligence, cloud migration, and exploding digital workloads. For Salesforce (CRM), whose software-as-a-service model fundamentally depends on robust, scalable, and cost‑efficient data center infrastructure, the dynamics captured in this analysis are of critical strategic importance. The data reveals an industry in hyper‑growth mode, but one that is simultaneously confronting severe capacity bottlenecks, escalating costs, environmental scrutiny, and an unprecedented wave of public and regulatory resistance. These trends, when viewed collectively, suggest that the infrastructure underpinning Salesforce’s platform is poised for a period of constrained supply, rising costs, and architectural innovation—all of which will influence the company’s operating margins, sustainability commitments, and global expansion.
A Construction Boom Fueled by AI
Data center construction spending surged by 26% in 2025 4, marking an acceleration that has continued into 2026. The United States alone hosts approximately 3,000 operational data centers, with a staggering 5,500 additional facilities currently under construction 13. This building frenzy is reflected in transaction volumes, which grew by about 52%—from $48 billion to $73 billion 34. Long‑range forecasts project the global converged data center infrastructure market to reach $49.34 billion by 2035 1,19, and European datacenter IT load, currently around 10 GW, is expected to more than triple to 35 GW by 2030 7. Such rapid expansion underscores the foundational role of data centers in the digital economy and creates a highly competitive environment for securing capacity—a factor that directly affects Salesforce’s cloud partners and, by extension, its service delivery costs.
Escalating Costs and the Modular Alternative
The economics of data center construction are shifting. Traditional “stick‑built” primary‑market facilities now cost $11–13 million per megawatt (MW) to construct 34. In contrast, factory‑built modular designs offer 45–55% cost savings, at $5–7 million per MW 34, and can be commissioned in just 90 to 120 days 34. Hyperscalers are projected to spend roughly $700 billion on infrastructure over the next few years 11,12, and behind‑the‑meter power generation has emerged as a major trend as developers seek to bypass grid constraints 31. For Salesforce, while it does not build its own data centers, the capital intensity faced by its infrastructure partners is likely to translate into higher colocation or cloud service costs over time, especially if supply tightens further. Modular solutions may offer a path to more cost‑predictable expansions, but their market share remains small.
The Energy and Water Challenge
Data centers are voracious consumers of electricity. Globally, if treated as a nation, they would rank as the 11th largest electricity user, consuming an estimated 448 terawatt‑hours in 2025 8,17. In the United States, data center electricity consumption already accounts for 4.4% of total demand 23 and is projected to triple to 12% by 2028 18,23. Ireland’s data centers already consume 20% of the country’s electricity 22. Cooling alone typically represents 38–40% of a facility’s energy use 6,7, while liquid cooling adoption can push mechanical, electrical, and plumbing (MEP) costs above 80% of construction expenses 10. Water consumption is equally staggering: traditional cooling towers evaporate around 2.6 million gallons of water per year per megawatt 26, and hyperscale facilities use 11,000–19,000 cubic meters daily 7. Importantly, roughly two‑thirds of a data center’s water footprint occurs upstream—in power production and semiconductor manufacturing 27. These metrics expose Salesforce’s supply chain to material environmental risks and may complicate the company’s Net Zero ambitions, especially as over 40% of new power for data centers by 2030 is expected to come from fossil fuels 27.
Pushing Boundaries: Floating, Subsea, and Space Data Centers
Land and grid constraints are pushing the industry toward unconventional architectures. Floating data centers, championed by shipbuilders like Samsung Heavy Industries, have gained momentum: a 50 MW‑class standard model received Approval in Principle from ABS and Lloyd’s Register in April 2026 9, and numerous global shipowners have made collaboration inquiries 9. However, these floating facilities face significant hurdles, including server instability from vibration, salt‑water corrosion, bandwidth limitations, regulatory permitting, ocean heating concerns, and unit capital costs of $100–200 million 9. Subsea deployments have been operationalized: Shanghai HiCloud Technology’s 24 MW facility offshore Lingang uses seawater cooling and offshore wind power 9, claiming a 22.8% reduction in power consumption compared to land‑based counterparts 9. Yet earlier pilots like Microsoft’s Project Natick are seen as commercially unviable 9. Space‑based data centers, despite fanciful projections, face insurmountable heat‑rejection physics—a single 20 kW rack requires a 20 m² radiator in constant shade 16—and are dismissed as at least 15 years from realization 3. For Salesforce, these novel form factors are unlikely to have a near‑term impact on its infrastructure footprint, but they underscore the sector’s desperate search for scalable, low‑cost capacity. Should floating or subsea models mature, they could offer geographically flexible, low‑latency nodes in coastal regions—relevant for edge computing and global connectivity.
Regulatory and Community Resistance
A powerful new constraint is the surge in local opposition. A Gallup survey found that 70% of Americans oppose data center construction in their communities 2,23. This sentiment has translated into concrete action: projects worth $64 billion have been blocked or delayed in the U.S. 23, with 20 such cancellations occurring in Q2 2025 alone 23. Several states are fast‑tracking permits for AI data centers—Mississippi, Louisiana, Wisconsin 5—while others are attempting to limit or ban new builds 14,29,32,33. Pennsylvania has permitted only 3 of approximately 100 proposals 24, and Governor Shapiro labeled over 90% of them as speculative 24. This environment is delaying grid interconnections: about 50% of completed data centers are waiting for power hookups, with years‑long wait times 15. For Salesforce, which relies on a geographically dispersed network of cloud providers, these headwinds could slow expansion into key markets and elevate colocation pricing. Moreover, the controversy could spill over into brand perception, especially for a company that heavily markets its sustainability credentials.
Labor Shortages and Execution Bottlenecks
The construction boom is colliding with a severe labor shortage. In 2025, 45% of data center contractors experienced project delays due to workforce gaps 10. The talent pool is aging: 32% of data center engineering personnel are over 60, while only 16% are under 30 10, and roughly 23,000 senior engineers retire each year 10. Compounding this, data centers compete with power grid buildouts and reshored manufacturing for 39 identical job categories 10. The overall U.S. construction industry faced a shortfall of approximately 439,000 workers at the end of 2025 10. Consequently, only about 5 GW of the 16 GW of capacity planned for 2026 operation is actually under construction 10. These execution risks are acute for hyperscale projects, where schedule slips can cost tens of millions per day 28. For Salesforce, any delay in its cloud partners’ data center projects could hinder its platform capacity to onboard new customers or support AI workloads, potentially affecting revenue growth and customer satisfaction.
Financialization and Market Valuations
The data center sector is increasingly financialized. Public REITs like Equinix (EQIX) and Digital Realty (DLR) trade at significant premiums to net asset value 34, reflecting strong investor appetite. AI data center construction debt is being packaged and sold as Baa‑rated investment‑grade bonds 30, indicating that capital markets are comfortable with the asset class despite rapid technological obsolescence—AI infrastructure assets are expected to depreciate within 2–4 years 12. Cost segregation tax strategies allow operators to accelerate depreciation and generate immediate tax shields 34. For Salesforce, though not a direct investor, this financialization suggests that the supply of capital for data center expansion remains robust, but also that investors are demanding premium returns, which could flow through to lease rates and ultimately to Salesforce’s operating expenses.
Strategic Implications for Salesforce
Salesforce’s entire business model—delivering customer relationship management, analytics, and AI agents via the cloud—rests on the availability of reliable, efficient, and affordable data center infrastructure. The cluster reveals a market in the throes of a super‑cycle, yet one riddled with constraints that are likely to persist for years. The direct implications can be mapped along three dimensions: cost structure, sustainability, and strategic flexibility.
Cost Pressures. The confluence of rising land, construction, power, and labor costs points to an upward trajectory for colocation and cloud service pricing. Hyperscale partners will need to pass through some of these increases, particularly in primary markets like Northern Virginia, which remains the world’s largest data center cluster 9,25. Modular construction offers a counterweight, and the migration toward lower‑cost states such as Oklahoma, North Dakota, and Wyoming 9 could moderate expenses in the long run. Nevertheless, in the near term, Salesforce may face higher infrastructure costs, which could pressure its gross margins unless offset by productivity gains or price increases.
Sustainability and Reputation. Salesforce has staked out ambitious climate goals, but the data center industry’s environmental footprint—from massive water use 7,26 to growing reliance on fossil fuels 27—creates a conflict. While hyperscalers are investing in renewable energy (Microsoft alone has secured 34 GW of clean power contracts across 24 countries 20), the reality on the ground often involves diesel backup generators and grid power from mixed sources. The increasing public backlash and the formation of initiatives like the Global Urban Data Centres Pact 21 mean that any association with unsustainable infrastructure could attract regulatory and reputational risk. Salesforce must actively engage with its infrastructure providers to ensure alignment with its sustainability commitments, or risk being ensnared in the growing data center opposition.
Strategic Flexibility. The rise of floating, subsea, and modular data centers opens new frontiers. While still nascent, these models could enable deployments in areas with limited land or grid access, potentially supporting Salesforce’s edge computing and low‑latency applications. However, the technical and regulatory hurdles remain substantial 9, and the capital intensity is high 9. More immediately, the industry’s pivot toward liquid cooling and higher rack densities (planned at 60 kW per rack 35) aligns with the GPU‑heavy workloads that AI demands, which Salesforce is aggressively pursuing with its Einstein GPT and autonomous agents. This transition will require infrastructure partners to retrofit or build new facilities, potentially creating short‑term capacity crunches but long‑term efficiency gains.
In sum, the data center sector is both a tailwind and a headwind for Salesforce. It provides the essential substrate for growth but introduces a multidimensional risk profile that the company must navigate through careful partner selection, contractual safeguards, and proactive sustainability initiatives. Infrastructure capacity will remain tight for at least 2–3 years, driven by labor shortages, grid interconnection delays, and local opposition, which could elevate cloud service costs and limit Salesforce’s ability to rapidly scale AI workloads in certain geographies. The shift to modular, high‑density, liquid‑cooled facilities is accelerating; while ultimately more efficient, the transition may disrupt existing capacity and require close collaboration with partners to ensure seamless access to next‑generation infrastructure. Environmental and regulatory risks are intensifying—Salesforce’s Net‑Zero commitments could be undermined by a supply chain still heavily reliant on fossil fuels and water‑intensive cooling, making proactive investment in sustainable practices essential to protect brand equity. Finally, the financialization of data center assets creates both opportunity and risk: the influx of institutional capital and innovative tax structures could spur more supply, but premium valuations and short depreciation horizons may lead to boom‑bust dynamics that indirectly affect Salesforce’s long‑term infrastructure stability.