17/06/25

The evolution of artificial intelligence, especially within the scope of generative AI, drives the demand for increasingly powerful and sustainable processing structures, particularly high-efficiency cooling Data Centers. As advanced models require ever-greater computing power, Data Centers must handle a growing volume of heat generated by servers.

This challenge is especially critical in infrastructures that use accelerators like GPUs and TPUs to train large language models and perform inference workloads. These environments demand innovative thermal solutions to ensure operational stability and energy efficiency.

As is widely known, high-performance AI servers operate under extreme thermal densities, making traditional cooling methods insufficient. Without effective cooling, overheating can jeopardize hardware reliability, shorten equipment lifespan, and dramatically increase energy consumption.

Beyond internal operational pressures, organizations are also dealing with a challenging climate scenario. Global warming and increasingly frequent heatwaves complicate server cooling, making it essential to develop environmentally sustainable strategies to preserve digital infrastructure resilience.

Learn more below:

Global landscape: AI expansion accelerates demand for high-performance computing infrastructure

The global AI infrastructure market is undergoing unprecedented expansion, with double-digit growth rates since 2019. According to IDC, in the first half of 2024 alone, investments in AI-specific servers grew by 105% compared to the same period in the previous year, accounting for an impressive 95% of total spending in the segment. During the same period, investments in AI-related computing and storage infrastructure reached $47.4 billion – a 97% jump compared to H1 2023.

This upward trend is backed by robust projections: the AI infrastructure segment is expected to exceed $200 billion in investments by 2028, largely driven by the growth of generative applications. Statista forecasts that this niche will move $62.72 billion by 2025, with a compound annual growth rate (CAGR) of 41.52%, reaching $356.05 billion by 2030.

These figures clearly point to a global race for computing capacity and efficient storage, especially to handle the massive data volumes and processing intensity of generative AI.

And the acceleration doesn’t stop there. Gartner estimates that by 2027, generative AI will account for 40% of initial network configurations in Data Centers – a significant leap from less than 5% in 2024. Additionally, Gartner analysts project that global IT services spending will hit $1.73 trillion in 2025, growing 9.4% over the previous year.

In this rapidly expanding scenario, high-efficiency cooling Data Centers are not only keeping up with growth but are becoming essential to ensure the performance, reliability, and scalability of the entire AI value chain.

Why high-efficiency cooling is essential in generative AI Data Centers

It is fair to say that the scalability of generative AI today depends directly on the thermal infrastructure of Data Centers. In addition to high energy consumption, AI servers operate under elevated temperature conditions, making standard heat dissipation methods insufficient.

As a result, conventional forced-air cooling is no longer enough. In AI-focused Data Centers, power densities per rack exceed 30 kW (and can reach up to 50 kW), requiring technologies capable of capturing and dissipating heat directly at the source.

Training large-scale models – such as LLMs (Large Language Models) – requires the continuous operation of thousands of parallel processing units, like GPUs and TPUs, under intense thermal loads. These environments quickly surpass traditional heat dissipation thresholds, demanding thermal solutions that combine precision, adaptability, and energy efficiency.

Without an efficient cooling system to counter hardware hyperthermia, overheating can not only compromise operational stability but also reduce the expected lifespan of complex, high-density equipment. Additionally, it can lead to unnecessary energy waste and cause environmental impacts due to heat concentration in the facility’s surrounding area.

Thermal efficiency: performance and sustainability

In this context, high-efficiency cooling Data Centers go beyond avoiding thermal throttling – they enable scalable operation without compromising system integrity or computational performance.

In addition to technical efficiency, cooling’s impact on PUE (Power Usage Effectiveness) has become a critical metric for sustainable operations. Traditional systems fail to keep pace with the variability and unpredictability of AI workloads, resulting in energy waste, increased operational costs, and a higher environmental footprint.

Advanced thermal strategies – such as active thermal containment, adaptive cooling, high-efficiency air systems, and hybrid air-liquid technologies – are redefining cooling as a competitive differentiator in AI-focused Data Centers.

---

READ ALSO: Understanding the Impact of Artificial Intelligence on Data Centers

---

AI in Brazil: accelerated growth and outlook

The Brazilian artificial intelligence market is gaining momentum, propelled by an increasingly mature tech ecosystem. According to Statista, the sector is expected to reach $4.8 billion in 2025, with a CAGR of 27.77% through 2030, totalling $16.34 billion.

This growth is supported by factors like increased digitalization, advancements in robotics, smart sensors, computer vision, and the expansion of generative AI, which is gaining traction across various industry sectors.

The rising demand for responsive, personalized solutions – such as AI-powered chatbots and virtual assistants – is driven by the growing connectivity of Brazilian consumers. Industries like healthcare, finance, and retail are leading AI adoption, attracted by the real-time capabilities and data-driven experiences these tools offer.

This transformation reflects a structural shift in consumer behavior and corporate strategies aligned with digital innovation in Brazil.

The national market also benefits from favorable regulatory and economic factors. The Brazilian government has implemented public policies to promote AI, especially in public services and digital management.

Coupled with a diversified economy, widespread internet access, and a relatively supportive regulatory environment, Brazil is emerging as an AI innovation hub in Latin America.

For both global and regional players, Brazil already represents a unique strategic opportunity for experimenting, scaling, and monetizing AI-driven solutions.

Brazil’s role in the global AI Data Center landscape

The strengthening of the artificial intelligence market is closely tied to advances in digital infrastructure. As demand for AI-based solutions grows, so does the need for robust, scalable, and efficient computing environments – a role fulfilled by Data Centers.

In this context, Brazil has stood out on the international stage as a strategic Data Center hub, thanks to its ability to attract infrastructure investment, growing digital maturity, and privileged geographic location in Latin America.

Statista estimates that the sector’s revenue will reach $5.63 billion in 2025, with a CAGR of 5.82% through 2029. Grand View Research projects the Brazilian Data Center market to hit $6.84 billion by 2030, driven by demand for cloud services, secure data storage, and low-latency connectivity.

Favorable infrastructure

With network infrastructure expected to account for over $3.6 billion by 2025, Brazil is on track to become the leading Data Center hub in Latin America. This progress is supported by local strategic factors: political stability, abundant natural resources, and growing renewable energy production. These attributes position Brazil as a viable and resilient alternative to global sustainability challenges, reinforcing the need for new developments to prioritize energy efficiency – like high-efficiency cooling Data Centers.

Beyond technical and economic foundations, the regulatory environment is also favorable. The country’s rapid digitalization, combined with the growth of e-commerce and online financial services, has increased the demand for secure, reliable data management and storage solutions. Data governance, compliance with laws like LGPD, and broadband infrastructure investments are essential pillars supporting this growth.

With over 210 million inhabitants and high internet penetration, Brazil is fertile ground for the next generation of high-performance, sustainable Data Centers.

---

READ ALSO: Infrastructure for artificial intelligence: how Data Centers are preparing for the future

---

DC SP04: ODATA’s new Data Center raises the bar for efficiency and sustainability in greater São Paulo

Opened in early 2025, DC SP04 marks a strategic milestone in ODATA’s expansion in Brazil. Located in Osasco, greater São Paulo, the new campus received a total investment of R$2.6 billion and will feature 48MW of IT capacity.

It will be the company’s fifth Data Center in the country and the first to implement Delta³ – a high-efficiency air cooling system developed by ODATA’s U.S.-based parent company, Aligned Data Centers.

This innovation comes alongside a firm environmental commitment: the facility will be fully powered by 100% renewable, self-produced energy.

---

“We are honoured to operate in Brazil with 100% renewable, self-produced energy, reinforcing ODATA’s ongoing commitment to sustainable and responsible energy practices. We are always looking for ways to maximize our services and promote sustainability.”

Ricardo Alário, CEO of ODATA

---

“Delta³ technology enhances our portfolio by offering clients rapid scalability, cutting-edge innovation, and the highest standards of energy efficiency,” adds Ricardo.

In addition to its technological innovation, DC SP04 is designed to be an AI-ready Data Center, with flexible infrastructure to accommodate either air or liquid cooling systems. This enables support for high-density racks, including GPU-based servers used in advanced AI applications.

The option for direct-to-chip liquid cooling reflects the architectural evolution of cloud environments, which were previously CPU-centric but are now increasingly driven by AI processing demands.

The launch of DC SP04 will generate positive impacts for the local economy and Greater São Paulo’s tech ecosystem. By addressing a critical capacity gap in the Osasco region, the facility strengthens the city’s role as a digital hub. The project combines high energy availability with best-in-class IT infrastructure, supporting job creation, attracting new investments, and accelerating digital transformation.

With this development, ODATA reaffirms its leadership in the evolution of sustainable Data Centers in São Paulo, fully equipped to meet the demands of the next computing era.

---

READ ALSO: Data Center cooling technology: maximizing energy efficiency

---

Delta³ Redefines Cooling Standards for High-Density Data Centers

Delta³ (pronounced “Delta Cube”) is a proprietary innovation developed by Aligned, ODATA’s parent company, designed specifically for mission-critical environments with high thermal requirements. It is a patented high-efficiency air cooling system capable of supporting power densities up to 50kW per rack.

Unlike traditional solutions that simply push cold air into rooms, Delta³ targets and removes heat directly at its source. This approach delivers greater energy efficiency and enables a more dynamic and adaptive thermal response to fluctuating client workloads.

This design allows high-efficiency cooling Data Centers to achieve significant gains in scalability and space utilization. The solution notably reduces overall cooling demand, lowering electricity consumption and shrinking the operation’s carbon footprint.

Furthermore, Delta³ easily integrates with liquid cooling systems, making it ideal for high-performance computing applications involving generative AI and GPU-intensive servers. This level of flexibility is essential for Data Centers handling critical and unpredictable workloads, such as those found in modern AI cloud architectures.

A strategic advantage is that Delta³ is manufactured in Brazil, reducing acquisition costs and accelerating implementation timelines – crucial factors for projects with aggressive delivery targets. At the same time, local production strengthens the national innovation ecosystem, driving investment and advancement in critical infrastructure technologies.

With this solution, ODATA enhances its technological delivery capabilities, offering clients a high-performance, next-generation cooling system aligned with global benchmarks in energy efficiency and sustainability.

---

Want to learn how high-efficiency air cooling can benefit your Data Center?