8 Min. Reading Time
Saudi Aramco and California-based startup Groq Inc. are jointly building the world’s largest AI data center in Saudi Arabia. With 19,000 LPUs by the end of 2024 and a planned capacity of 200,000 accelerators, the center aims to cover half of the global AI computing demand. Groq’s chips will be used not only for language models but also for high-performance calculations in the classical oil and gas business.
The Essentials at a Glance
- Saudi Aramco and California-based Groq plan to build the world’s largest AI data center in Saudi Arabia.
- The center is expected to have 19,000 LPUs by the end of 2024, with a long-term goal of scaling up to 200,000 chips.
- The facility aims to cover half of the global AI computing demand, particularly for language models and industrial applications.
- Groq’s accelerators are based on the Language Processing Unit type and will also be used in the oil and gas sector.
- The project is financed with $2.8 billion, with Saudi Aramco contributing a single-digit billion-dollar amount.
Strategic Shift from Oil to AI
Saudi Arabia was synonymous with crude oil for decades. Today, it is becoming a flagship for AI in the Middle East and North Africa. The state-owned company Saudi Aramco, once the backbone of fossil energy supply, is driving a profound transformation with the construction of the world’s largest AI data center. The country’s dependence on oil, which has made it one of the richest countries in the world, is being systematically reduced. Instead, the focus is on digital infrastructure and artificial intelligence as new growth drivers.
The initiative is led by Aramco Digital, a subsidiary of Saudi Aramco. Aramco Digital has partnered with California-based startup Groq Inc. The partnership between Saudi Aramco and Groq is not just a technological deal; it’s a signal to the market. The end of the oil era is in sight, and the future belongs to AI. According to Bloomberg and Golem, the project aims to serve not only regional but also global AI requirements, with an ambitious goal of covering half of the global AI computing demand.
The decision to partner with Groq is strategically sound. The company specializes in developing AI accelerators of the Language Processing Unit type and has expertise in Application-Specific Integrated Circuit. With a valuation of $2.8 billion after just eight years, Groq Inc. is a player that combines technological maturity with scaling potential. For Saudi Arabia, this is not just an investment in hardware; it’s a step to reposition the world for AI.
The financial framework underscores the significance. The investment amount is in the billions of dollars, demonstrating the country’s willingness to commit substantial funds to infrastructure. It’s not just about computing power; it’s about securing the economic future beyond raw material extraction. The collaboration with a US startup also showcases the openness to international technology partnerships.
Technical Details and Location Advantages
The technical foundation of the data center is formed by Groq’s LPUs. These are in-house developed chips currently produced by Globalfoundries in Fab Malta. By the end of 2024, 19,000 of these accelerators are to be installed. This already represents a massive computing capacity. In the long term, scaling to 200,000 LPUs is planned. This quantity will then occupy around 2,800 19-inch racks.
For comparison: A single LPU rack can process thousands of inference requests per second. This is more efficient than many GPU clusters for certain workloads. The physical dimension of the project is enormous. 2,800 racks require a correspondingly large hall and a complex power grid. The logistics for transporting and installing such quantities of hardware are a challenge in themselves. Nevertheless, the plan is concrete and timed.
The geographical location of Saudi Arabia offers decisive advantages. Extremely low energy costs are possible. These are fed by oil and increasingly by solar energy. This enables cost-efficient operation that is hardly achievable in other regions. At the same time, the climate presents a significant technical challenge. Cooling in the desert requires innovative solutions.
Groq must provide large quantities of cooling water. This serves to optimize free-air cooling and avoid thermal overload. Without these measures, operating the hardware would be impossible. The ambient heat would drastically reduce the performance of the chips. Therefore, the water supply is just as important as the power supply. Groq’s ambitions go beyond Saudi Arabia.
In parallel to the project in the Middle East, the company plans to expand in Norway. There, 21,600 LPUs are to be installed by the end of 2024. By the end of 2025, 129,000 are already planned. This means a fivefold increase in just twelve months. Worldwide, Groq aims to install a total of 1.5 million accelerators. These numbers show: The partnership with Saudi Aramco is not a one-off project. It is part of a global infrastructure strategy that industrializes AI computing power.
Language Processors are not just for Language Models
The name Language Processing Unit might suggest that Groq’s LPUs are primarily optimized for language models like LLMs. However, the application is much broader. According to a whitepaper by the company, the accelerators are also suitable for highly complex, deterministic calculations. This is particularly relevant in industrial applications. This flexibility was crucial for the state-owned company Aramco to support the project.
The LPUs are characterized by extremely low latency. They also offer predictable performance. These properties are essential for real-time analyses in oil and gas extraction. In contrast to GPUs, which rely on parallelism, the LPU architecture works sequentially. It explicitly controls the data flow via software. This enables more efficient utilization for specific workloads.
Examples include seismic data analysis or the optimization of drilling processes. Here, every millisecond of delay counts. Unpredictable response times can cause expensive downtime. The determinism of the LPU ensures planning security. The variety of applications that can be addressed with LPUs makes them a strategic tool for Saudi Aramco. It’s not just about offering AI computing power.
It’s about optimizing their core competence. Fossil fuel extraction is improved with AI. The technology is thus not used as a replacement but as an enhancer of the existing business model. And exactly that makes the combination of Saudi Aramco and Groq so effective. They connect two worlds that seem contradictory at first glance. Crude oil and AI stand side by side.
“Groq’s LPUs are not just a pure language accelerator – they are a deterministic engine for high-performance AI that also plays to its strengths in industry.”
AI also benefits the classic oil and gas business
The integration of AI into the classic oil and gas business is not a side effect of the project. It’s at its core. Saudi Aramco uses the computing power not only to offer external AI services. The company uses the computing power to optimize its own processes in oil and gas production. For example, AI can help evaluate seismic data faster and more precisely. This optimizes the site selection for new boreholes.
Predicting deposits or corrosion in pipelines also becomes more efficient through machine learning. The accelerators for these tasks come directly from the new data center. They are operated with LPUs from Groq. The architecture enables models to be executed in real-time. There is no need to wait for the latency of cloud APIs. This is crucial for applications where every second counts.
Furthermore, local processing reduces dependence on external providers. It strengthens data sovereignty. However, the benefits do not remain limited to the technical level. The AI initiative also strengthens Saudi Arabia’s geopolitical position. By positioning itself as a hub for AI infrastructure, the country positions itself as a technology leader. This happens in a region that has so far hardly been in the focus of global AI development.
The partnership with Groq is more than just a joint venture. It’s a statement. AI and crude oil are not opposites. They are two sides of the same economic transformation. By using the latest hardware in the oil industry, Saudi Aramco shows that traditional sectors can benefit from innovation. The synergies lie in increasing efficiency and reducing costs.
According to Bloomberg, this is an important step for diversifying the economy. Golem also highlights that hardware development plays a key role here. The combination of local energy and global technology creates a unique competitive advantage. Other nations must now react to keep up.
Frequently Asked Questions
What is the goal of the AI data center in Saudi Arabia?
The goal is to build the world’s largest AI data center, which will ultimately cover half of the global AI computing demand. This applies to both language models and industrial applications in oil and gas production.
What technology is used in the data center?
The center uses Language Processing Units from the company Groq. These accelerators are based on a special architecture for deterministic, low-latency AI calculations and are used for both language models and high-performance analyzes in the energy sector.
How much is invested in the project?
The project is financed with a total of $2.8 billion. A single-digit billion amount comes from Saudi Arabia, provided by the state-owned company Saudi Aramco.
What challenges are there in realizing the project?
The greatest technical challenge is cooling the system in the hot desert region. Groq must provide large amounts of cooling water to improve air cooling and avoid thermal overheating.
How does Saudi Aramco benefit from this project?
Saudi Aramco reduces its dependence on the oil business and invests in the future technology AI. At the same time, the company uses the computing power to optimize its own processes in oil and gas production. This ranges from seismic analysis to pipeline maintenance.
Source of title image: Pexels / Greg Gulik (px:349419)
