Groq is an Artificial Intelligence (AI) organization developing a novel and disruptive architecture and software designed to accelerate workloads at scale with ultra-low latency. Its LPU Inference Engine is built to process large language models at high speed. Groq seeks to enable all businesses to join the burgeoning AI economy, and its real-time AI inference technology has been used across finance, industrial automation, cybersecurity and scientific research for leading government labs. The company is growing rapidly as both a business and a technology.

Groq's growth created a demanding infrastructure problem. In the rapidly evolving, highly competitive world of AI, simply keeping up with computing demand is not enough. Interruptions in a real-time model would mean it is no longer truly real-time, so the company's colocation had to match the performance of its high-speed LPU Inference Engine.

Groq's data center sources were scattered, and it needed to consolidate them into a single high-powered facility. That facility had to scale effortlessly with shifting business needs, deliver ultra-high densities so Groq could grow wherever the AI market expanded, and offer liquid cooling capability for the future high-powered computing requirements of its evolving AI suite.

Security and compliance were equally important. Groq required robust on-site security, including two-factor authentication and controlled facility access, for itself and for its customers, along with monitoring and maintenance by technology experts around the clock. With Groq's CEO Jonathan Ross describing the AI industry as being in a space race, the company could not afford a partner that was just getting started.

TierPoint began with a comprehensive assessment of Groq's needs, pinpointing exactly what the company required in a data center partner to meet its users' needs now and as they continue to grow and change. Groq chose TierPoint's data center in Spokane, Washington, to house its LPU Inference Engine and support its ability to process large language models.

The Spokane facility provides ultra high-density colocation with rack densities that handle the most demanding workloads, including AI accelerators and advanced GPU computing. Hot and cold aisle configurations support Groq's current generation of systems, with the ability to add liquid cooling modules as its systems evolve and require more power and cooling. TierPoint's experts also supply data-driven recommendations to boost system performance and reduce costs through power optimization.

Security is built in through hardened on-site controls, dedicated cages for Groq's systems and comprehensive compliance certifications across TierPoint's data centers. Maintenance requests are handled swiftly, and TierPoint experts are on hand 24/7 to answer questions about maximizing the power of the facility. TierPoint operates 40 data centers in 20 markets connected by a coast-to-coast network and says it has been ready for high-density AI workloads for more than a decade.

• Scattered data center sources consolidated into a single high-powered TierPoint facility in Spokane, Washington
• Ultra high-density racks support the LPU Inference Engine, AI accelerators and advanced GPU computing
• Hot and cold aisle configurations in place today, with liquid cooling modules available as systems evolve
• Data-driven power optimization recommendations boost system performance and reduce costs
• Dedicated cages, hardened on-site security and compliance certifications protect Groq and its customers
• TierPoint experts available 24/7, with maintenance requests handled swiftly