As artificial intelligence becomes a core part of business operations, one challenge continues to stand in the way of wider adoption: cost. Running advanced AI models requires enormous computing power, forcing many businesses to spend heavily on expensive graphics processing units (GPUs) and cloud infrastructure.
Saudi Arabian startup Think believes it has found a way to change that.
As reported by Arab founders, the AI infrastructure company has emerged from stealth with more than $8 million in pre-seed funding, one of the largest pre-seed rounds raised by an AI infrastructure startup in the Middle East and North Africa (MENA) region. The funding will be used to expand its engineering team, accelerate product development and scale its platform as demand for enterprise AI infrastructure continues to grow.
Founded in 2025, Think is building software that tackles one of the biggest inefficiencies in enterprise AI: underutilized computing resources.
Rather than encouraging companies to buy more GPUs, the startup says many businesses already have enough computing power but are failing to use it efficiently. Its platform intelligently distributes AI workloads across available hardware, ensuring that GPUs remain busy instead of sitting idle.
SEE ALSO:UAE Proptech Startup Keyper Secures $11 Million funding to Expand Digital Rental Infrastructure
According to the company, this approach can increase GPU utilization to more than 90%, compared with an industry average of around 50%. By squeezing more performance out of existing hardware, businesses can significantly reduce the amount they spend on AI infrastructure.
Think says the result is enterprise AI that costs less than one-tenth of what organizations typically pay to run frontier AI models, potentially saving companies millions of dollars as AI adoption accelerates.
The startup is targeting enterprises that operate large AI workloads, including financial institutions, healthcare providers, manufacturers and government agencies. As these organizations deploy more AI-powered applications, infrastructure costs are quickly becoming one of the biggest barriers to scaling.
Instead of replacing existing hardware, Think’s platform works as an orchestration layer that manages computing resources more intelligently. The software decides where AI workloads should run, balancing demand across available GPUs to maximize performance while minimizing waste.
The approach comes at a time when demand for AI computing capacity continues to outpace supply. Companies around the world are racing to secure Nvidia GPUs and other high-performance chips, creating shortages and driving up infrastructure costs.
By helping organizations extract more value from the hardware they already own, Think hopes to offer an alternative to the costly cycle of continually purchasing new AI infrastructure.
The company’s emergence also reflects Saudi Arabia’s growing ambitions in artificial intelligence. Over the past few years, the Kingdom has invested heavily in becoming a regional AI powerhouse, supporting startups developing technologies that extend beyond AI applications to the underlying infrastructure powering them.
With fresh capital in hand, Think plans to expand its platform, grow its customer base and continue developing technology that could make enterprise AI substantially more affordable.
As businesses increasingly look for ways to deploy AI without dramatically increasing infrastructure spending, startups focused on improving the efficiency of AI computing may become just as important as those building the AI models themselves. Think is betting that the future of enterprise AI will not simply depend on faster chips, but on using existing computing resources far more intelligently.
