Leap 2026 investment rally exceeds $18 billion to localize the computing and AI stack in Saudi Arabia
Listen to this article
Read by Anchor
The activities of the “Leap 2026” conference in Riyadh marked a qualitative shift in Saudi Arabia's technology strategy, moving the focus from merely attracting global firms to building and localizing the full stack of computing and artificial intelligence. Investment announcements and tech partnerships surpassed $15 billion on the first day alone, followed by about $2.5 billion on the second day and $857 million on the third day, covering advanced data centers, semiconductors, networks, language models and human-skill development.
Cloud infrastructure and heavy-computing projects dominated the conference agenda, with Amazon Web Services confirming that it will launch its first cloud region in the kingdom by December 2026, backed by investments exceeding $5.3 billion to run workloads locally. It also expanded its partnership with Hiwmaen to provide up to 50 megawatts of compute capacity by 2028, using Trinium chips and Nvidia’s infrastructure, while making the Arabic “Allam” model available through the Amazon Bedrock platform and offering the Hiwmaen Fabric platform in the Amazon cloud marketplace.
The rapid buildup of computing capacity is reshaping the regional data-center map, driven by major alliances with chip manufacturers and infrastructure developers.Hiwmaen has entered a partnership with Together AI to build a data center with 250 MW capacity and allocate 55 MW, alongside a partnership with AMD and Cisco to operate a live computing architecture based on Instinct MI355 X processors and EPYC units, aiming to reach 250 MW within two to three years and target 1 GW by 2030 through their joint project. In Oxagon, Neom, Hiwmaen and DataVolt have begun construction of the first phase of an AI-ready data center with 100 MW of the planned 360 MW slated for completion by 2028. Hiwmaen also launched an AI cloud powered by Nvidia HGX B300 Blackwell Ultra systems, operating at over 90 % of its current 1.1 MW capacity in preparation to raise it to 35.1 MW by 2027. Meanwhile, Al-Muammar Information Systems announced a $1.2 billion investment to raise its data-center capacity to 192 MW, and the National Housing for Innovation announced an $800 million injection to develop the Khuzam Digital Valley to a capacity of up to 65 MW by 2033.
The capital influx extended upward into the software and model layer, with Hiwmaen announcing a strategic investment in Arabic.ai and Tarjama to develop intelligent agents for managing translation, documents, contracts and tenders, and a partnership with Reflection AI to host foundational models locally and sovereignly customize them. The company is also collaborating with Quora Systems to develop the Hiwmaen OS operating system, aimed at enterprises on a Linux environment to automatically route tasks to the appropriate compute environment, while Microsoft broadened its alliance to bundle Hiwmaen One with Copilot 365 targeting one million enterprise users, and prepared to launch AI-enhanced Windows PCs in September.
In the expanded applications arena, Adobe committed to investments and partnerships worth more than $4 billion together with the Ministry of Communications and Information Technology and Hiwmaen, providing over 27 million citizens and residents aged 13 and above with free access to Firefly and Express Premium features for 12 months by the end of 2026, while developing a visual generation model tailored to the local cultural context and supporting Arabic commands.
Localizing infrastructure and data gives organizations in the Gulf and the region a tangible reduction in inference cost and response time.This massive influx of local compute capacity enables engineers or technical decision-makers in banks and regional firms to shift sensitive workloads from distant foreign clouds to local data centers that comply with data-residency requirements, while leveraging language and visual models trained on the Arabic cultural and linguistic context, thereby lowering licensing and consumption costs and accelerating the integration of intelligent agents into commercial and service operations.
The ecosystem integrated major initiatives to build human capital, including a partnership with Google to empower one million university students with Gemini tools, training 100 thousand staff through the Amazon Academy, and establishing a $100 million investment fund between Ola Capital and Korea’s Future Play to link deep-tech opportunities, completing the infrastructure chain from power and silicon to models and operational expertise.