Riyadh negotiates Silicon Valley to run 14 gigawatts through Hyumann as part of the computing, capital and markets equation

The Saudi Ministry of Communications and Information Technology presented its strategy to compete with traditional cloud computing centers during the 2026 Innovation Leaders Forum in Silicon Valley, held under the slogan “Artificial Intelligence Forward: A Global Horizon.” Minister Abdullah Al-Sawaha told participants that the Kingdom is working with the private sector to develop computing capacity exceeding 14 gigawatts, aiming to lower operating costs for advanced AI companies and increase their efficiency amid rising global demand for processing power. The initiative relies on a unified model that combines infrastructure provision, capital injection, and opening market and client access channels, offering an integrated alternative for firms that usually turn to Western cloud providers.

The bet does not stop at building data centers; it also extends to integrating the national operator “Hyumann” into the core of software and model supply chains.The talks in California included a meeting with Sam Altman, CEO of OpenAI, to discuss deploying advanced models on the Hyumann infrastructure and developing national applications for priority sectors. Those discussions coincided with conversations that Al-Sawaha held with Michael Dell on integrating Dell AI Factory solutions, and with Scott Wu, CEO of Cognition AI, on expanding self-software-engineering tools and deploying the “Devin” agent, alongside a review of storage technologies with Kyle Fokuda of Solidigm to improve data-retrieval efficiency within the computing centers. These moves build on practical precedents that involved supporting the chip company Groq and Hyumann’s partnership with XAI to run Groq models locally.

The attraction of venture-capital funds in Silicon Valley was not aimed at financing construction and electrical-installation work, but at ensuring that software workloads are shifted to Saudi servers once they are completed. The rounds featured meetings with Himant Taniga, CEO of the General Catalyst fund, Shamat Balihabitiya, founder of Social Capital, as well as Ben Horowitz, co-founder of Andreessen Horowitz, and Alfred Lin, partner at Sequoia Capital. The message to these funds focuses on persuading portfolio start-ups to run their models and products on the Hyumann infrastructure, given that the project’s core financing relies on the state and local private sector rather than on external investment capital.

Expanding the infrastructure base takes a deeper step toward chip engineering, cooling, and ultra-high-speed interconnect networks.The meetings included discussions with Mark McQuaid of RCAI to deploy open “Trinity” models and custom enterprise models, and with Gabriel Kreman of Ingram to develop contextual learning and continuous-memory applications. On the hardware side, Charles Liang of Supermicro discussed accelerating the rollout of high-density servers that use liquid cooling, and Baron Car of Ascile AI explored high-performance network interconnects, while Rajiv Khimani of Velaura AI presented energy-efficient computing solutions, and Suk Hwan Lim of TenX Semi addressed the use of intelligent agents in chip design and verification to reduce reliance on imported design tools.

This shift directly affects decision-makers and technical engineers in the Gulf, Egypt and the Levant; localizing the hosting of large inference models and self-software-engineering tools reshapes operating costs for regional companies. If they lead a technical team or build an enterprise system, making computing workloads and agents available locally eases response-time constraints and compliance requirements for keeping data within regional borders, reducing the need to route data through U.S. or European servers, while simultaneously requiring development teams to acquire advanced skills in long-term system memory management and handling high-intensity processing environments.

These understandings, according to the forum’s official coverage, remain broad exploratory discussions that have not disclosed any binding deals, specific investment figures or final implementation timelines. The region’s real test is not measured by the gigawatt numbers announced at conferences, but by the national operator’s ability to integrate this complex mix of model providers, hardware and cooling networks into a single operational platform that proves its efficiency to developers and companies on the ground.