AMD, Saudi MCIT, and DCO Launch Open Developer Ecosystem to Advance AI Innovation

At LEAP 2026 in Riyadh, chipmaker AMD, Saudi Arabia’s Ministry of Communications and Information Technology (MCIT), and the Digital Cooperation Organization (DCO) formally launched a new ecosystem initiative aimed at accelerating AI innovation across the Kingdom and beyond. The announcement, published August 31 and confirmed by the Saudi Press Agency on September 1, goes beyond yet another memorandum of understanding: it pairs AMD’s open-source software stack and developer programs with MCIT’s national AI strategy and the DCO’s multilateral governance framework, creating what the three entities describe as an “open developer ecosystem” for AI.

What Was Actually Announced

The core of the initiative is a structured combination of three pillars, each contributed by a different partner:

• AMD’s AI Developer Program — the company’s integrated portfolio of AI-focused computing technologies, including access to its hardware platforms and the open-source ROCm software ecosystem for AI application development.
• DCO’s AI Ethics & Responsible Governance training — courses and tooling designed to embed responsible AI practices in the next generation of builders.
• MCIT’s national pipeline — connecting developers, startups, and digital talent to technical expertise, open innovation resources, and learning opportunities aligned with Saudi Vision 2030.

The inaugural workshop took place at LEAP 2026 on Monday, August 31, where attendees were introduced to the AMD AI Developer Program, the ROCm stack, and DCO’s governance curriculum. The choice of venue is deliberate: LEAP has become the flagship event for Saudi Arabia’s technology ambitions, and this initiative signals that the Kingdom’s AI strategy is now explicitly focused on the software and talent layer, not just data centers and chips.

Why It Matters: The Software Layer of Sovereign AI

For the past two years, Saudi Arabia’s AI story has been dominated by infrastructure headlines — multi-billion-dollar GPU clusters, sovereign data centers, and partnerships with hyperscalers. But infrastructure alone does not create an AI economy. The bottleneck for most emerging AI ecosystems is not compute; it is developers who know how to use it.

That is the gap this initiative targets. By anchoring the program in ROCm — AMD’s open-source alternative to NVIDIA’s CUDA — the partners are making a deliberate strategic bet. CUDA remains the dominant AI software platform, and its closed nature has become a geopolitical and economic chokepoint. AMD has been investing heavily in ROCm’s maturity (the stack received its own major update, ROCm 10.0, just days before LEAP), and getting national developer programs built on an open stack creates a pool of talent that is not locked into a single vendor’s ecosystem.

For AMD, the commercial logic is straightforward: every developer trained on ROCm is a future customer for AMD’s Instinct accelerators, and every startup that builds on the stack strengthens AMD’s position in a market where NVIDIA still holds the overwhelming share of AI compute. For Saudi Arabia, the initiative helps “translate Saudi Arabia’s investment in AI technology into new applications and services,” as the press release puts it — converting capital expenditure into domestic capability.

The Multilateral Angle: 16 Nations, 800 Million People

The DCO’s role gives the initiative reach far beyond Saudi Arabia. Founded in 2020, the Digital Cooperation Organization is the world’s first standalone intergovernmental organization dedicated to the digital economy, bringing together the ministries of communications of 16 member states representing nearly USD 3.5 trillion in GDP and a combined market of over 800 million people — more than 70% of whom are under the age of 35.

Following the Saudi inaugural engagement, the DCO will work with AMD and interested member states to explore expanding the AI capability-building and knowledge-sharing program. In practice, that means the ROCm training and governance curriculum piloted in Riyadh could be replicated across the DCO’s footprint, which spans the Middle East, Africa, and Asia. For smaller member states that cannot negotiate bilateral deals with chipmakers individually, the DCO offers a collective bargaining channel — and for AMD, a single partnership that opens 16 markets at once.

What Leaders Said

Dr. Lisa Su, chair and CEO of AMD, framed the initiative around builders: “The real power of AI will come from what people build with it. Together with MCIT and DCO, we are expanding access to AMD’s open software and AI platforms and helping developers, startups and engineers across Saudi Arabia build the skills and capabilities to turn ideas into real-world applications and services.”

His Excellency Abdullah bin Amer Alswaha, Saudi Arabia’s Minister of Communications and Information Technology, tied the program to national talent development: “AI leadership is built by empowering people to build, innovate and compete. With the support of our leadership, our collaboration with AMD and the Digital Cooperation Organization will expand access to advanced AI technologies, developer capabilities and open innovation ecosystems — enabling developers and startups in Saudi Arabia to turn ideas into solutions and scale them globally.”

Deemah AlYahya, Secretary-General of the DCO, emphasized the multilateral dimension: “Through our collaboration with MCIT and AMD, we are equipping developers, startups and digital talent with open technologies and world-class capability building opportunities while building lasting pathways to share knowledge, technology and expertise across DCO Member States.”

The Open-Stack Calculus

The initiative also lands at a moment when the AI software layer is becoming a contested strategic arena. Open-source inference stacks such as vLLM have made it dramatically easier to serve models on non-NVIDIA hardware, and AMD’s ROCm has matured to the point where national-scale deployments — like the AMD-HUMAIN collaborations announced over the past year — can realistically run large-scale AI workloads without CUDA. Training developers on these stacks early, at national scale, compounds that advantage over time.

There is a governance angle as well. By pairing technical training with DCO’s AI ethics and responsible governance course, the initiative embeds a policy framework directly into the developer pipeline — an approach that anticipates the direction of AI regulation in many DCO member states, where adoption is growing faster than local governance capacity.

What to Watch

The inaugural workshop is a starting point, not a finished program. Key indicators of whether this becomes substantive will be: concrete numbers on developers trained and startups engaged; whether additional DCO member states formally adopt the curriculum; and whether AMD’s developer cloud access is included as part of the program at meaningful scale. As with every LEAP announcement, the gap between ambition and execution is where these initiatives succeed or quietly fade.

Still, the direction is clear. Saudi Arabia is systematically building every layer of an AI stack — chips and data centers through HUMAIN, models through partnerships with OpenAI and others, and now, with AMD and the DCO, the developer talent layer on top. The open-source choice is the detail worth remembering: in the geopolitics of AI compute, the software stack is the new battleground, and Riyadh just picked a side.