By Venkat Ramaswamy and Krishnan Narayanan
The digital divide has long been an important measure of global inequality. Who is connected, and who is not? Who gets to produce, program, and profit in the Information Age? But a far more consequential rift now looms: the Intelligence Divide.
At LEAP 2025, and again in his keynote at IGF 2024, Saudi Arabia’s Minister of Communications and Information Technology, H.E. Abdullah Al-Swaha, issued a clear call: “The Intelligence Age must break the Digital Divide.”
But more important, it must not deepen it. And certainly not replace it with a new one—where AI capability, creativity, and control are monopolized by the few.
The solution, we propose, lies in the concept of Co-Intelligence—a vision of human–AI collaboration rooted in distributed intelligence, relational ecosystems, and sustainable value creation.
Let’s explore how Saudi Arabia is attempting to operationalize this through its AI infrastructure strategy—particularly via HUMAIN, PIF investments, and a broader governance reimagination—and what challenges must be addressed to truly democratize AI.
HUMAIN: Saudi Arabia’s AI Infrastructure Moonshot
HUMAIN, launched in partnership with NVIDIA and backed by Saudi Arabia’s Public Investment Fund (PIF), is not merely another data center project. It is a national initiative to build the AI infrastructure layer of the future, offering 500 MW of AI compute power initially and scaling to 1.9 GW by 2030.
Tareq Amin, CEO of HUMAIN notes that “HUMAIN is Saudi Arabia’s play to participate in the full AI value chain—from chips to models to solutions.”
Jensen Huang, CEO of NVIDIA, described these deployments as “AI factories”—facilities where energy is transformed into AI tokens, and those tokens are assembled into models, applications, and embodied intelligence. The implications go far beyond GDP or tech sovereignty. It creates the capacity for every industry—healthcare, education, energy, manufacturing—to locally adopt and adapt agentic and physical AI. And for local developers, researchers, and entrepreneurs to contribute to global AI ecosystems on their own terms.
From Infrastructure to Intelligence as a Public Good
This is not just an economic strategy. It’s a governance experiment. At IGF 2024, Al-Swaha outlined the necessity of AI Public Infrastructure (AIPI) as a next-generation evolution of Digital Public Infrastructure (DPI). These shared, open, secure platforms must address three new divides:
Compute Divide — only a handful of nations currently have the AI compute capacity to train or fine-tune large models
Data Divide — quality, contextualized data is unevenly distributed and often inaccessible to local innovators
Algorithmic Divide — pretrained models often encode biases, are optimized for high-resource languages, and reflect alien cultural values
This is precisely where the concept of Co-Intelligent Digital Public Infrastructure (Co-DPI)—as developed in The Co-Intelligence Revolution—offers a leap. Traditional DPIs were static, transaction-oriented platforms designed for service delivery. In contrast, Co-DPIs are dynamic, learning infrastructures powered by Tokenized Digital Intelligences (TDIs). They don’t just facilitate access; they enable co-creation of value, learning, and intelligence across ecosystems.
Built using the PIEX lens—Platforms, Interactive Engagements/Impacts in Ecosystems, and life-Experiences—Co-DPIs can:
Orchestrate multi-sided, multi-stakeholder interactions at scale
Facilitate feedback-rich, AI-powered public services across health, education, finance, mobility
Evolve through collective intelligence, rather than top-down programming
Ensure relational sovereignty—where governance is local, transparent, and co-owned by the people it serves
In Saudi Arabia’s context, Co-DPIs could empower citizens, public agencies, and developers to continuously shape and refine AI systems—grounded in local values, languages, needs, and aspirations. The iCARe ethics center, the Sedaya-UNESCO collaboration, and platforms like ALLaM and MetaBrain could serve as foundational modules in a larger Co-DPI stack, with AI infrastructure being made accessible to local developers via tools like the Qualcomm AI Hub and hybrid AI playgrounds—where agentic models can reason, interact, and evolve based on context. Saudi Arabia’s AI capabilities should also include and empower its women - according to Stanford's AI Index Report 2025, it was the only country where the relativeAI skill penetration rate was greater for women than men.
Why Co-Intelligence Is the Antidote
Co-Intelligence reframes AI not as a central brain but as a distributed relational capacity. It invites:
Human + machine synergy across platforms, ecosystems, and life-experiences
Open, diverse, and culturally grounded knowledge systems
Shared design and adaptive orchestration of AI through Tokenized Digital Intelligences (TDIs)
Applied in Saudi Arabia’s context, this means AI that reflects the values of Arabic-speaking societies, supports domain-specific industries like energy or logistics, and creates feedback loops of capability-building—from talent to trust to transformation.
This is how the Intelligence Divide can be prevented—not through trickle-down diffusion, but through Human-AI Co-Creation.
But Major Challenges Remain
Despite Saudi Arabia’s strong positioning, success will depend on tackling governance head-on with several key tensions:
1. Bias and Guardrails
H.E. Al-Swaha noted that it is important to ensure that no synthetic data is modeled that excludes one group over another, and that algorithms are helpful, honest, and harmless—especially across different cultures, languages, and values? What’s the line between safety and censorship?
2. Sustainability
AI’s energy demands are vast. If inference expands to edge devices and robotics, will we have the resilient, affordable energy infrastructure to support it without undermining climate goals?
3. Sovereignty vs. Openness
Can shared infrastructures also respect national data rights, cultural integrity, and decentralized control? Is AI sovereignty compatible with AI collaboration?
4. Capability Mismatch
Notwithstanding an admirable aspiration to close a “10 million-person skills gap” in AI, data science, and cybersecurity—especially in the Global South, how do we navigate a transition to AI infrastructure that lands even before talent pipelines are ready?
A Call for Co-Creating Inclusive AI Futures
Saudi Arabia’s AI ambition—powered by HUMAIN, PIF, Sedaya, and global partnerships—is a promising model. The Kingdom has jumped from just 7% women in tech to over 35%—surpassing the Silicon Valley, EU, and G20 averages. As H.E. Al-Swaha emphasized, this is not just a milestone for the region but a signal to the world. This gender empowerment forms a crucial piece of the broader strategy to close all divides—across geography, access, and opportunity. The Digital Cooperation Organization (DCO), representing 10% of the world’s population (800 million people), is leading efforts to create a truly inclusive digital future. Led by Dima Al Yahya, the DCO’s commitment to "leaving no one behind" is visible in its generative AI capacity-building initiatives across the Global South.
But massive disparities still persist: 2.5 billion people in the Global South remain unconnected; only 15% of the world’s economy is digital. Per capita digital value creation still skews heavily toward the Global North. And globally, there remains a shortage of 5 million cybersecurity professionals, 3 million AI-data specialists, and still a long road ahead in closing the gender digital divide.
Beyond AI Factories to AI Commons
Beyond building AI factories, we must also cultivate an AI commons—a shared space of learning, co-creation, and empowerment. And one of the most urgent frontiers in this commons is EcoAI Literacy.
As we propose in The Co-Intelligence Revolution, EcoAI Literacy is not merely about teaching people how to use AI, but fostering a new kind of ecological and ethical consciousness in how we co-design, govern, and engage with AI. It helps individuals, communities, and institutions:
Understand the energy and environmental implications of AI usage and infrastructure
Make informed choices about sustainable AI adoption, especially in edge and industrial contexts
Promote stewardship over ownership, aligning digital intelligence with local, planetary, and intergenerational well-being
Navigate trade-offs between compute, inference, and climate costs with relational awareness
By embracing EcoAI Literacy as part of its national curriculum and talent pipelines, Saudi Arabia can build not just a digitally fluent population, but a co-intelligent citizenry—capable of shaping AI systems that serve people, place, and planet.
We believe that by adopting the deeper ethos of Co-Intelligence—a commitment to equity, relational design, and dynamic learning across people, machines, and ecosystems—the Intelligence Age can mitigate a recreation of power asymmetries from the Digital Age, and even leap over them.
The future is not something we await—it is something we must co-create with AI, in partnership with one another and Earth.