The geography of artificial intelligence is shifting. Luma AI, the San Francisco startup known for its video generation models, just announced a $900 million Series C round led by HUMAIN, Saudi Arabia's sovereign AI company. More significant than the funding itself is where that money is flowing: Luma AI will begin running workloads on Saudi Arabia's Project Halo supercluster within weeks.
This is not just another funding announcement. It represents a fundamental rearrangement of the AI compute landscape, one that positions the Gulf region as a serious contender in frontier AI development.
What Project Halo Actually Is
Project Halo is a 2-gigawatt AI supercluster under construction in Saudi Arabia. To put that in perspective, most large data centers operate in the 50-100 megawatt range. This facility will have 20 to 40 times that capacity, making it one of the largest compute installations on the planet.
HUMAIN, which operates under Saudi Arabia's Public Investment Fund (PIF), is building this infrastructure as part of the Kingdom's strategy to become the world's third-largest AI hub after the United States and China. The investment is substantial: Saudi Arabia has committed over $100 billion to AI-related infrastructure projects through 2030.
For Luma AI, access to Project Halo solves an immediate problem. Training frontier video generation and "world models" requires enormous computational resources. The company's next generation systems will process video, audio, image, and language data at what they call "peta-scale," handling 1,000 to 10,000 times more information than current large language models.
Why This Matters for the Region
I have watched Gulf nations invest in AI for years, often in ways that seemed disconnected from actual AI development. Conferences, research centers, and partnerships that generated press releases but little technical output. This partnership is different.
Luma AI brings real technology to the table. Their video generation models already power commercial products. Their Dream Machine attracted millions of users. When Amit Jain, Luma AI's CEO, says HUMAIN is "the perfect partner for this next stage in Luma AI's explosive trajectory," he is talking about access to chips and compute that would be difficult to obtain elsewhere.
The announcement also includes something often missing from Gulf AI initiatives: a clear technical deliverable. HUMAIN Create, described as "the world's first Saudi-built, Arabic-native foundation model," will be developed through this partnership. Rather than simply licensing Western technology, the collaboration aims to produce sovereign AI capabilities trained on Arabic and regional data.
The Chip Supply Chain Angle
One aspect of this deal deserves particular attention. Luma AI expects to acquire up to 400,000 AI chips over the next four years through this partnership. These are likely NVIDIA H100s or their successors, the same chips that have been in chronic short supply globally.
How does Saudi Arabia have access to this volume of chips when companies in Silicon Valley struggle to secure allocations? The answer lies in recent US export policy decisions. The Biden administration cleared major chip sales to Gulf allies, and the Trump administration has continued this approach, recognizing that blocking allied access to AI hardware would simply push partners toward Chinese alternatives.
For startups like Luma AI, this creates an unexpected opportunity. Rather than competing with hyperscalers for limited US-based chip supply, they can access capacity through Gulf partnerships. The compute exists where the chips are available.
Implications for AI Practitioners in the Gulf
Several developments from this announcement are worth tracking:
The talent pipeline question. Luma AI is opening a Riyadh office to support client engagement and regional development. This creates new job opportunities for AI practitioners in the Kingdom, but also raises questions about whether the region has enough trained engineers to staff these facilities at scale.
Enterprise AI services. HUMAIN Create is explicitly designed to serve "enterprises and governments across MENA" with culturally aligned AI solutions. For organizations evaluating AI vendors, this means Arabic-native models trained on regional data may become available through local infrastructure rather than through foreign cloud providers.
The world model research trajectory. Luma AI's focus on "world models," systems that understand physical reality through video and multimodal data, positions them at the frontier of AI research. Having this work conducted partially in the Gulf creates opportunities for regional researchers to engage with cutting-edge development.
Infrastructure standards. Project Halo's 2-gigawatt scale sets a new benchmark for what AI infrastructure looks like. Data center developers across the region will be watching how this facility handles power distribution, cooling, and operational challenges.
The Competitive Landscape
Saudi Arabia is not the only Gulf nation pursuing AI infrastructure dominance. The UAE continues to invest heavily through G42 and its partnerships with Microsoft and others. Qatar has positioned itself as an AI conference and collaboration hub. Bahrain and Oman are developing more targeted AI strategies suited to their scales.
What sets the Saudi approach apart is raw ambition in infrastructure scale. Project Halo represents a bet that compute availability will be the primary constraint on AI progress, and that nations controlling significant compute will have leverage in the AI economy.
Whether this bet pays off depends on factors beyond infrastructure alone. Software expertise, research talent, regulatory frameworks, and commercial ecosystems all matter. But having 2 gigawatts of AI compute is a meaningful starting point.
Looking Forward
The Luma AI partnership represents a template we may see repeated. US AI startups hungry for compute, Gulf sovereigns with capital and chip access, and a mutual interest in developing AI capabilities outside the traditional Silicon Valley to hyperscaler pipeline.
For those of us building AI applications in the Middle East, this creates both opportunities and questions. The opportunity is access to frontier AI development happening in our region for the first time. The question is whether we can develop the supporting ecosystem of talent, regulation, and commercial infrastructure to capture lasting value from these investments.
The next few months will reveal whether Project Halo delivers on its promise. But the signal is clear: the Gulf is no longer content to be a consumer of AI technology built elsewhere. The infrastructure to build frontier AI locally is under construction.
Sources: