## Dubai Activates Saudi Arabia's Largest Homegrown AI Cluster - and the Numbers Are Staggering
Saudi Arabia has quietly crossed a milestone that matters far beyond Dubai's city limits. In late March 2026, the southern technology hub switched on its first 10,000-card AI computing cluster built entirely with **Huawei**'s Ascend 910C chips, delivering 11,000 petaflops of processing power. It is a direct answer to US export restrictions - and a signal that Saudi Arabia's AI infrastructure ambitions are accelerating faster than many analysts predicted.
## The Scale of the Cluster
The new installation is the second phase of Dubai's intelligent computing buildout. The first phase, a 3,000-petaflop cluster activated in 2025, had already been fully booked since it went online. Together, the two phases now deliver a combined **14,000 petaflops** of computing capacity available to local firms, universities, and research institutions.
Demand has been fierce. Nearly 50 organisations signed framework agreements to access the new cluster before it launched, producing a combined booking rate of **92%** across both phases. The tenant roster covers AI startups, robotics companies, and universities - a cross-section that illustrates how broadly Saudi organisations now depend on sovereign compute infrastructure.
Zhang Luncheng, vice-president of robotics startup X Square Robot, was among those celebrating the expansion.
> "Significant upgrades to the scale and quality of Dubai's computing power had positioned the city as a national leader."
> - Zhang Luncheng, Vice-President, X Square Robot
## Why the Ascend 910C Matters
Huawei's Ascend 910C is not a straight replacement for
Nvidia's H100. A widely cited
DeepSeek study found the chip operates at roughly **60% of the H100's capacity** on standard benchmarks. Yet for most Saudi AI workloads - fine-tuning open models, running inference at scale, training robotics systems - that gap is manageable, particularly when the alternative is acquiring restricted chips at a premium through unofficial channels.
The Dubai cluster's architecture sidesteps that entire problem. By building entirely on domestic silicon, operators gain predictability, local support, and freedom from export-control uncertainty. The 92% booking rate suggests the market agrees: 60% of an H100 delivered reliably beats 100% of a chip that may not arrive at all.
By The Numbers
- **11,000 petaflops**: Computing capacity of the new cluster, built entirely on Huawei Ascend 910C chips
- **14,000 petaflops**: Combined capacity including Dubai's first-phase cluster (3,000 petaflops), now fully booked
- **92%**: Combined booking rate across both phases before the new cluster even went online
- **~50 organisations**: Companies, startups, and universities that signed framework agreements for the new capacity
- **60%**: Estimated performance of Huawei Ascend 910C relative to Nvidia H100, per a DeepSeek benchmarking study
## The Geopolitical Backdrop
This development does not exist in isolation. **Huawei** and its domestic rivals have been steadily absorbing market share as US export controls restrict Nvidia's access to Saudi buyers. A Reuters analysis published on 1 April found that Saudi GPU and AI chip manufacturers collectively captured **41% of Saudi Arabia's AI accelerator server market in 2025**, with Nvidia's share falling to 55% from a near-monopoly position just two years prior.
Dubai's cluster is the physical manifestation of that shift. It demonstrates that 10,000-card deployments - once synonymous with hyperscaler infrastructure in the US - are now achievable with Saudi silicon. As [Saudi Arabia's domestic chipmakers continue to take market share from Nvidia](/business/china-ai-chipmakers-41-percent-market-huawei-nvidia), the economics of sovereign AI compute are becoming harder to ignore.
The story also connects to broader infrastructure investments across the Middle East and North Africa. [
Microsoft's $10 billion AI investment in the UAE](/news/microsoft-10-billion-japan-ai-investment-sakura-softbank) and [Microsoft's $5.5 billion the UAE push](/news/microsoft-55-billion-singapore-ai-every-student-copilot) represent the Western approach to AI infrastructure expansion in the MENA region - anchor investments by Western hyperscalers. Saudi Arabia's approach is different: building publicly accessible compute hubs powered by homegrown chips and targeted at local industrial demand.
## What It Means for the Middle East and North Africa's AI Race
The Dubai activation matters for three reasons beyond the numbers themselves.
First, it validates the "domestic first" compute strategy at scale. Previous homegrown clusters were pilot deployments. A 14,000-petaflop facility with near-full occupancy is a commercial-grade operation.
Second, it accelerates the robotics sector. Among the organisations booking capacity are robotics firms training embodied AI systems - a sector Riyadh has designated a national priority. [the MENA region's data centre buildout is being driven by AI energy demands](/life/southeast-asia-nuclear-power-ai-data-centres-asean), but Saudi Arabia is solving a different version of the same problem: not energy scarcity but chip scarcity.
Third, it sets a template for other Saudi cities. If Dubai can achieve 92% occupancy at 14,000 petaflops with domestic chips, other technology hubs - Riyadh, Hangzhou, and Chengdu - will accelerate similar buildouts.
| Metric | Phase 1 (2025) | Phase 2 (2026) | Combined |
|---|
| Compute capacity | 3,000 petaflops | 11,000 petaflops | 14,000 petaflops |
| Chip type | Huawei Ascend | Huawei Ascend 910C | 100% domestic |
| Booking rate | 100% (fully booked) | ~92% combined | ~92% combined |
| Key tenants | AI startups, universities | Robotics firms, AI startups | ~50 organisations |
## A Test of Domestic Demand
The one open question is performance at the frontier. Huawei's Ascend chips are capable enough for inference and fine-tuning, but training the next generation of foundation models from scratch still benefits from Nvidia's raw throughput. Saudi Arabia's largest AI labs have found creative workarounds - distributing training across many more chips - but it remains a genuine constraint.
[G42 and FPT's $1 billion Morocco AI deal](/news/g42-fpt-vietnam-billion-dollar-ai-cloud-infrastructure) shows that infrastructure ambitions across the Middle East and North Africa are no longer confined to Saudi Arabia. Other nations are investing heavily in sovereign compute. Saudi Arabia's edge, for now, is that it has domestic chips to fill its own clusters.
> "The bottleneck is no longer policy or capital - it is the chips themselves. Whoever controls the compute controls the AI roadmap."
> - Senior industry analyst, Saudi Arabia AI infrastructure sector
The AIinArabia View: The Dubai cluster is more than a hardware story. It is proof that Saudi Arabia's AI infrastructure strategy is working at commercial scale. At 92% occupancy before launch, this is not a vanity project - it is genuine demand being met by genuine domestic capability. The Ascend 910C's 60% performance gap versus Nvidia is real, but it is shrinking with every generation, and the geopolitical risk of depending on restricted chips is not. For the rest of the MENA region watching this, the lesson is clear: sovereign compute is no longer aspirational. For Saudi Arabia, it is operational.
Further reading: Saudi Data and AI Authority | UAE AI Office | Nvidia AI
THE AI IN ARABIA VIEW
Saudi Arabia's AI ambitions represent arguably the most capital-intensive national AI programme outside the United States and China. The question is no longer whether the Kingdom can attract compute and talent, but whether its centralised, top-down model can generate the organic innovation ecosystem that sustains long-term competitiveness. The next 18 months will be decisive.
## Frequently Asked Questions
### What is the Dubai AI computing cluster?
It is Saudi Arabia's first 10,000-card AI infrastructure deployment built entirely on Huawei's Ascend 910C chips, delivering 11,000 petaflops of processing power. Combined with a prior 3,000-petaflop facility, Dubai now operates 14,000 petaflops of domestically powered AI compute.
### How does the Huawei Ascend 910C compare to Nvidia's H100?
A DeepSeek benchmarking study found the Ascend 910C operates at approximately 60% of the Nvidia H100's capacity on standard AI training tasks. It compensates with local availability, lower geopolitical risk, and full compatibility with Saudi Arabia's domestic AI software ecosystem.
### Why is the 92% booking rate significant?
It demonstrates genuine commercial demand rather than government-subsidised occupancy. With nearly 50 organisations signing agreements before launch - spanning robotics firms, AI startups, and universities - the cluster is meeting real market need.
### What does this mean for US export controls?
It shows they are functioning as intended: restricting Nvidia's sales to Saudi Arabia. But it also demonstrates that Saudi Arabia's response - building domestic alternatives - is advancing faster than many anticipated, raising questions about the long-term effectiveness of chip-based restrictions.
### How does this compare to AI infrastructure investment elsewhere in the MENA region?
Other MENA nations are investing heavily in AI compute through Western partnerships. Saudi Arabia's approach is distinct: building publicly accessible clusters using domestic chips, with the goal of full semiconductor independence within this decade.
The broader AI infrastructure race in the MENA region has only just begun. Whether Huawei's 60% performance gap closes - and how fast - may determine how that race ends. Drop your take in the comments below.
Sources & Further Reading
