HUMAIN adds Arabic AI, live compute and a NEOM data centre

In one week, HUMAIN announced plans to bring its ALLAM Arabic models to Microsoft Foundry and Microsoft 365 Copilot, switched on AMD and Cisco AI infrastructure for customers in Saudi Arabia, and began building a 100 MW data-centre phase with DataVolt at Oxagon.

Microsoft plans to bring ALLAM to Foundry and Copilot

Microsoft and HUMAIN announced a long-term strategic collaboration to support AI adoption in Saudi Arabia and beyond. Its first planned technology milestone is ALLAM, HUMAIN’s family of Arabic language models.

The companies intend to make ALLAM available through Microsoft Foundry, so developers and enterprises can build, customise and deploy AI applications and agents on it. They also plan to let organisations create specialised agents in Microsoft 365 Copilot that draw on ALLAM’s Arabic capabilities.

Beyond the models, HUMAIN’s AI specialists are set to work alongside Microsoft’s Forward Deployed Engineers and customer teams to identify use cases, integrate AI into workflows and move projects from experimentation into production. The two companies also say they will explore deeper cooperation across productivity, devices, models and infrastructure, together with joint go-to-market work.

• Status: announced as plans. The release does not give availability dates for ALLAM in Foundry or Copilot.
• Tareq Amin, CEO of HUMAIN: “This collaboration begins with plans to bring ALLAM to a number of key Microsoft AI solutions.”
• Brad Smith, Vice Chair and President of Microsoft: “AI delivers the greatest value when it becomes part of the way people work every day.”

AMD and Cisco infrastructure is live and serving customers

AMD, Cisco and HUMAIN announced that AI infrastructure built by their joint venture is now running in Saudi Arabia and serving HUMAIN customers in the Kingdom and beyond. The systems combine AMD Instinct MI355X GPUs and AMD EPYC CPUs with Cisco Silicon One-based AI networking, including Cisco N9000 Series switches and 800G optics.

HUMAIN offers the capacity as GPU-as-a-service for model training and inference, aimed at governments, enterprises, research institutions and developers.

The next phase covers up to 250 MW built on AMD Instinct MI400 Series GPUs, with deployment planned to begin in 2027 and capacity expected to start coming online in the second half of that year. The partners say the joint venture remains on track to deploy up to 1 GW of AI infrastructure by 2030.

• Live now: MI355X-based systems on Cisco AI networking, sold as GPU-as-a-service.
• Next: up to 250 MW on MI400 Series from 2027. Long-term: up to 1 GW by 2030. Both are targets, not installed capacity.
• Dr. Lisa Su, Chair and CEO of AMD: “Bringing AMD Instinct systems online in Saudi Arabia is an important milestone in our work with HUMAIN.”
• Chuck Robbins, Chair and CEO of Cisco: “The region is moving quickly from AI investment to delivery.”

HUMAIN and DataVolt start building at Oxagon

HUMAIN and DataVolt, a Saudi developer and operator of digital infrastructure, are jointly developing 100 MW of a 360 MW AI-ready data centre. The facility is the first phase of DataVolt’s planned AI campus at Oxagon, NEOM’s industrial city on the Red Sea coast, which is part of a wider planned 1.5 GW campus.

Construction has started, and the first 100 MW is expected to be available in 2028. HUMAIN contributes its capabilities across the AI stack; DataVolt brings experience developing, financing and operating large-scale digital infrastructure. Oxagon offers pre-zoned industrial land, energy resources and subsea cable connectivity to Europe and Africa, and the facility is expected to integrate renewable energy and new cooling technologies.

• Tareq Amin, CEO of HUMAIN: “AI ambition only matters when you can turn it into infrastructure, compute and capacity that customers can use.”
• Rajit Nanda, CEO of DataVolt: “AI infrastructure at this scale demands more than capital and capacity.”

Arabic models need capacity close to home

Each announcement fills a different gap. The NEOM project and the AMD–Cisco joint venture add sovereign compute capacity inside the Kingdom, and the AMD–Cisco systems show part of that capacity is already serving customers. The Microsoft collaboration is about reach: putting an Arabic-first model where developers build and where employees already work.

For Saudi organisations, the practical effect is more choice. Teams that need Arabic language understanding could, once the integrations ship, reach ALLAM through familiar Microsoft tools rather than a separate platform, while infrastructure buyers gain more in-Kingdom options for training and inference.

Capacity and models are the start, not the outcome

Better Arabic models and more local compute make useful AI easier to build in Saudi Arabia. The value still comes from connecting them to real workflows, under the permissions, approvals and audit trails an organisation already relies on. That is the layer where most projects succeed or stall.

Sources

• Microsoft — Microsoft and HUMAIN announce long-term strategic collaboration (26 August 2026): https://news.microsoft.com/source/emea/features/microsoft-and-humain-announce-long-term-strategic-collaboration-to-enable-ai-transformation-in-saudi-arabia-and-beyond/
• AMD — AMD, Cisco and HUMAIN expand Saudi Arabia’s AI infrastructure as AMD Instinct systems go live (31 August 2026): https://newsroom.amd.com/news/amd-cisco-humain-expand-saudi-arabia-ai-infrastructure/
• Cisco — AMD, Cisco and HUMAIN expand Saudi Arabia AI infrastructure: https://newsroom.cisco.com/c/r/newsroom/en/us/a/y2026/m08/amd-cisco-and-humain-expand-saudi-arabia-ai-infrastructure.html
• NEOM — HUMAIN partners with DataVolt to develop 100 MW of a 360 MW AI data centre in NEOM (31 August 2026): https://www.neom.com/en-us/newsroom/HUMAIN-and-DataVolt-to-develop-AI-data-center-in-NEOM