Mastiska, a UAE-based semiconductor startup founded by chip-industry veteran Suresh Sugumar, has raised USD 10 million in seed funding to accelerate its push toward sovereign AI chip development. The round was backed largely by Gulf sovereign-wealth networks and early supporters of enterprise hardware in the region.
The company plans to build data-centre-grade AI inference accelerators for governments, enterprises and institutions across the Middle East, South Asia and the broader Global South. Mastiska’s first commercial product will be an FPGA-based accelerator card. The long-term roadmap points toward fully custom AI silicon.
The startup operates a fabless model with an AI-engineering team in Abu Dhabi and a VLSI chip-design centre in India. It also commits to offering full audit access to its hardware designs for clients that need sovereignty, security and transparency for sensitive workloads.
Why this funding round matters for the region
The raise marks a step forward in the region’s ambition to build independent AI infrastructure. With compute demand rising and global chip supply constrained, Gulf countries are exploring in-region alternatives to reduce reliance on US and Chinese hardware. Mastiska aims to position itself as a provider of “sovereign silicon”, giving governments and regulated industries hardware designed and controlled within the region.
There is also a clear technology story. Starting with FPGA-based accelerators allows early customers to run AI inference workloads efficiently without waiting for custom chips. If Mastiska succeeds in delivering its own silicon, it could offer hardware tuned for large language models, computer vision systems and national-scale AI deployments.
For the UAE, this signals a shift from importing high-end hardware to building foundational compute components locally. It also highlights ongoing efforts to grow semiconductor capability and attract deep-tech engineering talent.
Who stands to gain
Governments, sovereign agencies and regulated sectors such as defence, energy and finance may benefit from regionally controlled AI compute options. Enterprises, AI startups and research institutions across the Global South could also gain more predictable access to inference hardware that is easier to procure and integrate than foreign alternatives.
The semiconductor talent pools in Abu Dhabi and India will feel the impact as Mastiska expands its engineering teams, builds partnerships with universities and works with data-centre providers and industry partners.
What comes next for Mastiska
Mastiska plans to begin pilot deployments of its FPGA accelerator cards in 2026. In parallel, the company will push ahead with design work for its first custom AI inference chip. A significant expansion of its Abu Dhabi and India teams is expected over the next year.
Key markers to watch include early performance benchmarks, partnerships with data-centre operators or government agencies and the pace of additional sovereign funding. The company has momentum. Its challenge is turning capital into working silicon that can compete in a global market.
