AI semiconductor startup Mastiska has secured $10 million in seed funding to develop sovereign AI hardware guided by open-source foundations and auditable design practices.
The UAE-based fabless semiconductor company will focus on data-centre inference accelerators, starting with deployable FPGA cards built around proprietary IP and verified tooling.
Backed primarily by GCC sovereign wealth funds, the AI semiconductor startup plans to target buyers seeking supply-chain independence from U.S. and Chinese chip vendors.
AI semiconductor startup: What You Need to Know
Mastiska raises $10m to ship deployable FPGA inference accelerators and build a UAE-led, auditable path to sovereign AI hardware.
Partner offers for secure, scalable AI infrastructure
- Tenable Vulnerability Management – continuous exposure visibility for data centres.
- Bitdefender Business Security – layered protection for AI workloads.
- 1Password Business – secure credentials across engineering teams.
- IDrive Cloud Backup – reliable, encrypted backups for critical IP.
- Auvik – network monitoring to keep inference clusters online.
- Tresorit – end-to-end encrypted file sharing for chip design.
- Plesk – simplify secure app hosting on private clouds.
Seed funding backs a UAE-led fabless strategy
The seed round will finance data-centre-class inference accelerators and establish a UAE-based fabless semiconductor company that emphasises verifiable design. Mastiska will set up model creation in Abu Dhabi and VLSI design in India, aligning engineering with sovereign requirements.
The AI semiconductor startup intends to give government and enterprise customers audit access to chip designs to support cybersecurity assurance and standards compliance.
The company plans to serve regions prioritising autonomy and supply assurance, including the UAE, the wider GCC, Southeast Asia, BRICS economies, and the Global South.
This positioning avoids the U.S.–China export-control crossfire while offering an alternative built for transparency and predictable procurement.
Building sovereign AI hardware with auditable design
Mastiska’s strategy is to anchor sovereign AI hardware in sovereign silicon, pairing open-source tooling with proprietary IP that remains auditable for qualified buyers.
Transparent review aims to mitigate system-level risks associated with opaque accelerators and toolchains.
The AI semiconductor startup highlights growing attack surfaces in AI supply chains, citing independent research on AI cyber threat benchmarks and prompt injection risks. Its approach seeks to strengthen verifiability, reduce black-box dependencies and support rigorous assurance reviews at chip and board levels.
Regional initiatives underscore the relevance of resilient digital infrastructure, from new data centre deployments in Africa to MENA automation programmes.
Related developments include a new data centre in Namibia and a robotic fulfilment centre in Kuwait, reflecting rising demand for secure compute capacity close to users.
Product roadmap: deployable FPGA accelerators first
The first commercial product from the AI semiconductor startup will be custom FPGA accelerator cards with Mastiska IP. These are positioned for immediate deployment in data centres to run AI inference efficiently, rather than serving as proofs of concept.
The roadmap prioritises performance per watt, predictable latency and operational readiness, giving customers a near-term alternative while longer-term ASIC programmes mature.
Alongside silicon, the model team is investigating brain-inspired architectures and modified transformers to improve energy efficiency without sacrificing accuracy.
Open-source tooling will be used where it accelerates development, reproducibility and auditability across the accelerator stack.
Market focus and geopolitical positioning
Mastiska will concentrate on buyers pursuing diversified supply chains and transparent audit rights.
The AI semiconductor startup will work with public sector entities, regulated industries and cloud operators that need sovereign AI hardware aligned with local standards, data residency and security controls.
This stance mirrors broader regional digitisation and AI rollouts, including Arabic-language AI platforms and telecom upgrades.
In a related trend, Arabic AI customer experience platforms and regional cloud expansions show growing demand for controllable, verifiable compute.
Implications for AI infrastructure sovereignty
For governments and enterprises outside the U.S.–China corridor, the AI semiconductor startup model offers clearer audit trails, supplier diversification and closer collaboration on feature roadmaps.
It can also provide cost visibility and alignment with local compliance regimes while developing regional skills across modelling and VLSI.
Challenges remain. Scaling from FPGAs to custom silicon requires substantial capital, foundry access and software ecosystem traction.
The company must meet stringent performance benchmarks and navigate export controls, supply constraints and developer adoption to win production deployments at scale.
Build secure, high‑performance AI infrastructure with these tools
- Tenable Nessus – gold standard vulnerability assessment for critical systems.
- Foxit PDF Editor Pro – secure document workflows for engineering teams.
- Tresorit for Business – compliant, encrypted collaboration on sensitive designs.
- Plesk Obsidian – streamline secure app hosting on private infrastructure.
- Optery – protect executive identities with automated data broker removals.
- EasyDMARC – enforce email authentication to stop phishing at source.
Conclusion
Mastiska’s seed round advances a UAE-based path to sovereign AI hardware built on auditable, open foundations. The AI semiconductor startup will ship deployable FPGA accelerators first.
As a fabless semiconductor company serving the GCC, BRICS and Global South, it aims to balance transparency, performance and supply assurance while avoiding the U.S.–China chip race.
Execution will be judged on shipping timelines, performance-per-watt and trust guarantees. If successful, the AI semiconductor startup could help nations reduce foreign dependency and scale regional capability.
Questions Worth Answering
Who led the $10m seed round?
- GCC sovereign wealth funds provided the bulk of the capital; individual investors were not disclosed.
What will Mastiska ship first?
- Custom FPGA accelerator cards using proprietary IP for production AI inference in data centres.
How does Mastiska differ from U.S.–China chip vendors?
- It targets sovereign buyers with audit rights, open-source foundations and neutral regional positioning.
Where are engineering teams based?
- Model development is in Abu Dhabi, with VLSI design in India, collaborating across the stack.
Is the platform open source?
- Core IP is proprietary but built on open-source tools to enhance verifiability and auditability.
Which markets are prioritised?
- The UAE, GCC neighbours, Southeast Asia, BRICS economies and the Global South.
What risks could slow adoption?
- ASIC development costs, foundry access, export controls, and the need to meet strict performance targets.
About Mastiska
Mastiska is a UAE-based AI semiconductor startup founded in 2024. It operates as a fabless semiconductor company focused on verifiable, sovereign AI hardware.
The company designs data-centre-grade inference accelerators and next-generation AI chips leveraging open-source foundations to support auditability and security assurance.
Engineering teams in Abu Dhabi and India develop models and VLSI designs for customers across the GCC, South Asia, Southeast Asia, BRICS and the Global South.
About Suresh Sugumar
Suresh Sugumar founded Mastiska in 2024 to advance sovereign AI silicon built from the UAE.
He leads strategy to pair model innovation with VLSI design and auditable engineering processes.
Under his direction, the company prioritises deployable inference accelerators and trusted design for sovereign buyers.
Level up your stack: CloudTalk, KrispCall, and LearnWorlds—build, connect, and train with enterprise-grade reliability.