Armenia's first tech megaproject loan: $300M for Firebird AI datacenter
What's the deal? Six Armenian financial institutions signed a $300 million syndicated loan to finance Firebird AIDealroom has a profile for this one. Try Dealroom →'s artificial intelligence datacenter, the high-tech ministry announced on March 27. Ardshinbank, ACBA Bank, Evocabank, Fast Bank, C-Quadrat Ampega Asset Management Armenia, and Amundi-ACBA Asset ManagementDealroom has a profile for this one. Try Dealroom → are backing the facility.
The debt covers a significant share of the project's first phase, valued at roughly $450 million. Fast Bank is contributing $30 million to the total.
What's the endgame? Firebird AI is building what the consortium describes as the region's most powerful computing and data processing center, designed to support AI workloads. A second phase would bring total investment to around $4 billion, adding 41,000 NVIDIA GB300 GPUs to reach 50,000 — enabled by US export licensing announced during Vice President JD Vance's February visit to Yerevan.
Why now? The project ties Armenia into US technology supply chains, building on AI and semiconductor agreements signed in Washington in 2025. Delegations from Armenia's High-Tech Industry ministry, the Central Bank, and the US Embassy attended the signing.
It is the first time Armenia's banking system has used a syndicated structure for a large-scale technology investment, the ministry said, and the largest corporate lending transaction in the country's history. Ardshinbank acted as initiator and lead financial institution, forming the six-member syndicate.
High-Tech Industry Minister Mkhitar Hayrapetyan said the deal reflects broader economic ambitions. "This is an investment not only in the Firebird AI company, but also in the development of Armenia's technological sector, innovation, science, and economy," he said.
What could go wrong? Fitch Ratings, which affirmed Armenia at BB- with a Positive Outlook, noted the datacenter's investment could reach 12% of GDP — an upper estimate, not committed spending. The scale concentrates risk: delays in chip deliveries, power supply, or customer contracts would register at the macro level.
The signal: The round sits near the top of AI debt financings in its class, ranking in the 95th percentile among all-time US AI debt deals. For a small economy, the involvement of both commercial banks and asset managers signals that domestic capital can be mobilised at scale — reducing reliance on overseas financing for compute-intensive infrastructure.
Read more: sevanbusiness.am
Image credit: Generated with Gemini
