The announcement of roughly USD 3 billion in public funding intended to catalyse around USD 10 billion of private investment in data centres and AI is a significant policy signal—not an automatic grant for every project. Investors should separate infrastructure-heavy data-centre economics from AI-software economics before valuing an opportunity.
Data centres: power before narrative
Test available MW, grid connection timing and redundancy, energy price, PUE, cooling/water, fibre, land, permits, CAPEX per MW, build schedule, financing and customer contracts/pre-leases. A cheap site without credible power and connectivity is not a viable data-centre thesis.
AI companies: recurring revenue and compute economics
Measure ARR/MRR, retention and usage, GPU/cloud/model/API cost, gross margin after compute, CAC, sales cycle, data/IP rights and dependence on one model/provider. Rapid user growth can still destroy cash if inference and acquisition economics are weak.
Do not put public support into the base case before eligibility
Treat the announcement as policy direction. Cash grants, incentives or exemptions belong in the model only after an open programme, eligible entity/activity/location, qualifying expenditure, obligations and official approval are verified.
Two different diligence packs
For infrastructure, require grid/power studies, design/capacity evidence, permits, customer contracts and CAPEX schedule. For AI software, require cohorts, recurring revenue, compute cost, security/privacy and IP. Sharing a headline theme does not make the diligence identical.
Use the research in a decision
Investor Tools · Feasibility Studies · Investment Marketplace · Incentive Matcher
Official source and review
Editorial review: 11 September 2026. Verify transaction-specific legal, tax, licensing and incentive conditions at the date of execution.
Istanbul Real Estate Market Guides