On 15 June 2026, at the Türkiye Artificial Intelligence Summit in Istanbul, President Recep Tayyip Erdoğan and Minister of Industry and Technology Mehmet Fatih Kacır announced Türkiye’s Artificial Intelligence Action Plan, covering 2026-2030. The plan is built around four pillars: Recognise (AI literacy), Benefit (data access and public-sector use), Produce (domestic models and entrepreneurship), and Manage (regulatory framework).
Most coverage led with the headline numbers: 10 billion dollars in private-sector-led infrastructure investment, AI literacy training for 5 million citizens within two years, 1 gigawatt of data centre capacity by 2030. For a manufacturing SME, the real question is different: which door does this plan actually open, and what needs to be in place before walking through it?
The parts that touch an SME directly
Two mechanisms in the published material reach manufacturing SMEs directly:
- AI vouchers — a support mechanism intended to make AI tools more accessible for SMEs. Health, energy, and smart manufacturing are named among the priority areas.
- National Data Library — a plan to open at least 2,000 public datasets from critical areas such as health, agriculture, defence, and e-commerce to researchers and entrepreneurs.
Alongside these, an AI Growth Fund and a National AI Research Fund are mentioned; both lean toward startups and R&D, though a more active ecosystem tends to benefit SMEs indirectly too.
Worth stating plainly: as of this writing, the voucher programme’s eligibility rules, funding caps, and application calendar have not been published. The plan is a framework document — implementation details will likely follow through communiqués and calls from the Ministry of Industry and Technology and KOSGEB. Once applications open, check the current terms directly on sanayi.gov.tr and kosgeb.gov.tr.
The plan in numbers
| Target | Figure |
|---|---|
| AI literacy workshops | 81 provinces |
| Citizens trained within two years | 5 million |
| Advanced AI specialists trained | 10,000 |
| Applied AI professionals trained | 100,000 |
| Public datasets to be opened | 2,000 |
| Data centre capacity by 2030 | 1 gigawatt |
| Private-sector-led infrastructure investment | $10 billion |
| Share of public investment programme allocated to AI | 2% |
The table gives a sense of scale. For a manufacturing shop floor, what matters more is a narrower question: which data is being collected on your own line, and where does it go.
What the smart-manufacturing priority actually means
Naming “smart manufacturing” as a priority area is not accidental. Demand forecasting, inventory optimisation, predictive maintenance, and anomaly detection are use cases already in demand across Türkiye’s manufacturing base. AI-assisted ERP and production-planning tools are becoming more common in this space; both local and international ERP vendors have started adding demand-forecasting and maintenance-prediction modules to standard packages.
But for a model to learn and produce a useful forecast, it first needs regular, clean data at sufficient volume. Maintenance logs kept on paper, production counts that come from an operator’s memory rather than a sensor, inventory data scattered across separate spreadsheets — building an AI model on top of that is like adding a floor to a building with no foundation.
Before applying for a voucher: the data-infrastructure requirement
KOSGEB’s existing SME Digital Transformation Support Programme runs on a similar logic: TÜBİTAK TÜSSİDE’s DDX digital-maturity assessment measures where a business stands and which investment should be prioritised. It’s reasonable to expect the AI voucher programme to apply a comparable maturity criterion — though without a published communiqué, that’s an expectation based on past programme design, not a confirmed requirement.
In practice, preparation moves in three steps:
- Measure — Use machine-to-machine (M2M) and IoT connectivity to turn production, energy, and maintenance data into a regular stream. If sensors aren’t in place yet, decide first what data to collect, at what frequency, and where it will be stored.
- Transform — Move scattered spreadsheets into a single source of truth; use an ERP and process automation (RPA) to make the data flow systematic. An AI model can’t produce meaningful output without a reliable data pipeline underneath it.
- Sustain — Turn the collected data into reports usable both for internal decisions and for support-programme applications.
A business that has already completed these three steps can put together an application in days once a voucher or support programme opens. A business without a data foundation ends up spending more time on pre-programme preparation than on the programme itself.
What the National Data Library means in practice
The datasets the National Data Library plans to open — weighted toward health, agriculture, defence, and e-commerce — may not be directly usable for most manufacturing SMEs; that list leans closer to the research and startup ecosystem. There’s an indirect effect worth noting, though: as public data-sharing culture and standards mature, sector-specific data-sharing models (through supply chains or industry associations) could develop over time as well. The concrete step available today is improving the quality of your own operational data, rather than waiting for an external data ecosystem to arrive.
İkiz Eksen’s approach
İkiz Eksen doesn’t expect manufacturing SMEs to build this foundation alone. We run the measurement-to-software-to-compliance chain end to end: field data collection (M2M/IoT), ERP and process automation as part of digital transformation, and the AI-assisted reporting layer that sits on top of that data once it’s structured. Through the Qera track record, we support more than 550 organisations across 15+ sectors, with 100+ ERP integrations, running on Microsoft Azure infrastructure. The work is carried out across Türkiye.
When an AI voucher or a comparable support programme becomes relevant to your business, we can assess together whether your data infrastructure is ready. Take a look at our solutions or get in touch directly.
Frequently Asked Questions
When will the AI voucher programme open for applications?
As of this writing (August 2026), the voucher programme’s application calendar and eligibility terms haven’t been officially published. The plan is a framework document; follow announcements on sanayi.gov.tr and kosgeb.gov.tr for implementation details.
Can a small manufacturing workshop actually benefit from this plan, or is it aimed at large companies?
The plan’s text explicitly names SMEs as a target group, with smart manufacturing among the priority areas. As with any support programme, though, eligibility rules (NACE code, scale, digital maturity) will be decisive — the manufacturing (NACE C) requirement seen in past KOSGEB programmes is a reasonable precedent.
Do I need an ERP system before I can apply for an AI voucher?
There’s no published communiqué making this a formal requirement. In practice, though, an AI model needs regular, digitised, accurate data to produce anything useful. An ERP isn’t mandatory, but a discipline of data collection and storage is.
Does this plan replace the KOSGEB Digital Transformation Support Programme?
No, they’re separate programmes. KOSGEB’s existing digital transformation support continues; the AI action plan appears to add a new layer on top of it — vouchers, the data library, fund mechanisms. The exact relationship between the two will become clearer as implementation communiqués are published.
Does opening my data to an AI programme create KVKK (data protection) risk?
The public data library programme doesn’t require your business to share its own data — the published datasets are public-sector data aimed at researchers and entrepreneurs. When you do share your own operational data with a third party (a consultant, a software vendor, a support institution), KVKK’s data-processing agreement and purpose-limitation principles always apply.