Bulgaria is an intriguing model to study, as it is a country that recognized it would not be France or Germany, nor Turkey or Brazil or South Korea in traditional defense. However, it sees itself as having the potential to exist as a regional niche in the fields of artificial intelligence and computer science. It chose to use artificial intelligence as an entry point for the defense industries, which I consider a smart and unconventional model. It differs from the idea of copying a successful model and transferring it as is, as this idea often fails.
Let’s contemplate Bulgaria first to understand what we are talking about. Bulgaria is a country with an estimated population of about 6.4 million people, and it is a member of the European Union and NATO. The ratio of R&D expenditure from GDP is only 0.8%, which is considered low compared to Europe’s ambitions. Nonetheless, its global ranking in the Global Innovation Index for 2024 stands at 38th place, which is an advanced position considering its resources. I do not wish to cite examples from other countries so that the topic is not interpreted personally, but this position is considered very good for Bulgaria and is higher than many countries that have greater resources and capabilities than it does.
Historically, Bulgaria is known for its strength in mathematics and computer science and for low-cost, high-quality software (and it even surpassed Russia at times). It also has one of the highest ratios of women working in information technology within the European Union. This is part of the Bulgarian narrative regarding their talent base. I don’t know if this information will be useful, but talking this way reinforces the general European perception.
In terms of the defense industry, Bulgaria has a medium-sized industrial base. Its traditional activities focus on producing small arms and ammunition, as well as some armored vehicles (similar to Egypt). Its defense budget is small compared to France, Britain, Germany, Italy, or even Turkey.
However, it is essential to focus on this specific point, as what I will mention now is the result of almost five years only, and we will see how it achieved a significant leap in such a short period.
In December 2020, Bulgaria laid out a strategy announcing the adoption of “Artificial Intelligence Development Vision 2030.” This strategy was issued by the Bulgarian Academy of Sciences and the Ministry of Transport and Information Technology and Communications.
Now let us count the successful projects that have emerged from this vision, and we will talk about them one by one in a simplified and rapid manner:
INSAIT
GATE
DIANA-NATO Test Center
BgGPT
EU AI Factory
This strategy is based on several key pillars, and instead of delving into its precise technical details, it can be said that its essence revolves around education, training, scientific research, and supporting innovation and startups.
If we talk about the most important institute in this initiative, it is the INSAIT Institute. Let us reflect on how an institute established only three years ago has become competitive with prestigious American scientific institutions.
The institute was established in 2021, and it was officially opened in 2022 within Sofia University. It has forged partnerships with ETH Zurich and EPFL, which are among the strongest universities in Europe in the fields of computer science and artificial intelligence. It enjoys scientific supervision from experts connected to ETH and EPFL, IST Austria, MIT, UC Berkeley, Yale, and Princeton.
The institute receives government funding of about 85 million euros over ten years, in addition to private contributions from major companies like Amazon Web Services, Google, DeepMind, SiteGround, and VMware, amounting to 15 million dollars.
The institute offers scholarships for doctoral students worth 36,000 euros annually, which is an excellent figure compared to the cost of living in Sofia, making it a real competitor to countries like Germany and France. These incentives have attracted distinguished talents, including Professor Luc Van Gool, one of the most cited researchers in the fields of computer vision (CV) and machine learning (ML) globally.
Seven research papers from the institute have been accepted at the CVPR 2025 conference, an achievement that places it at the forefront of Central European institutions, ahead of many Middle Eastern countries, and approaching prestigious institutions like Caltech and King’s College. All this for an institute that is not more than three years old.
In 2022, the DIANA NATO Program selected the institute to be a center for big data and artificial intelligence testing. This is an unconventional entry for a small country in the defense industry, through a different model based on a single specialized technology that was developed at a remarkable speed.
They developed the BgGPT program, a massive local language model, and then followed it up with a developed version called BgGPT-Gemma-2, specifically trained on Bulgarian data. It can be described as a "national ChatGPT," used in the tax authority, civil registry, police, and other sensitive sectors of a security nature.
Currently, Bulgaria has about 90 companies working in artificial intelligence with market-ready products. These companies have received funding of around 52.7 million euros in 2024 alone, while the funding value in 2023 reached approximately 101 million euros.
Thus, Bulgaria, the small resource-limited country, now holds 8.4% of the AI companies operating in fintech (FinTech) in the Cental Asia and Eastern Europe region.
It has also applied its expertise in artificial intelligence across multiple fields, including mail, shipping, legal services, civil registry, automated agriculture with robots, as well as police and investigation services, and other vital areas.
It is important to note that AI applications in Bulgarian companies are still below the European average; however, there is rapid and noticeable development in this field.
The success of this experience stems from a smart choice of the right location within the global technology chain. Bulgaria did not say, "We will create everything from computer chips to weapons," but rather decided to focus on a specific field and strive to become distinguished in it to the extent that large countries themselves would find themselves in need of it.
This logic is the same that other small countries have adopted, allowing them to excel in precise fields; like Estonia in digital government, Ireland in software, and Taiwan in semiconductors.
The essence of the idea is that you are not required to compete in all fields, but instead create a clear added value that makes you an indispensable element within any larger industrial or defense system.
Bulgaria has chosen to rely on its local universities, its European partnerships, and its low cost of living, which has enabled it to build a scientific system that produces real, advanced, and needed results.
In short, Bulgaria has done two things. First, it did not adopt a ready-made successful model to implement, because this idea is flawed. Many studies confirm that "localization" is an essential element for success, and countries must build their own model that fits their circumstances, culture, industrial and knowledge base. The most notable example of that is the Gulf countries' attempts over many years to create a "Gulf Silicon Valley," efforts that have not succeeded until today.
The second aspect is its recognition of the importance of investing in education and scientific research, which has allowed it to develop its capabilities in line with its possibilities, without building unrealistic dreams that are hard to achieve.
If we want to draw lessons from this experience, they are clear and straightforward: Small countries are not obliged to enter the race of "who has the strongest tank or the largest aircraft factory," but they can enter through the door of knowledge and precise technology, building with a clear vision, intelligently, and at the lowest possible cost.
The right question is not "How do we become like such and such a country?", but rather "In which field can we excel according to our capabilities and circumstances?" Based on the answer to this question, all educational and scientific policies should be built, for success does not always require massive resources, but requires a correct choice and a genuine willingness to invest in minds before factories.
This is an AI-generated English translation. The original text is in العربية
PhD Researcher in Defense Contracting and Cooperation Policies
