Kazakhstan needs its own technologies. Does it have any chance of getting them now?

Kazakhstan is the biggest consumer of new technologies in Central Asia. We can say this for sure because we see more and more new startups in the country that rely on new business models based on innovations. However, is Kazakhstan ready to develop technologies and not just consume them?

It’s not a good strategy to put money in the sphere of new technologies without creating solid scientific infrastructure locally. Such a strategy would lead nowhere but to a dead end. For example, the cost of just one round of machine learning has increased from $0.01 in the 2000s to $10,000 now. These services will cost about $500,000 by 2030, according to analysts. This is why only big companies or research institutions can work with these technologies. However, those companies that invested in their own processors, like TPU Google, Graviton AWS or Yitian Alibaba, can spend less on machine learning.

There are no businesses in Kazakhstan that have any experience developing AI, although local companies might have used such systems developed by third-party companies. That is why it is worth taking a look at research institutions that already have achieved some progress in this field. For instance,  Kazakh Speech Corpus 2 (KSC2) is the first industrial-scale open-source Kazakh speech corpus developed by Nazarbayev University, similar to several other projects by the Kazakh-British Technical University.

According to a survey conducted by the Ministry of Education and Science last year, about half of postdoctoral students stop publishing scientific articles and start working either as IT specialists or teachers right after graduation. At the same time, the vast majority of postgraduate students combine their education with a full-time job.

The scientific basis is necessary because any business project might be ruined by changes in browsing technology (ChatGPT is a vivid example of this trend). In this case, it would need to switch to another business model. In order to avoid such risks, big companies actively invest in DeepTech projects. Google Ventures, for example, is prone to balancing between DeepTech and quick profit projects.

These might be Enterprise- and Consumer-projects based on AI/ML/E-commerce/Fintech or more advanced startups linked to innovative knowledge in natural science, psychology, etc. For instance, such projects are involved in the production of crops with minimal water consumption, the creation of protein out of the air and new aromas with the help of AI.

When I talk about the necessity of the creation of infrastructure, I am not talking about just buildings and labs. Any infrastructure needs people who are going to maintain it. For example, the AI lab of the Mohamed bin Zayed University of Artificial Intelligence has employed world-class data engineers who maintain the smooth work of servers and software that local researchers rely on. Some estimates show that ChatGPT server operation costs about $100,000 a day or $3 million a month (excluding the salary of employees).

Scientific infrastructure is necessary for the sustainable development of Kazakhstan. Perhaps, it doesn’t make sense to switch from technology consumption to the creation of the AI industry, but it does make sense to develop new promising directions in biotechnologies and the sphere of energy. Of course, it would be naïve to wait for rapid results. This process would take decades before any tangible results appear.