Artificial intelligence has ceased to be a laboratory experiment and has become the foundation of a new economy. If just a few years ago AI investments were associated mostly with private companies, today governments have actively joined the race. The USA, the EU, China, Japan, and the United Arab Emirates are competing to build the most powerful data centers, purchase thousands of GPUs, develop their own models, and even form state-level alliances with corporations. The amount of investment is no longer counted in billions — but in trillions of dollars.
Anthropic — one of the leading AI developers and a rival to OpenAI — is launching a large-scale infrastructure project worth $50 billion. The plan includes building specialized data centers in Texas and New York. The partner is GPU provider Fluidstack. The project will create 800 permanent jobs and more than 2,000 temporary construction positions, with the first facilities going live in 2026. This is only part of Anthropic’s strategic plan, aligned with a massive U.S. government initiative involving $1.4 trillion worth of projects. Why have governments accelerated investments so dramatically?
AI is becoming critical infrastructure — like gas, electricity, and telecom
Ten years ago, artificial intelligence was simply a tool for automation or “smart features” in smartphones. Today it is the foundation of security, defense, the economy, healthcare, energy, and scientific development. What once required thousands of analysts is now handled by a model capable of processing billions of data points in seconds. Governments realized: whoever controls computing power — controls innovation, and therefore the future.
It is important to understand the meaning of “AI infrastructure”: it is not just software or algorithms. It includes massive data centers, thousands of high-performance GPUs, specialized cooling systems, power substations, fiber-optic networks, and data delivery systems. Without this, it is impossible to train large language models, computer vision systems, or autonomous robots.
A new tech race: not between companies — but states
Competition between OpenAI and Anthropic is just the tip of the iceberg. In reality, the rivalry is between the USA, China, and the EU. The U.S. is issuing grants and government loans to stimulate data-center construction domestically. China is investing in national AI clusters with its own GPU architectures. The European Union is building digital sovereignty — a program aimed at developing its own cloud platforms and local models to reduce dependency on American corporations.
The reason is obvious: if computing power is located outside the country, strategic autonomy is at risk. This is not only about economic concerns but also national security. Governments cannot afford to have their core AI systems operate on foreign servers.
AI as a source of new economic growth
According to PwC, artificial intelligence will generate up to $15.7 trillion of global GDP by 2030 — more than the combined GDP of the EU. And governments want their share not through finished products, but by controlling the fundamental infrastructure.
Investing in data centers and GPUs creates thousands of high-tech jobs, stimulates local manufacturing, energy development, and network expansion. This forms an ecosystem that gives a country long-term strategic advantage. In this sense, the Anthropic project is not just a business plan — it’s an expression of the U.S.’s new economic policy.
GPUs and data centers: why are they so expensive?
The main costs aren’t just processors. To host thousands of GPUs, one must build a facility with multi-megawatt power lines, uninterrupted power systems, cooling, and full redundancy. A single modern AI data center may consume more electricity than a city of 30–50 thousand people.
The GPUs themselves are extremely expensive: one Nvidia H100 card costs around $30,000–$40,000 — and there are hundreds of thousands of them in industrial clusters. That’s why $50 billion is not an exaggeration but the real price of joining the “AI-state club.”
Robots, autonomous systems, and the “brain” of the economy
Alongside infrastructure, the demand for physical AI applications is rising. Humanoid robots, logistics systems, smart factories — all require models capable of operating in real time, processing streams from thousands of sensors. One can imagine an AI cluster as the “brain of the country,” and the robotic industry as its “hands.” Without powerful data centers, neither autonomous transport nor robotized manufacturing can exist.
Why this matters to everyone — not just governments
Artificial intelligence already influences energy pricing, optimizes logistics in supermarkets, and helps doctors with diagnoses. In the next 5–7 years, access to computational resources will become as critical as access to electricity or the internet. The key question for every nation: will it be able to provide its business sector, science, healthcare, and defense with its own computational infrastructure?
Ukraine and its path toward AI infrastructure
Ukraine is also moving in this direction. The first government initiatives for data-center development are emerging, and businesses are gradually adopting machine learning and generative AI models. This requires reliable hosting platforms, energy independence, and access to fast data channels.
What this means for Ukrainian businesses
While global powers invest trillions in international AI clusters, Ukrainian companies must ensure they have a “solid rear” — data centers capable of keeping their services online without interruption. This is especially important under constant energy risks, cyberattacks, and increasing demand for computing.
That’s why many Ukrainian projects host their infrastructure at Server.UA — a data center with backup power, diesel generators, autonomous cooling systems, and high-speed connectivity. It already provides the foundation that the rest of the world is building for tens of billions: stable infrastructure ready for the era of artificial intelligence.
Your company may not be launching its own AI model yet — but if you work with big data, logistics, e-commerce, or fintech, you already need next-generation infrastructure. And it’s better to secure it now than when it becomes a matter of survival.
Leave a Reply