Alexander Kopylkov: Why the Next Wave of AI Infrastructure Is Being Built Outside Silicon Valley

(MENAFN- MeUp Ltd)

The Bay Area captured 76% of U.S. AI funding in 2025. The other 24% is quietly building the infrastructure layer that the next decade runs on.

In 2025, the world invested $202.3 billion in AI, nearly double the $114 billion invested the year before. Most of it, $122 billion to be precise, landed in the San Francisco Bay Area. Alexander Kopylkov, venture capital investor and early backer of Anthropic, Telegram, and Canva, has been paying close attention to the $43 billion that left the United States entirely. Because that is where the next wave of AI infrastructure is being built.

Europe moved fast. The EU launched InvestAI, targeting €200 billion in AI investment by 2030. In March 2025, the EU selected six new AI factories for development across Austria, Bulgaria, France, Germany, Poland, and Slovenia, each backed by EU and national funding. France alone attracted more than €5 billion in AI-specific startup funding last year, making it one of the fastest-growing AI ecosystems in the world. Europe is accelerating from a standing start, and the infrastructure is following.

The Gulf went even bigger. Abu Dhabi launched MGX, a sovereign AI investment vehicle backed by Mubadala with a target of up to $100 billion in assets. The UAE is also constructing one of the world's largest AI research and chip design centers. In Southeast Asia, late-stage AI funding surged 140% in early 2025, led by Singapore, which now hosts around 650 AI startups and has secured dedicated AI labs and regional commitments from OpenAI, Google, and NVIDIA. Brazil reserved 500 megawatts of power for a sovereign AI facility and set a target of up to $350 billion in data center investment over the next decade. Every one of these governments has a clear economic return in mind.

"In regions like the Gulf and Central Europe, you often find all three conditions aligned," says Kopylkov. "Government capital reduces early risk, industrial demand creates an immediate customer base, and the infrastructure gap means early movers capture years of structural advantage."

Kopylkov looks at three things when evaluating an AI infrastructure bet outside the U.S.: how much government capital is sharing the early risk, how deep local enterprise demand runs, and how large the gap is between the compute available and the AI actually being deployed. Where all three line up, there is an opportunity. He identified the same pattern when analyzing why enterprise AI deployments stall: the bottleneck is rarely the technology. It is the infrastructure that carries the technology to the end user.

Here is what makes this especially interesting. McKinsey projects that by 2030, more than half of all AI computing will go toward inference, not training. Training is what happens when an AI model learns from billions of data points. It requires massive centralized computing power and enormous capital, mostly concentrated in the U.S. Inference is what happens when a model answers your question, flags a defect on a factory line, or reads a medical scan. That can happen anywhere. India's Yotta Data Services already runs an AI cloud with over 16,000 NVIDIA H100 GPUs. Cassava Technologies is deploying 12,000 GPUs across five African countries. The infrastructure gap is closing fast, and the investors who move now capture the early-mover advantage.

Alexander Kopylkov's earlier analysis of why 90% of AI startups fail found one quality shared by every survivor: cost discipline built in from day one, not added on after things get expensive. The same principle applies to geography. The most durable AI infrastructure bets outside Silicon Valley solve real enterprise problems in their region. They do not try to out-spend California.

For founders and investors, Kopylkov's advice is practical: start with supply chain adjacency. Which industries near you are digitizing right now? Where is demand for AI tools outpacing the infrastructure to deliver them? That gap is the opportunity worth sizing.

Capital concentrates until the returns compress. The €200 billion mobilizing across Europe, Abu Dhabi's $100 billion MGX vehicle, and Southeast Asia's 140% late-stage surge are early signals of a redistribution that the inference shift will accelerate. The window for early movers is still open. Not for long.

MENAFN18062026008100017288ID1111275522