In the current investment landscape of 2026, traditional economic rules seem to have been suspended. A recent report by 36Kr on the AI ecosystem in China and internationally highlights a paradoxical phenomenon: companies with zero revenue and products still in beta are achieving valuations reaching or exceeding $10 billion. The key to this phenomenon lies not in balance sheets, but in the resumes of their founders.

The Great Exodus from Big Tech

We are witnessing a massive brain drain from companies like Google, Meta, Baidu, and Tencent. Top executives and researchers, who once formed the backbone of these giants' AI departments, are choosing the uncertainty of a startup. But why are investors risking such vast amounts of capital on these individuals? The answer lies in the belief that Artificial General Intelligence (AGI) is a 'winner-takes-all' game played by a select few. Venture Capitalists (VCs) aren't just buying shares in a business; they are buying access to the expertise capable of building the next frontier Large Language Model (LLM).

In China, the so-called 'Four New AI Tigers'—Moonshot AI, Zhipu AI, MiniMax, and Baichuan—were all founded by former tech executives or distinguished academics. Despite the fact that monetization remains a distant dream, funding flows abundantly. Investors fear missing out on the next 'OpenAI' far more than they fear losing their capital in a failed venture.

Compute Cost as a Barrier to Entry

One reason for these astronomical valuations is the immense cost of entry into the industry. Training cutting-edge models requires thousands of GPUs and electricity worth hundreds of millions of dollars. When a startup raises $1 billion, the majority of that sum goes directly to Nvidia or cloud providers. This creates a unique cycle: high valuations are necessary to cover operating costs, which in turn justifies the need for even more capital.

  • Investors are betting on 'Human Capital' as the ultimate asset.
  • The US-China geopolitical rivalry is accelerating investment in domestic AI models.
  • Lack of revenue is considered secondary to achieving technological supremacy.

Is AI the New Dot-com Bubble?

The comparison to the dot-com crash of 2000 is inevitable. Back then, a company's value was measured by 'eyeballs'; today, it is measured by model parameters and researcher pedigree. However, there is a fundamental difference: AI is already producing tangible productivity gains, even if these haven't fully translated into profits for the startups developing them.

"We are not investing in a software company; we are investing in the infrastructure of the next industrial revolution," states a leading market analyst.

The unanswered question is how long this 'growth without profit' model can be sustained. As interest rates remain relatively high and pressure for results increases, these startups must soon prove they can turn intelligence into sustainable revenue. The market of 2026 is unforgiving: the 'unicorns' that fail to find a functional business model risk becoming the 'dinosaurs' of tomorrow.