BroadChain, April 27, 20:16. On April 27, the long-rumored deal between Manus and Meta finally came to a close, but the outcome was not the market's anticipated "continued review" or "supplementary materials." Instead, it was a direct ban on investment and a requirement to cancel the transaction. The weight of this decision far exceeds the commercial level; it conveys a stance. For Meta, missing out on acquiring one company is not the end of the world. The $2 billion loss for Manus is also not the first bubble burst in the AI industry. However, in the startup sector, stories of failed fundraising, collapsed M&A, and valuations dropping to zero are common, but this time it is entirely different.
In the past, many entrepreneurs were accustomed to viewing their companies as purely commercial entities: a good product drives user growth, valuation increases, and capital exits—this is the natural path. But AI is different from the internet businesses of the previous generation. AI is not a new application category, nor is it a smarter office tool; it is becoming the key to winning the next round of competition. Whoever masters model capabilities controls the next generation of software entry points; whoever possesses intelligent agent products may dominate the next generation of workflows; whoever builds AI infrastructure and application ecosystems gains an advantage in the future division of industrial labor. This is not just a setback for a particular company, but a collapse of the old world model.
PART.01 The Old Model Fails: From Capital Game to Strategic Asset. Over the past decade or more, a mature script has formed in the minds of Chinese entrepreneurs: people in China, market in China, engineers in China, products developed in China, but financing in US dollars, entities in the Cayman Islands, listing in the US, and offices relocated to Hong Kong, Singapore, or Silicon Valley when necessary. This system has operated for years, with the underlying premise being: China needs growth, the US needs assets, capital needs exits, and entrepreneurs need stories. As long as all parties cooperate in the gray zone, allowing the company to grow, investors to exit, and founders to escape difficulties, it is considered a success. The core contradiction of that era was not "who owns the technology," but "how to make the company grow, capital exit, and sustain development." As long as this logic held, many gray areas could be tolerated.
But AI has now been placed within a competitive framework. Model capabilities, engineering talent, training data, inference systems, intelligent agent products, and commercialization entry points—each can be regarded as a strategic asset. If one still uses the worldview of internet companies from the 2010s to handle AI companies in 2026, they will inevitably hit a wall. The problem for many is not a lack of judgment, but that their world model has not been updated. They think it is a capital game, but the table has changed. In the past, the main counterparts were investors, users, exchanges, and M&A lawyers; now, they must also deal with security reviews, export controls, technology boundaries, and geopolitical competition. This is not just a change in China; the US is also transforming. In the past, global capital believed in efficiency: capital flows to cheaper talent, companies move to larger markets, and projects list at higher valuations. The underlying logic of globalization was resource allocation efficiency. But today, the underlying logic of globalization is becoming boundaries: technology has boundaries, data has boundaries, computing power has boundaries, capital has boundaries, and talent mobility is also starting to have boundaries.
PART.02 Location Determines Everything: The Survival Rule in the AI Era. Many in the business world like to talk about capabilities: product capability, financing capability, growth capability, organizational capability, technical capability, and narrative capability. But in the AI era, what is more critical than capability is location. Which country you are in, which ecosystem, which link in the industry chain determines your ceiling. The Manus case shows that even with excellent products and a strong team, if the location is wrong, it can be crushed by geopolitical and regulatory forces. What entrepreneurs fear most is not a lack of effort—they all know how hard it is to work—but giving their all to run forward, only to find halfway that the map has changed.