China’s decision to block Meta’s $2bn purchase of Manus is not just a failed acquisition. It changes how investors need to price AI companies. The old assumption was straightforward: the best assets would draw the highest bids, and the market would set the value. The Manus case cuts across that. In AI, a company may now be worth less than it looks if politics, regulators and national industrial priorities make it harder to buy, harder to move and harder to exit.
That is the financial point behind the block. China’s National Development and Reform Commission said it would prohibit foreign investment in Manus and ordered the parties to cancel the transaction. Manus had already shifted its headquarters and core team to Singapore after a US-led funding round, but Chinese authorities still reviewed the deal, involved multiple regulators and restricted two co-founders from leaving the country while the process was under way. The message is plain enough: changing a company’s address does not necessarily change the state’s view of what it is.
That goes straight to valuation. Buyers, founders and venture investors usually focus on product strength, strategic fit, user growth and revenue potential. Manus adds another variable that is much harder to model cleanly: political ownership risk. A company can attract a rich offer and still find that the route to closing is blocked because regulators decide the technology, talent or founder base is too important to move into foreign hands. At that point, value is no longer just what a buyer will pay. It is also what governments will allow.
That is bad news for more than Meta. Meta may now have to unwind a complex transaction, sell the asset on or settle for some more awkward structure if Beijing keeps pressing. But the broader warning is for founders and investors across AI. If a sale to a foreign buyer can no longer be treated as a clean exit, the exit maths changes quickly. Fewer realistic buyers usually mean weaker pricing, however attractive the product may be.
The bigger shift is that AI M&A is starting to look less like ordinary tech M&A and more like a strategic sector with political gates around it. Manus was attractive because it sat in one of the market’s hottest areas: AI “agents” that can operate software, manage files and carry out tasks rather than simply answer prompts. In normal market terms, that is exactly the sort of capability a large platform would want to own. Beijing has now made clear that it sees something else in the same transaction: loss of domestic technology, loss of talent and loss of control. That is not the kind of objection that disappears with a cleaner legal structure.
The buyer list gets shorter once governments start thinking that way. Local capital, politically acceptable capital and more complicated ownership structures become more useful. Direct acquisitions by large foreign platforms become less certain. That can pull valuations in two directions at once. Strategically important AI companies may look more valuable in theory because states want to hold on to them. At the same time, they can become less valuable in practice because the number of buyers who can actually complete a deal is smaller.
China is not acting in isolation here. The US has spent years treating TikTok as a national-security problem tied to ownership and control, and Reuters reported in January that TikTok reached a deal for a new majority American-owned joint venture to avoid a ban, with ByteDance retaining only a minority stake. That is a different legal route from the Manus case, but the underlying point is similar: once a technology platform is seen as strategically sensitive, ownership itself becomes political.
The UK has not gone as far as either Beijing or Washington in high-profile AI ownership fights, but it is moving in the same general direction of tighter scrutiny. Reuters reported in March that Britain was refining its investment-screening rules to keep pace with national security risks, while the government’s own response on the National Security and Investment Act made clear that advanced AI remains one of the sectors where deals can trigger closer review. The British approach is less theatrical, but it still pushes in the same direction: strategic technologies are no longer assumed to be open to any buyer who can write the cheque.
That is where the Manus episode becomes more important than the deal itself. If the US is willing to force ownership changes in the name of security, if China is willing to block a foreign acquisition outright, and if the UK is steadily tightening the screening framework around advanced technologies, then investors are looking at a market that is changing shape. AI assets are still valuable, but they are not freely tradable in the old sense. Their value now depends much more heavily on where they were built, who controls them and which governments think they matter.
For Meta and its rivals, that raises the cost of keeping up. Buying your way into the next important corner of AI may become harder just as the competitive pressure to do so gets stronger. If cross-border acquisitions in agentic AI run into political blocks, the largest groups may have to rely more on internal development, minority stakes or partnership structures that stop short of outright control. That slows some deals, complicates others and makes the whole race more expensive.
The market lesson is clear enough. Investors have spent the past year valuing AI companies as though scale, scarcity and strategic excitement would eventually justify almost any price. The Manus case is a reminder that the route from promising technology to monetisable asset is less smooth than that language suggests. Ownership, integration, founder mobility and exit can all now be interrupted by the state. That is not background noise. It goes straight to what an AI company is worth and to whom it can realistically be sold.
So China’s move is not just about Meta. It is a warning aimed at the next wave of deals. The strongest AI assets will not necessarily end up in the hands of the biggest global buyers simply because those buyers can pay the most. In this market, political permission is starting to sit much closer to the centre of valuation.
More from Finance Monthly: Anthropic Amazon Compute Deal: What the $100 Billion Commitment Says About AI Infrastructure Power
