ST Dongyi: How to Build an AI Empire Without Money and Working Servers

When a company's affairs are going very poorly, heavy artillery in the form of hyped technologies comes into play. The Chinese market is not an exception here, but rather a trendsetter. The firm *ST Dongyi, which once dealt with interior design and now struggles for survival in bankruptcy proceedings, suddenly decided that the AI computing market would be its salvation.

In a recent statement, management acknowledged that plans to transform into an IT giant are hanging by a thread, and investors should moderate their appetites. The essence of the story is that the company is attempting what is called an asset injection related to computing power. This is a standard move for Chinese companies with the *ST prefix, which means risk of delisting.

If the core business is mired in debt, you need to urgently find "new oil." In 2024, this oil became teraflops and graphics processors. However, Dongyi honestly (or under pressure from regulators) admits that the success of this venture depends on whether they can complete the reorganization and whether supervisory authorities will approve it.

For now, it all looks like an attempt to build a skyscraper on a swamp without a foundation. The main asset being bet on is an intelligent computing center in Heihe. It sounds solid, if you don't read between the lines.

It turns out that this center is a "gift" from the majority shareholder, which hasn't even been finished yet. The facility exists in a half-assembled state, and launching it requires huge capital injections, which the sinking company simply doesn't have. In an industry where Nvidia H100s cost like small islands and delivery times stretch for months, promises to finish the data center "sometime later" look at the very least ironic.

Even if we assume that a miracle happens and server racks in Heihe are filled with hardware, a question arises: who will compute there? The cloud computing market in China is oversaturated. Giants like Alibaba, Tencent and Huawei wage price wars, driving down the cost of computing power to a minimum.

For a small player with huge debts and questionable reputation, it will be extremely difficult to convince serious model developers to trust them with their data. The risk that the cabinets in the data center will stand empty is enormous, and this is a direct path to even greater losses. This situation highlights a broader problem in the industry — "AI-washing" or an attempt to pass off wishes as reality.

We've already seen this with blockchain and metaverses, when any company that added a fashionable word to its name got a short-term spike in stock price. But AI requires real infrastructure, extremely complex logistics and an insane amount of electricity. Simply announcing yourself as a "computing power provider" is not enough, especially when a bankruptcy procedure looms in the background and an unfinished concrete frame in the steppe.

The bottom line: will the Dongyi case be a lesson for the market, or will we see dozens more such "transformations" from builders to GPU suppliers? For now, it looks like a last attempt to save face before complete collapse.