Lessons from a Chinese AI unicorn’s failed IPO

Unisound, a Beijing based AI company focused on speech recognition, has decided to terminate its IPO process. Surprising for some, but predicted for people in the industry. 

The company was valued at CNY8 billion – with an annual turnover of CNY 200 million, that means a Price to Sales ratio (P/S) of 40. That’s the valuation before IPO – how do you expect backers and investors to make money from this IPO?

A P/S of 20 would have been much healthier, giving both the team as well as investors room to grow. 

Besides, the company was making a loss exceeding CNY 300 million a year, which means it runs a CNY 500 million annual cost structure. 

That is scary – especially we know that AI companies ARE NOT ABLE TO productise their offerings very easily. A lot of customisation is needed for each customer deployment, meaning they can’t scale up their revenues without dramatically increasing costs. 

The company ran into its current debacle because initially everyone, including the founders and investors, placed too much hope on it – resulting in a valuation that was too high to be justified by business performance, even in the current capital environment of abundant liquidity. 

What should AI companies have done, or do now? Well, instead of continuously investing in more and larger client projects, one approach might be to use their high valuation to acquire more traditional businesses serving the same market. 

You can’t expect those businesses to modernise themselves more efficiently to unlock their value – an acquisition will force them to do that. It is also easier than building up organically for AI companies.  

Just like what internet companies and D2C brands have been doing – scale, acquire, and unlock value.