China memory giant CXMT’s US$10 billion IPO is a flop

WHAT makes an initial public offering great? One might say SpaceX’s recent IPO was a success. Global investors were fervently chasing its shares during book-building, but the actual debut was orderly.

Trading was smooth and the stock finished 20 per cent higher on the first day, giving speculators a nice little bonus to go home with. Elon Musk, meanwhile, was handed with US$86 billion to fund his grand ambition, which ranges from Mars exploration to orbiting artificial intelligence data centres in space.

By this standard, the listing of China memory chipmaking giant CXMT is a failure.

Its shares jumped 466 per cent on its debut trading on Monday (Jul 27), after raising as much as US$9.8 billion on Shanghai’s Star board, which only hosts companies that align with “national strategies” and demonstrate breakthrough technologies. CXMT’s first-day pop is too extreme.

Analysts have broadly agreed that the offering is underpriced. The IPO values the company at 2.4 times book, a lot less than its global peers and Star-listed foundries Semiconductor Manufacturing International Corporation and Hua Hong Grace Semiconductor.

In addition, CXMT is a strategic asset. It is China’s only viable contender to break into the global dynamic random-access memory (Dram) market, dominated by Samsung Electronics, SK Hynix and Micron Technology. It surely deserves some scarcity premium.

This raises the question of why CXMT is leaving billions of dollars on the table. In theory, a listing on the relatively new Star exchange means that its IPO price is not set by regulatory restrictions, which often end up undervaluing hot tech firms, but through a market-friendly book-building mechanism.

Underwriters and company management agree on the final price after comprehensive inquiries with institutional investors. So what was lost in this process?

Money on the table

A lot of politics are at play. Half of the offering was reserved for strategic investors, which included the central government’s various social security funds as well as state-owned insurance giants such as China Post Life Insurance.

With lock-up periods lasting at least a year, these conservative but powerful investment entities naturally demand that they will be able to profit by the time they can sell their shares. This means the offer price has to be low enough.

On the other side of the ledger, CXMT does not have a powerful backer to work out a good selling price. The chipmaker does not even have a controlling shareholder.

Before the IPO, the city government of Hefei, where CXMT is headquartered, held a stake of more than 36 per cent, but via multiple platforms to “reduce administrative intervention”, indicated data compiled by research outlet Gavekal.

The China Integrated Circuit Industry Investment Fund, a state-owned venture capital vehicle, is the second-largest shareholder, followed by the provincial government of Anhui. All of them are passive investors that do not want to stick their necks out.

Seen in this light, it is no surprise that CXMT’s offering is so detached from market sentiment. The Star board may have moved away from a bureaucratic registration-based IPO system, but it is still not market-friendly.

Unfortunately, this IPO leaves CXMT handicapped right out of the gate. Memory chipmaking is a capital-intensive business. The US$10 billion IPO may seem like a lot, but pales next to rival SK Hynix’s US$26.5 billion US listing earlier this month.

SpaceX and CXMT make an interesting comparison. It is clear that the US-China AI arms race has moved on to financing, and that access to vast amounts of capital has become a strategic advantage.

In this regard, China is well behind, even though its stock market is big and liquid enough. The process is still too political – and that means its national champions cannot fundraise at fair prices. BLOOMBERG