SMIC: right time, right place?

After one of the fastest ramp-ups in history, Chinese foundry SMIC positions its IPO valuation second only to TSMC.

Semiconductor Manufacturing International Company (SMIC) has set out with one of the richest valuation targets of a China listing in recent years, but tech specialists believe it may be achievable thanks to the $1.5 billion IPO's fortuitous market timing.

With a syndicate consensus price to book range of 2 to 2.5 times post deal or 2.5 to 3 times pre deal, SMIC will rank second only to global leader TSMC, which was trading at 3.2 times 2004 book at Tuesday's close. At these levels, it is being pitched at a significant premium to both UMC and Singapore-based Chartered Semiconductor, which are respectively trading at 1.8 times 2004 book and 1.4 times 2004 book.

At the time of its IPO in October 1999, Chartered Semi was priced at 2.2 times book.

Most of SMIC's early stage fundraisings averaged 2.5 to three times book and initially many tech bankers thought an IPO at these levels was far too ambitious for what is only a four-year old company. However, many now believe this valuation is not unfeasible in the context of an upswing in both the tech and China cycles. Together the two should play off against each other to the benefit of the deal's overall momentum.

And while some investors are beginning to worry that pure China plays are becoming increasingly overvalued, they may derive comfort from the fact that the tech sector is at a much earlier stage of an upswing. Having believed ASPs (Average Selling Price) would turn in the third quarter of 2003, for example, most industry players now think they will start rising in the second quarter of 2004.

Analysts also argue that an upturn should continue through to the end of 2005. As UBS wrote in a recent research report, "We believe the foundry industry remains in the recovery phase. We are now seeing tight capacity utilization at the leading players and overflow orders trickling down to the smaller players. We see this as a necessary condition for a rise in pricing power."

TSMC Chairman Morris Chang also recently forecast that global semiconductor revenues would rise 26% in 2004 and 10% in 2005.

At the height of previous upswings, TSMC has pierced five times book and consistently traded above four, implying significant upside from current levels.

China vaults SMIC at premium to UMC

The second and third main factors determining SMIC's valuation premium are its scarcity value and strategic location. While UMC may currently be more profitable than SMIC, it is not based in China where the latter has first mover advantage ahead of all its rivals.

According to IC Insights, China's share of the global IC market is expected to increase from 13.7% in 2003 to 23.5% by 2008. One year ago, 88.6% of these chips were imported and even though semiconductor capacity is forecast to grow by 44% per annum over the next five years, there will still be a significant gap between domestic supply and demand.

One of the most popular sayings among those working in the capital intensive tech industry is that timing is everything and SMIC appears to have got its timing exactly right. Having purchased much of its asset base during an industry downturn, it stands poised to ramp-up output just as the cycle is turning.

SMIC already achieves same margins and ASP as Chartered: future growth higher

Few are willing to dismiss the incredible achievements of Richard Chang since he founded SMIC only four years ago. The former head of TSMC-acquired WSMC has created a profitable company in record timing. The statistics are impressive on any count.

Having posted a negative gross margin of 17.9% in June 2003, SMIC had turned this round to a positive margin of 21.2% by the end of the fourth quarter. Some way ahead of it, TSMC stood at 39.3%, with UMC on 28.1% and Chartered 22%.

To do it, SMIC broke the mould of the Chinese semiconductor industry, where most players have created a niche serving companies looking for trailing edge technologies of 0.25 micron and above. SMIC, by contrast has swiftly ramped up to the most sophisticated process geometries in the hope of fatter margins.

At the end of the fourth quarter, it reported that 10.4% of its sales were wafers of 0.13 micron. TSMC was deriving 18% of its sales from 0.13 micron or below at this point, while UMC was down at a much lower 12% and Chartered 10%.

SMIC has been able to achieve this because of a strong roster of technology partners such as Elpida and Infineon. Few of its partners appear constrained by the Wassenaar agreement signed by over a dozen nations, which is supposed to limit advanced technological transfers to Communist countries.

But it is not the adoption of advanced technologies, which really bothers SMIC's rivals. It is the threat of undercutting them on price, in the process bringing down everyone's margins.

Some industry veterans say that SMIC is still signing contracts that undercut the biggest foundries by 20% to 40%. The company itself, on the other hand, reported a 90.8% increase in its ASP during 2003.

At the end of the year, SMIC says its ASP was $910 compared to $911 for Chartered, about $1,100 for UMC and about $1,300 for TSMC

The fourth quarter saw all four foundries increase their utilization rates as chip demand started to grow. By the end of the year TSMC was running at 101% capacity, SMIC 97%, UMC 96% and Chartered 71%.

This in turn helped SMIC turn its first ever quarterly profit during the fourth quarter of the year and reduce its overall net loss from $102.6 million at the end of 2002 to $66.1 million at the end of 2003.

During the fourth quarter it posted a profit of $10.2 million, compared to a net loss of $43.2 million for Chartered. Because they ship so many more wafers, UMC and TSMC were far more profitable, with UMC reporting a fourth quarter net income of $202 million and TSMC $483 million.

In terms of shipments, TSMC shipped about four million wafers during 2003, UMC about two million, Chartered 606,000 and SMIC 153,000.

Too good to be true?

Critics of the company say its IPO faces two main challenges. Firstly there remains considerable uncertainty about the sustainability of the current industry cycle.

As one tech banker comments, "If the US economy falls over, then so will the tech sector. With a deficit consistently running at 5% of GDP, there are those who wonder how real the US economic recovery really is."

Secondly, not all analysts are convinced ASPs will increase dramatically over the course of 2004. A number argue that foundries will have less pricing power during the current upswing than they have in the past.

Others counter that since IDM's continued to under-invest during the last industry downturn, they will outsource far more production to the foundries during the current upturn, thereby increasing their pricing power. IDM's account for 90% of the global semiconductor market and to date, have outsourced about 20% of production to the foundries.

And thirdly there are those who doubt SMIC itself. "The numbers seem almost too good to be true and what if they are?" says the head of technology at one investment bank. "There are those who think the cost structure is just too low to have got SMIC up to that level of technological sophistication so quickly."

During 2004, SMIC intends to increase capacity from 49,000 eight-inch wafers per month to 114,750 by the end of 2004.

To do so it will spend $1.9 billion during 2004 and a further $1.4 billion during 2005. It currently has three eight-inch fabs in Shanghai, one eight-inch fab in Tianjin (purchased from Motorola in October 2003) and three 12-inch fabs under construction in Beijing.

IPO structure

Global co-ordinator for its IPO is Credit Suisse First Boston, with Deutsche Bank as joint bookrunner. Under them, Citigroup and Nomura are co-leads in the international tranche, alongside ABN AMRO and DBS as co-managers. Nomura will separately run a POWL (public offer without listing) in Japan.

In the US tranche, Citigroup is again an underwriter alongside Bear Stearns, WR Hambrecht & Co, Needham & Co, SG Cowen, Think Equity and Thomas Weiss & Partners. In the Hong Kong IPO, Bank of China and SinoPac will complete the syndicate.

The deal has a split of about two-thirds new shares and one-third old shares. The company has a very fragmented shareholding base and bankers say it will maintain careful control over the exit of existing shareholders.

Post listing, SMIC will have a freefloat of about 25% and a market capitalization of about $6 billion. This will easily surpass Chartered's $2.6 billion, rank half of UMC's $14 billion and a long way behind TSMC's $40 billion.

Formal roadshows will begin on February 23, with the Hong Kong public offer scheduled to run from March 8 to 11. Listing is scheduled for March 17 in New York and March 18 in Hong Kong.

Share our publication on social media
Share our publication on social media