Abu Dhabi intends to add Chartered Semiconductor to Global Foundries

The government of Abu Dhabi, acting through the vehicle of its Advanced Technology Investment Company (ATIC) added a second significant foundry operation to its portfolio today with the announcement of its intent to purchase Chartered Semiconductor of Singapore for approximately US$ 3.9 billion. The offer represents only a 14.2 percent premium to the 30-day volume-weighted average trading price of Chartered on the Singapore exchange, according to the companies. The two companies have signed a definitive agreement, and so only await shareholder and Singaporean High Court approval. But since Temasek Holdings, the holding company which owns and manages direct investments by the government of Singapore, has already committed to vote its roughly 62 percent of shares for the deal, it appears that the odds are in favor of the transaction. Issues, should they arise, would likely come from other places, such as the European Union.

The acquisition would add Chartered to ATIC’s earlier purchase of the Dresden R/D and production fab and the under-construction New York fab of AMD to form Global Foundries. The companies have said that Doug Grose, CEO of Global Foundries, would head the new combined operation, with current Chartered CEO Chia Song Hwee taking the position of COO in the new operation. The intent is to integrate the two foundries into a single global operation, the companies said.

On paper, integrating the two foundries should be an exercise in complementarity. Global Foundries has inherited the formidable but probably under-funded process development capability at Dresden, plus the futures play of the incomplete fab in New York. It also has access to the family of processes, including 45 nm, based on IBM’s Silicon-on-Insulator (SoI) technology. But from its history as essentially a captive fab for AMD, Global Foundries has little experience in the very different business of being a foundry.

Chartered, in contrast, has stayed primarily on a bulk-silicon process and the IBM-driven Common Platform Alliance roadmap. That effort has Chartered announcing existence, if not availability, of a 32 nm high-k/metal-gate process and plans for 28 nm. The company has struggled financially and lagged giant TSMC in market share and R/D budget, but has none the less deep experience and significant market share in the pure-play foundry business.

But in practice, putting the two pieces together could prove a serious challenge for Chia, who will have responsibility for the integration. About the only things the two organizations have in common are close R/D cooperation with IBM and, now, common ownership. Beyond that, they are two organizations with very divergent cultures. Global Foundries already combines the unique culture of the Dresden fab, tied to the city that birthed and supports it, and the very different culture of the semiconductor complex growing in the shadow of IBM in up-state New York. Add to this the inimitable AMD internal culture, and Global Foundries is already, shall we say, cosmopolitan. But Chartered, which some senior semiconductor executives have described as uniquely Singaporean, will bring an entirely new way of approaching operations, human resources, and financial management to the organization.

In this mix, the common relationship with IBM won’t be much help. Ultimate responsibility for bringing the cultures together is likely to fall on Grose, and on the new owners at ATIC. But Grose himself only joined AMD in 2007, coming from 25 years at IBM, most recently managing technology development and manufacturing. So Grose represents yet another culture, the legendarily hermetic one of IBM. And his management at ATIC represents still another culture, the consistently hands-off, financially-oriented world of Gulf sovereign wealth funds.

These cultures will find simply communicating to be an interesting exercise. But they may also find that there are unresolved conflicts in fundamental objectives. Global Foundries, for example, will in the end succeed or fail not on the bulk-CMOS foundry business, but on its success in serving AMD with advanced high-volume SoI processes, in AMD’s treadmill effort to keep up with Intel. Chartered, on the other hand, must live in the far more complex world of a rich mix of processes and an even richer mix of small- to large-sized lots of a pure-play foundry. Global Foundries teams have in the past had the freedom to compromise accessibility and usability in order to achieve price and performance on just a few products. Chartered must make processes as accessible as possible across a wide range of applicaitons, and often performance and die size are secondary issues to design ease and negotiated price.

The role ATIC can play in blending these objectives remains to be seen. In the past when governments have invested directly in foundries, it has been to create a foundry business within their borders. This was the history of Chartered. But ATIC appears to be investing as a purely financial move, hoping to achieve cash return on investment or a successful exit for owners—Abu Dhabi—unlikely to every see a foundry on their shores. A purely financial investor may turn out to have very different goals from a country attempting to buy into the semiconductor industry. But only time will tell.