2010년 1월 18일 월요일

Analysis: Intel not likely to buy major FPGA vendor

Embedded Systems

Analysis: Intel not likely to buy major FPGA vendor

Posted:18 Jan 2010

Intel Corp. wants to move beyond its microprocessor base to the digital and Internet connectivity markets but contrary to expectations, the chip giant's transition into new IC segments will not involve huge and disruptive acquisitions.

However, if Intel moves towards acquiring a major programmable logic IC company—Altera Corp. or Xilinx Inc.—as speculated by an industry analyst recently, it would be inflicting a grievous wound not only on current core computing and server microprocessor business but also killing off any future chances it may have to make a significant impact on the embedded processor market.

Fortunately, even though Intel's top executives preach paranoia they are not suicidal. Acquiring Altera or Xilinx, for which a potential buyer would have to pay a hefty premium—EE Times estimates in the range of $8 billion to $10 billion—would be a financially crippling move but also an operational and integration nightmare.

The hefty financial bill is one critical reason for ruling out the idea of Intel making a move for either of the two leading FPGA vendors. There are other reasons but first, let's explore the financial implications of such a move.

Many observers focus on Intel's huge cash hoard when weighing possible acquisition actions by the company obscuring other critical issues of concern to the board of directors.

Strictly on valuations alone, Intel's board of directors would be crazy to approve a deal to purchase Altera or Xilinx. True, with almost $13 billion in cash and short-term securities—not to mention slightly more than $9 billion in longer-term investments and other long-term assets—as at the end of the September quarter, Intel certainly has the means to fund a major acquisition on its own.

If necessary, the company's ultrahigh investment rating could come in useful if it needs to borrow from the financial market or raise additional funds by selling shares.

However, what will Intel get for the nearly $8 billion to $10 billion it would have to spend to gain control of Altera or Xilinx and would the benefits of such a union outweigh the challenges of integrating either of these enterprises into its operations? Furthermore, the entire speculation about Intel acquiring a company of the size of Altera or Xilinx ignores the history of the world's biggest semiconductor company.

Intel has made significant acquisitions in the past, the latest being its purchase in June 2009 of Wind River Systems Inc. for $884 million, as part of plans to diversify operations.

This token approximately $1 billion acquisition by a company of Intel's heft cannot be likened to a multi-billion dollar deal. Even though Intel spent a lot more over the last decade in efforts to broaden operations, the company till date, is not known for large and radical moves.

Even today, despite any perceived desperation that might prompt or necessitate the acquisition of companies the size of Altera or Xilinx, Intel's conservative management style does not favor such a move. While Intel will undoubtedly stir the acquisition waters in the near future, it won't be towards the direction of Altera or Xilinx.

Altera, Xilinx: Buyout targets?
It is also being speculated that private equity firms or other investment entities, including chip vendors, could dive in deep to push for consolidation in the industry by making a play for companies the size of Altera or Xilinx.

Even this is an unlikely development in today's market. We should expect some surprising developments and other not-so-shocking moves, including additional investments by Abu Dhabi government-controlled entities as they try to consolidate their actions in the foundry and semiconductor market.

Nevertheless, we are unlikely to see anything similar to the deals that took companies like Freescale Semiconductor and NXP private. Those deals are still being worked through by buyers and it is generally acknowledged in the industry that the conditions under which Freescale and NXP were taken private by equity investors are radically different than today's economic and financial environment.

Still, any leverage buyout companies—or even Intel should the board of directors in a moment of delusion decide to go for either FPGA companies—will have to pay $8 billion to $10 billion for the entities depending on the premium demanded by stockholders and management.

Shareholders in Altera and Xilinx would probably expect a premium of 30 percent to 40 percent to hand over control of the companies to interested investors. Based on the closing share price of the two companies' as at the end of trading on Jan. 13, a 30 percent premium would put Altera at approximately $8.5 billion and Xilinx at $8.7 billion.

With a more likely 40 percent premium—I say this because Altera and Xilinx are two well-managed and highly competitive companies—a buyer would have to pay $9.2 billion for Altera and $9.4 billion for Xilinx. Is there anyone out there who still sees this scenario coming to pass in today's muddled financial environment?

Cash cows
As in every situation, there are other factors that might sway a buyer. In the case of the two FPGA vendors, a potential buyer would not need to secure financing for these huge sums because both Altera and Xilinx are cash spinning enterprises with huge liquid assets than can help offset the purchase price.

In Xilinx's case, a buyer offering a 30 percent premium would need to raise only $7 billion because the company closed its September quarter with $1.52 billion in cash and short-term securities, long-term investments and other long-term assets of $527 million and only $353.5 million in long-term debt, giving it net liquid assets of $1.67 billion.

Altera's situation is slightly different though equally appealing. The company had $1.36 billion in cash and short-term securities at the end of the September 2009 quarter and other long-term assets of $49.3 million. After deducting long-term debt of $500 million from Altera's cash hoard, a potential buyer would be able to tap the company for almost $1 billion.

Why would shareholders of Altera and Xilinx sell their investments and why would a venture firm or corporate buyer shell out billions for these companies considering these are not distressed or operationally inefficient enterprises? Their gross and operating margins, for instance, are the envy of many in the semiconductor industry.

It's difficult therefore to see how and where a private equity buyer or a corporate investor like Intel could squeeze out additional value out of companies that already run gross margins in the range of 65 to 67 percent and operating margins of 29 to 31 percent.

Also, a corporate buyer would be certain to disrupt the delicate balance of the FPGA market, a segment Altera and Xilinx have managed to parcel out among themselves.

Anyone buying either company, especially a fellow semiconductor company would have to spend the first years struggling to integrate the companies, keep critical employees and pacify OEMs and other supply chain partners, including component distributors Arrow Electronics Inc. and Avnet Inc., which each represent a major chunk of the two companies' businesses.

After more false starts and awkward stumbles, Intel will eventually transform its operations and diversify into whatever new markets the company see opportunities but that road is unlikely to pass through the headquarters of either Altera or Xilinx in San Jose, California.

- Bolaji Ojo
EE Times

This article was printed from EE Times-Asia located at::
http://www.eetasia.com/ART_8800595584_499495_NT_9d194e3d.HTM

댓글 없음: