SAN JOSE, Calif. — The proposed merger of NXP and Freescale creates a big company but not a substantially different one. With the exception of its size, the combined giant will have few new tools to ensure its survival in a consolidating chip industry.
Assuming the deal closes as expected before the end of the year, executives are confident they can shave $200 million off annual 2016 costs largely from administrative overlap and expanded buying power. The savings could expand to a maximum $500 million/year at some undetermined point in the future, executives said. However, they are not prepared yet to set any targets for accelerating growth in revenues or profits.
As part of the deal, NXP will sell its high-performance RF unit in a nod to its overlap with a stronger unit at Freescale and to avoid regulatory issues. Rick Clemmer, chief executive of NXP and planned CEO of the merged company, said he does not anticipate any plant closures.
The guidelines suggest layoffs will be limited, especially among engineering ranks. However, executives have yet to work out many details, including overlap in the two companies’ ARM Cortex-M microcontroller lines.
The combined company would have more than $10 billion in annual revenues, becoming the world’s ninth largest semiconductor company. It would easily surpass rivals STMicroelectronics and Renesas at about $7.3 billion each, however it would still be a notch below Texas Instruments and Toshiba.
Clemmer correctly highlights the merged NXP would be a leader in automotive chips (see figure, right). The sector is growing rapidly as cars become more digital, networked and electrically powered, but it still represents one of the relatively smaller slices of the electronics pie.
Even with the merger, the new NXP would have only a 13% market share, a shade above the nearest competitors at 11 and 9%. “The area of most [product] overlap [in the deal] is likely for automotive MCUs,” said Christopher Rommel, analyst for market watcher VDC Research Group.
Executives also highlight the company would be a close second, by its own count, in general-purpose microcontrollers. VDC has a different tally, pegging the new NXP at a 17% share versus 25% for Renesas, the market leader.
However you count it, the MCU space is highly fragmented and hotly competitive. All the players are adopting ARM Cortex-M cores, making differentiation increasingly difficult.
Like the fragmented MCU business, the two companies serve a broad set of embedded markets. Outside of greater heft in automotive, the merger does nothing to provide a sharper or new focus.
Clemmer might decide to divest Freescale’s high-end communications business although he said nothing overtly to suggest such a move. When an analyst asked about the segment he noted, “we’ve been trying to move off the [process] technology treadmill and Freescale is more in line with it… From our due diligence at a very limited level so far, it is still complementary and fits together nicely.”
The reality is the communications business is becoming the largest slice of the diverse electronics pie. However it is one where Intel is gaining sway is comms apps move to software running on x86-based systems. Like many embedded chip vendors, Freescale trying to shift to ARM-based SoCs and find ways to differentiate itself here.
It’s not clear if Clemmer puts the Freescale comms SoC segment into his overall category of high performance mixed-signal chips he sees as the company’s future. From the old NXP point of view, the merger is good because it reduces the company’s dependence on relatively low margin so-called standard products from 22% to 12% of revenues.
Analyst Rommel of VDC sees a “potential downside” to the deal if NXP does not maintain Freescale’s focus on software initiatives.
In my opinion, Freescale’s increased focus on software solutions and partners has helped them gain share in recent years. Freescale's growth was inhibited for a few years because they did not properly incubate the software and tool assets they acquired with Metrowerks. It is critical the combined entity maintain its investment in these types of enabling software and tool solutions.
Indeed one of the big problems of the semiconductor industry in general is in getting paid for the value of the software it is expected to provide with its chips.
To read the rest of this article, visit EBN sister site EETimes.