NXP, private equity, and the future, continued

About this time last year, we explored the creation of NXP—the company formed when private equity players purchased Philips Semiconductors. At that time we left some questions open about how the very long-term business of generating revenue from the underlying technology value of a semiconductor company would fit with the borrow-and-flip tendencies of the private equity world. Not surprisingly, NXP’s management assured us, then and since, that private equity money was patient money: that in fact the investors were encouraging NXP management to embark on a growth-through-acquisition strategy, not a cost-cutting strategy.

Not surprisingly, times have changed. Perhaps it is that the collapse of the global credit markets at which the private equity companies fed at will has left them a little short-tempered. Perhaps it is simply that even patient money has its limits. But in April, NXP spun off a portion of its operations into a widely-publicized joint venture with ST Microelectronics. Today we had a chance to talk with one of the executives who remain with NXP, Pierre-Yves Lesaicherre, senior vice president and general manager of the standard ICs business line.

In the transaction, NXP sent its two most R/D-intensive product lines, mobile communications SoCs and personal communications SoCs, to the joint venture. Technology that went to the joint venture in the deal includes baseband processors for cellular handsets, Bluetooth and USB connectivity IP, the RF technology acquired from Silicon Laboratories, and the GloNav GPS IP. In exchange, NXP received 20 per cent equity in the operation and a cash payment of $1.55 billion.

But the facts and figures leave out the why. And a good part of the why has to do with cash flow. "We at NXP were spending about $450 to $500 million a year on R/D," Lesaicherre explained. "ST was spending a similar amount. But neither of those expenditures by themselves was enough to keep up on all the developments necessary to stay in the handset SoC business. Companies like TI were spending twice that. By moving those two R/D organizations into the joint venture, we have created an entity with the critical mass to keep up in that market."

More to the point for the investors, perhaps, the act also moved a huge expense item off of NXP’s income statement, freeing cash flow for other things, such as interest payments and fees to private equity partners. The Multi-Market Semiconductor group in which Lesaicherre works, for instance, runs R/D expense at about 7 percent of sales. "That is a little low," the VP says. "In the past, with the huge demands of the mobile products, our home, automotive, and multi-market product lines have perhaps not had all the R/D investment they should have. We would like to change that now."

That change will start right away, apparently, as NXP’s owners have decided to invest at least some of the $1.5 billion in cash in R/D and in unspecified acquisitions. "We are planning to put R/D money into microcontrollers, our analog-mixed-signal products, and our complex discrete semiconductor lines," Lesaicherre said. "We may also look at acquisitions in these areas."

The new strategy appears to be one of targeting only markets where a leading, second-place or third-place position can be had without engaging in massive R/D outlay. An example of such an opportunity might be the discreet diodes and transistors, in which Lesaicherre hopes to be the market leader this year, on a pretty slim R/D budget. An exception might be the microcontroller market, which is sufficiently fragmented that there are very profitable niches for minority players, Lesaicherre said.

In these commodity-like markets, cost and service loom large. Accordingly, while NXP will pursue a fabless strategy for advanced CMOS products that are more feature-dependent, it will continue to be an integrated manufacturer for commodity products. "It is absolutely vital to maintain control of your costs in these markets," Lesaicherre said. "That means owning your own fully-depreciated fabs. You can’t be paying 10 or 12 percent of silicon costs to a foundry for their profit margin. And we will maintain our own back-end assembly operations. Packaging costs are incredibly important in these markets. Of our 7,000 employees world-wide, about 5,000 are in back-end operations."

The new strategy certainly seems more familiar for private-equity-held companies: seek market-leading positions with as little expense as possible, and spend on acquiring positive cash flows in existing markets rather than on basic technology. All of the routes to a happy future for the investors start out with strong cash flow growth.

So what will the future bring for the reconfigured NXP? Certainly more focus on its proven high-volume commodity products. Certainly acquisitions that can make it a broader-line supplier, able to help customers reduce their vendor counts. And certainly not a new fab. Wafers at 65, 40, and 32 nm will come from TSMC, not from an NXP fab.

But what about product lines? It appears that there are still some lines within the new Multi-Market Semiconductors group that are at the edge of, if not outside, these criteria. One might be the TV tuner products for digital TVs and set-top boxes. This group uses unique technology on unique processes, and it tends to be rather R/D-intensive. Another possible spin-off? Of course the executives would not comment.

And there is the microcontroller group itself, now the technology flagship of the organization. The microcontroller people have deep joint-development relationships with ARM, and are the leaders in adopting new CMOS processes, now that the SoC developers are elsewhere. They are tied to what recent data says is the highest-margin big foundry in the CMOS logic business. Plus, they are segment leaders, not market leaders. And they are in an area that consumes R/D expense at an appreciable rate, especially as 32-bit ARM-based microcontrollers more and more closely resemble ASSPs rather than MCUs, more and more requiring detailed application knowledge.

It seems unthinkable that NXP, with a heritage including microprocessor pioneer Signetics, could part from the microcontroller business. But in the cold logic of private equity, nothing is really unthinkable. Everything is just the present value of a cash flow, and everything sits on the same table.