Acquisitions are boring. 

We all think so.  That’s why we went into engineering.

So why should we talk about ChipX grabbing Oki’s US ASIC business?  Is it the unusual “Private US company acquiring part of a Japanese technology giant” aspect?  Is it the enigmatic cloud surrounding privately held semiconductor companies in general?  Is it seeing a company that’s bullish on ASIC technology when worldwide ASIC starts are lagging?  Did you notice that we used the word “bullish”?  You can tell we’re in financial land now.

For those of us that dally more in clock skew than acquisition terms, here are the highlights of the deal (closed-captioned for technology professionals).  In addition to acquiring the Oki US ASIC operation (that means they get standard cell lines ranging from 0.25 micron down to 0.13 micron), ChipX has signed an agreement with Oki for additional services, including turnkey foundry for former Oki lines (that means that Oki had a turnkey fab that built their existing parts including wafers, packaging, and test.  This fab is older and taps out at about .22 micron, however, so ChipX will continue to work with UMC for 130nm and 90nm development).  ChipX also gets access to Oki’s embedded array technology (giving them new embedded array lines that range from 0.5 micron down to 0.15 micron – conveniently leaving off right where ChipX’s new 0.13 micron embedded array picks up) and IP portfolio (nothing special here, this is marketing stuff, just move on along).  As part of the agreement, the entire Oki US ASIC engineering team is joining ChipX.  (20 or so experienced Oki people swap their business cards for ones with a ChipX logo.)  Oki keeps a stake in the operation by becoming a ChipX shareholder.  (More business stuff – no technical implications in particular.)

The acquisition considerably boosts ChipX’s position in the ASIC market in the US.  Previously focused on structured ASIC, the company will now have a complete range of technology options and a very large number of successful designs under their newly collective belts with over 1500 structured ASIC and over 400 embedded array designs.  The company will also enter new territory for them in the high-end market (10M gates/550MHz) but will benefit from Oki’s experience and expertise as all the US Oki engineers are transitioning as-is to ChipX.

For those of us embarking on most of our custom design work in FPGA, ChipX’s new size and breadth offer a wide range of options for cost reduction when our products succeed and head toward higher volume, demanding lower device cost.  With their unique USB 2.0 and PCI Express capabilities, a lot of designs that required high-end FPGAs to realize can be migrated to much lower unit cost structured ASIC technologies.  Likewise, structured ASIC designs that are part of products successfully hitting higher volumes can be migrated to even lower-cost standard cell platforms.

ChipX is already very experienced in the FPGA conversion business.  “About 70% of our designs start in FPGAs,” says Elie Massabki, VP of marketing at ChipX.  “Some larger companies start directly in ASIC and structured ASIC implementation, but the majority use FPGAs, at least for prototyping.”

The acquisition clearly moves ChipX out of the realm of “boutique” suppliers of structured ASIC and FPGA conversions and into the mainstream of ASIC technology suppliers.  By adding Oki’s ASIC customer base, they pick up a substantial amount of street cred in the ASIC arena, removing a lot of the cloud of uncertainty over the company’s health in the process (privately held companies are harder to monitor for business risk). 

The company is also adept at mixed-signal with a number of mixed-signal blocks that can be thrown into your design.  They’ve made the analog part an emphasis, and with increasing levels of integration in today’s high-volume designs, bringing a little of your mixed-signal action onto an ASIC can result in a significant savings.  Their Analog IP portfolio includes such things as 10MHz to 1GHz PLLs, 12bit 400KHz ADC, and 3.3V to 2.4V regulators.

“We find a lot of customers using our products as what we call ‘SideChips’ to add product-model-specific features and standards adaptation to a variety of products,” Massabki continues. “The SideChip sits alongside the main ASIC and provides capabilities such as bridging, bug fixes, and other features that are too risky to add to the original ASIC.”   In some cases, the SideChip may be extending the useful life of a legacy ASIC design, bringing it up to date with the latest standards such as PCI Express.

The design tool flow for both ChipX and Oki’s lines are based on standard 3rd-party EDA tools.  Although some work will be required to standardize the design flows exactly for the two parts of the business, the lack of proprietary tools in the mix is a good sign for both existing and potential customers.

Speaking of customers, ChipX says they have gone to great lengths to protect customers already engaged in design work with Oki US.  “All of the engineers are moving to ChipX,” Massabki observes, “and we made a priority of creating a seamless transition for companies already doing business and designs with Oki.”  That should be helped considerably by the lack of any major relocation of engineering resources.

When you go to buy something, you’ll find that the ChipX sales channel has been augmented with some of Oki’s direct sales as well.  Sales managers previously supporting Oki’s ASIC business have been migrated to ChipX, creating a hybrid structure where some accounts will be handled by direct sales and others by the traditional ChipX rep channel.  (If you bought Oki stuff or ChipX stuff before – chances are, you’ll probably still get it from the same guy.)

February 20, 2007

[back to top]