Like Intel, Microsoft buys scraps of parallel-processing startup

Contact: John Abbott

Despite a fair bit of talk about how important it is to demystify the art of parallel programming now that multiple cores and threads have become mainstream in x86 computing platforms, the actual level of activity has been surprisingly low. Over the last few years we’ve identified no more than a dozen small development tools vendors active in this area – some of them focused on the high-performance computing (HPC) sector – that appeared to have some prospect of success. And the companies with the most at stake in seeing better performance levels from new-generation CPUs (notably Intel and Microsoft) don’t seem to have been working particularly hard on the problem, either.

Perhaps, however, that’s starting to change. True, the number of startups is declining rather than expanding, but as they fail their assets are being acquired by larger vendors. One of the first to go was PeakStream in June 2007, snagged by Google after raising $22m in VC funding. But Google had no interest in sharing what it had bought. It withdrew PeakStream’s commercial product and began using it internally to boost the performance of its own software. Just last month Intel – currently in the process or rolling out six- and eight-core microprocessors – revealed that it had quietly picked up two small companies: RapidMind and Cilk Arts. And now Microsoft has announced, equally quietly, that it has purchased the technology assets of Interactive Supercomputing (ISC).

ISC had raised around $18m in VC funding over its four years of life, from Ascent Venture Partners, CommonAngels, Flagship Ventures, Fletcher Spaght and Rock Maple Ventures. It’s perhaps a bit of a stretch to call what ISC was doing mainstream, since it was focused on the HPC market. Its Star-P development environment let users create software models on their desktops using off-the-shelf packages from which parallel code could be automatically generated. The company claimed it could cut months from software development lifecycles. But Microsoft is talking about integrating ISC’s technology into its own products and using it for desktop computing as well as clusters. ISC CEO Bill Rock will bring over a team of experts to join Microsoft’s New England Research & Development Center in Cambridge, Massachusetts. Microsoft says it will continue to support existing Star-P users but won’t continue to sell the product in its current form.

Where’s the hurry in Oracle’s reach for Sun?

Contact: Brenon Daly

Having gotten the all clear on this side of the Atlantic, Oracle is now waiting for the EU to sign off on its pending purchase of Sun Microsystems. And the company will have to wait a bit longer. The European Commission has a deadline of September 3 to determine if the deal would violate antitrust measures. If the body decides that it does, a subsequent probe could potentially drag on into 2010.

Granted, there’s a lot at stake in Larry Ellison’s plan to use the acquisition of Sun to turn Oracle into a systems vendor, as opposed to a company that just sells software. (Provided the transaction goes through, Oracle will be in a position to hawk Solaris and Linux servers, all running its own database, middleware and application software on the boxes.) And, as the largest tech buy since Hewlett-Packard purchased EDS in May 2008, Oracle’s $7.4bn reach for Sun is clearly not nickel-and-dime M&A.

But the pace of the review by regulators is absolutely glacial. Consider this fact: It took Oracle just two months to fully negotiate its purchase of Sun, according to proxy material. (Sun chairman Scott McNealy spoke with Ellison about a possible deal in late February; the companies announced the transaction on April 20.) More than twice that amount of time has elapsed since Oracle announced the deal – and regulators in Europe are still mulling it over.

Tech buyers shop locally

-by Yulitza Peraza, Brenon Daly

Although the Delaware Court of Chancery was slated to rule this week on Emulex’s poison pill, the court punted on the decision. In postponing the ruling on the poison pill, which has been a key part of Emulex’s defense against the unwanted advances of Broadcom, the judge indicated that the two sides may well be able to work out a deal over the next week. Broadcom, which took its bid public on April 21, recently extended the deadline of its tender offer until July 14. The extension came as Broadcom also raised its bid to $11 per share for Emulex, up from $9.25. That added about $150m to the price of Emulex, which is currently valued at some $912m. As we noted earlier, Broadcom’s initial offer essentially valued Emulex where it was trading last October.

Unsolicited offers for tech companies, while increasing, are still relatively rare. However, in one regard, Broadcom’s bid for Emulex is rather typical. Scouring our data, we noticed a significant trend among California tech vendors: they tend to shop locally. That’s certainly true for these two southern California firms, which are only about 10 miles from one another. In the last seven years, about half of total tech M&A spending by California-based buyers went toward acquiring other Golden State tech companies. We would add that the ‘shop local’ trend isn’t limited to California, which stands as the most-developed tech region in the world. It’s also true on the other side of the country, where tech vendors based on the East Coast have spent more on acquiring neighboring tech firms than they have on companies from anywhere else.

There are a number of reasons for this trend, both formal and informal. For starters, the two sides are more likely to have a number of connections, sharing financial backers or board members, for instance. Additionally, executives at the companies may belong to the same local tech organizations or business groups. (Or, more informally, they may frequent the same restaurants or belong to the same clubs.) In some ways, our finding flies in the face of the oft-repeated notion that the world is flat, with business flowing around the globe without regard to borders or geography. That may well be true in some aspects. But when it comes to M&A, business is still largely done locally.

Geographic tech M&A, 2002-2009

Acquirer state/region Target Number of deals Percentage of total deals Total value Percentage of total value
California All 2,389 100% $247bn 100%
California California 879 37% $126bn 51%
East Coast All 2,391 100% $282bn 100%
East Coast East Coast 758 32% $83bn 30%

Sales of tech assets are on the rise

-by Thomas Rasmussen, Yulitza Peraza

At a time when both M&A volume and deal values have declined dramatically, the relative volume of asset sales continues to rise. There are two main contributors to this. First, companies are under increasing pressure to focus on their core operations, so they’re looking to divest underperforming business units. And second, cash-burning startups often find their venture backers unwilling to sink more money into them, resulting in wind-down sales of the intellectual property they had developed.

For the first quarter of 2008, the volume of asset sales represented some 15% of total announced transactions. That number doubled in the first quarter of 2009 and has even inched up a bit in April. About one out of every three transactions announced so far this year has been an asset sale.

For all the talk of unbridgeable valuation gaps, however, we would note that the buyers often get a sharp markdown on the price of the assets. Consider Artistdirect’s acquisition of SafeNet’s MediaSentry unit this month. SafeNet, which originally paid $20m for the division in 2005, wanted the MediaSentry assets off its books before the end of the first quarter, and Artistdirect’s new management was happy to fork over less than $1m for the unit. We understand that the deal closed within a few weeks. Or look at semiconductor startup Nethra Imaging, which picked up the assets of Ambric for an estimated $1m this month. Ambric had received an estimated $30m in funding, but when investors refused to step up with another round, the startup had little choice but to sell.

Asset sales spike

Period Volume of asset sales, as % of overall M&A
April 1-24, 2009 32.4%
Q1 2009 31.8%
Q4 2008 19.9%
Q3 2008 17.4%
Q2 2008 16.2%
Q1 2008 15.8%
Q4 2007 13.7%

Source: The 451 M&A KnowledgeBase

Going it alone can be expensive

Contact: Brenon Daly, Henry Baltazar

Wall Street hasn’t been particularly supportive of tech companies that turn down unsolicited offers and opt to go it alone. Shares in a number of the targeted firms are currently changing hands at less than half the level that the would-be suitors were willing to pay for them. To wit: Microsoft was reportedly set to pay in the mid-$30s for each share of Yahoo, which is now trading in the mid-teens. And having spurned a $16-per-share unsolicited bid from Cadence Design Systems last summer, Mentor Graphics stock is now trading at about $7.

We mention that bit of cautionary history because there’s another showdown brewing. Broadcom, advised by Banc of America Securities, recently offered $9.25 for each share of Emulex, giving the unsolicited bid a total equity value of $764m. (As it often does, Goldman Sachs is advising the target.)

Broadcom’s bid values Emulex where it was trading last October. On an enterprise value basis, the proposed transaction values the maker of storage networking gear at just 1.2x its trailing 12-month (TTM) sales and 5.5x TTM EBITDA. Emulex investors want a richer valuation and have pushed the stock above $10 since the offer was unveiled. Broadcom has vowed to take the unsolicited bid directly to shareholders if the Emulex board rebuffs it. On its conference call Monday discussing fiscal third-quarter results, Emulex said only that it was ‘thoroughly’ reviewing Broadcom’s offer.

From Broadcom’s point of view, it’s understandable why it would want its fellow southern California-based company. If the deal goes through, Broadcom would get a foothold in a few interesting storage markets such as host bus adapters (for both standard servers and blade servers) and embedded storage processors for disk arrays. Broadcom sells Gigabit Ethernet and 10-Gigabit Ethernet products, but is not a player in the SAN market. With network convergence growing in popularity, Broadcom would also benefit from Emulex’s fiber channel technology and its new Fiber Channel over Ethernet adapters.

Revenue refresh through M&A

Contact: Brenon Daly

Having recently lost several key Asian customers in the brutally competitive digital TV market, Trident Microsystems went shopping in Europe last week to rebuild its top line. On its trip, Trident got a pretty good bargain. It will hand over some seven million shares, valued at roughly $10.3m, to Switzerland’s Micronas Semiconductor for three consumer product lines. We understand that the three units were generating more than $100m a year in sales.

The purchase comes at a crucial time for Trident. Revenue at the company has plummeted from $258m in the past fiscal year, which ended last June. With two quarters and guidance for the third quarter already in the books, revenue at Trident has totaled just $61m. That implies sales for the current fiscal year could well be just one-quarter the previous fiscal year’s figure. Along the way, Trident has slipped from a profitable business to a cash-burning one. Its difficulties haven’t been lost on Wall Street, which values the debt-free vendor at about $100m, just half its current cash level.

Trident’s pending pickup of the three Micronas units should help, both on the top and bottom lines. (Union Square Advisors worked with Trident while Micronas went with hometown bank Credit Suisse Securities.) The company indicated that the acquisition should put revenue at $35m for the quarter that ends in September, essentially matching the previous year’s level. More importantly, the combination will boost Trident’s earnings from the very start and slow its cash burn. The firm will have more to say about the deal, which will dramatically expand its portfolio and end markets, when it reports fiscal third-quarter earnings later this month. Meanwhile, we would note that Trident shares are slightly above where they were when the vendor announced the acquisition

WDC goes SSD

-Contact Thomas Rasmussen

The market for solid-state-drive (SSD) technology is heating up. As an increasing number of consumer and enterprise products (including servers, desktops, laptops and netbooks) incorporate the technology, some old-line technology companies are looking to expand their SSD offerings. Western Digital acknowledged that last week by acquiring SSD vendor SiliconSystems for $65m in cash after about a year of on-and-off talks. (It was Western Digital’s first purchase since its $1.14bn acquisition of Komag in mid-2007.) On the other side, SiliconSystems had taken in just $14m in venture capital since its inception in 2002 from Miramar Venture Partners, Rustic Canyon Partners, Samsung Ventures America, Shepherd Ventures and SanDisk.

We understand that SiliconSystems generated about $50m in trailing 12-month (TTM) sales, meaning Western Digital paid about 1.3x TTM sales for the startup. This is in line with historical averages for the space, but comes at a time when the median valuation for venture-backed startups has been nearly cut in half. In the first quarter of 2009, the median valuation in a sale for a VC-backed tech company sank to just 2.1x TTM sales, compared to 3.8x TTM sales during the same period last year. (See our full report on first-quarter M&A.)

SiliconSystems will be re-branded as Western Digital’s Solid-State Storage business unit and will be headed by former CEO Michael Hajeck, who used to run STEC Inc’s enterprise SSD division. The importance of this relatively small acquisition should not be underestimated. Having essentially become a player in the SSD space overnight, Western Digital has taken the first step toward securing its future survival. With $1.4bn in cash, we wonder if Western Digital will continue to use acquisitions to expand in this market. Possible targets are Hajeck’s former employer STEC, which we previously speculated might be on sale, as well as Smart Modular Technologies. There are also a few potentially disruptive startups out there worth looking at such as Pliant Technology, Texas Memory Systems and Fusion-io.

Western Digital M&A

Date announced Target Enterprise value Price to sales multiple
March 30, 2009 SiliconSystems $65m 1.3x*
June 28, 2007 Komag $1.14bn 1.1x
July 24, 2003 Read-Rite $172m 1.0x

Source: The 451 M&A KnowledgeBase *451 Group estimate

IBM-Sun: Nothing but March madness?

Contact: Brenon Daly

Maybe the speculation around IBM buying Sun Microsystems was nothing more than a bit of March Madness. When reports surfaced last month that a deal could be in the works, Sun’s long-ailing shares soared from about $5 to nearly $9 in a single session. (At the time, we also looked at what a potential pairing of the tech giants might mean.) And it wasn’t just sporadic trading that powered the mid-March move. More than 160 million Sun shares traded the day after The Wall Street Journal carried its report on initial talks, meaning volume was eight times heavier than average.

It turns out that anybody who bought the stock from then until last Friday is now underwater. (Or to continue our NCAA basketball terminology, they’ve had their bracket busted.) Both the WSJ and The New York Times reported Monday that a deal – even at a lowered price – may be off the table. Sun shares gave up one-quarter of their value in Monday afternoon trading, falling to about $6.50 each. Volume was again several times heavier than average.

Amid all these reports of tough negotiating and ‘recalibrated’ deal terms, we’re reminded of the five-month saga of one public company buying another public company last year. In mid-July, Brocade Communications unveiled a $3bn offer for Foundry Networks, paying nearly all of that in cash and only a tiny slice in equity. As the equity markets plunged last October, the two sides agreed to lower the deal value to $2.6bn by trimming the cash price and removing the equity component. (Brocade shares had been cut in half during the time from the announcement to the readjustment.)

Now, the combined Brocade-Foundry entity, which has existed since mid-December, has a total market capitalization of just $1.5bn. In fact, my colleague Simon Robinson recently speculated that Brocade may be attracting interest from suitors. One of the names that has popped up? IBM, which would get an instant presence in the networking market. And if Big Blue is done with Sun (as reports suggest), then perhaps the company will just shift its M&A focus.

How do you say ‘please come back’ in Korean?

-Contact Thomas Rasmussen

When SanDisk released its dismal earnings this week, dismayed shareholders hastily headed for the hills. The exodus caused SanDisk’s stock to plunge 25%. In the fourth quarter of 2008, the flash memory giant lost $1.6bn, pushing its total loss for the year to $2bn. This red ink from operations was exacerbated by the company’s $1bn of acquisition-related write-downs stemming from its $1.5bn acquisition of msystems in July 2006. In the days following the dire news, SanDisk has been trading at a valuation of around $2.2bn. That’s a far cry from the $5.6bn that Samsung offered for SanDisk in September.

To put the decline in perspective, SanDisk’s three largest outside shareholders – Clearbridge Advisors, Capital International Asset Management and Capital Guardian Trust, which collectively own more than 15% of SanDisk (as of September 30) – suffered a paper loss of more than $700m since the day Samsung walked away from the proposed deal. Given this, we wouldn’t be surprised if shareholder ire forced SanDisk to reconsider its strategic options this year. On its earnings call this past Monday, the company reiterated that its board is indeed open to deal with any interested parties, which begs the inevitable question: Who might be willing buyers?

With private equity largely stymied and longtime partner Toshiba repeatedly stating that it’s not interested in a deal, Samsung is still the most logical fit. It has the cash, has shown a willingness to pay a solid premium, and would integrate well with SanDisk’s overall portfolio of products. In addition to its valuable intellectual property assets (which would eliminate those ugly royalty fees) and flash and solid-state drive lineup, SanDisk would instantly give Samsung the second-largest share of the music player market, behind only Apple. Perhaps it’s time for SanDisk CEO Eli Harari to brush up on his Korean, or at least learn how to say ‘please come back’ in that language.

Intersil: Doubling down in Austin

Contact: Brenon Daly

Intersil’s purchase of Zilker Labs last week had more than a few echoes of its pickup of D2Audio last July: same buyer, same banker, same backyard and even a shared backer at the acquired company. Both Zilker Labs and D2Audio are based in Austin and drew venture money from Dallas-based Sevin Rosen. (We understand that Al Schuele, Sevin Rosen’s lone VC in Austin, participated in funding both companies.) On the exit, boutique firm Pagemill Partners advised both Zilker Labs and D2Audio.

Despite the similarities between the exits of Zilker Labs and D2Audio, the companies had virtually nothing to do with each other up until that point. D2Audio makes digital audio power amplifiers, and primarily serves the consumer market. We estimate that Intersil paid around $25m for D2Audio. Intersil’s more-recent purchase of Zilker Labs added power-management technology to its existing portfolio. We estimate that Intersil paid about $18m for Zilker Labs, which raised some $33m in backing.

Intersil’s 2008 acquisitions

Date Target Target’s headquarters
December 18, 2008 Zilker Labs Austin
September 30, 2008 Kenet Woburn, Massachusetts
July 28, 2008 D2Audio Austin

Source: The 451 M&A KnowledgeBase