Category: semiconductors

Contact: Scott Denne

SoftBank Group digs deep into its treasury for a bet that won’t pay off for several years. The company will spend $32.4bn to acquire ARM Holdings, a provider of chip designs for the mobile ecosystem. SoftBank will hand over $22bn of its own cash and fund the remainder of the all-cash deal with a bridge loan that it expects to repay with the proceeds of its sale of Supercell and a chunk of its Alibaba stock (both transactions were previously announced). That will leave SoftBank, which finished its recent fiscal year with negative free cash flow of $4bn, with about $2.5bn in cash and $25bn in debt.

The acquisition is the second-largest semiconductor deal, edged out by Avago’s $37m purchase of Broadcom last year. Despite the wave of large-scale consolidation in the chip industry over the past two years, $30m-plus chip pairings are rare. The third-largest transaction, the take-private of Freescale Semiconductor, was announced almost a decade ago and was just over half the size of today’s deal.

SoftBank will pay 20.9x trailing revenue for ARM. That’s the first time any company has cracked the 20x mark in a $1bn-plus chip acquisition. Even 10x has only been passed on two previous occasions, according to 451 Research’s M&A KnowledgeBase. As a supplier of intellectual property, not the chips themselves, ARM has a stronger profit margin compared with other chip vendors. That accounts for some of the high multiple. Still, the roughly 46x EBITDA multiple is one of the highest among such transactions.

Part of the rationale for the deal – and the valuation – is built on the emerging Internet of Things (IoT) opportunity. As a major licenser of system-on-chip technologies, ARM stands to play a major role in that market. And SoftBank, as a provider of wireless connectivity services in both Japan and the US, anticipates that substantial synergies will develop among the companies’ offerings, although it admits that such synergies won’t generate meaningful revenue or cost savings for many years.

That said, the overall growth of IoT will provide tailwinds for ARM to grow into its valuation with or without synergies. According to 451 Research’s Market Monitor, service providers globally will post $11bn in annual revenue servicing M2M connections. That number will nearly triple by the end of 2020.

For more real-time information on tech M&A, follow us on Twitter @451TechMnA.

Contact: John Abbott Scott Denne

Sixteen months after its direct rival and onetime parent Emulex was swallowed up by giant chipmaker Avago, QLogic is set to also become part of a more diversified silicon company. Networking and communications semiconductor firm Cavium has agreed to acquire QLogic in a deal that values the target at approximately $1bn. Over the past few years, Cavium has been showing an increasing interest in enterprise and cloud datacenter infrastructure, looking beyond networking into the server and storage sectors. It says there’s little product overlap and plenty of synergies in the combination. Cavium has taken a long, hard look at QLogic’s product portfolio and plans to immediately kill off several legacy product lines when the transaction closes to boost QLogic’s tepid growth into the double-digit range.

Both Emulex and QLogic needed to become part of larger organizations to survive and prosper. Their key positions as suppliers to the server and storage OEM market made them highly desirable properties within more diversified chipmakers, where cross-selling opportunities are everywhere. And with increased activities focused on converged infrastructure, further opportunities are emerging. There is also a clear need for Cavium to diversify. Nokia and Cisco are its two biggest customers, while Alcatel-Lucent (now merged with Nokia) was its fifth-largest client. This isn’t quite as dangerous as it sounds, as there are many design wins spread across the different divisions of those vendors. However, it’s in Cavium’s best interest to extend its business at scale into the datacenter and storage sectors, and to diversify both its customers and revenue sources.

Cavium will pay $15.50 ($11.00 in cash and 0.098 of a share of Cavium) per QLogic share. The deal value is $1.4bn and after backing out QLogic’s cash, it gives the target an enterprise value of about $1bn. Combined, the companies generated $870m in revenue over the past 12 months, with just over half coming from QLogic. Cavium will fund the purchase with a $750m loan, $400m in new Cavium equity and the remainder in cash. The transaction is expected to close in Q3 2016.

This is the latest in a string of acquisitions that shows little sign of slowing, despite an overall deceleration of tech M&A this year. QLogic’s sale marks the eighth $500m-plus semiconductor deal of the year and puts 2016 on pace to best last year’s record tally of such transactions. The level of consolidation and remaining number of chipmakers that can command that kind of valuation point to an impending slowdown.

For more real-time information on tech M&A, follow us on Twitter @451TechMnA.

Contact: Mark Fontecchio

Go big or go home has been the mantra in the semiconductor business of late, and there is no better example than Avago Technologies’ $37bn reach for Broadcom today. The purchase price is more than double the next-biggest chip deal that we’ve tracked, with the combined company becoming one of the largest global suppliers of semiconductors.

With four transactions for nearly $45bn, Singapore-based Avago has become the most active acquirer in the semiconductor sector since 2013 in volume and value, according to 451 Research’s M&A KnowledgeBase. Its two largest – for Broadcom today and LSI in 2013 – have diversified Avago’s chip portfolio with storage and networking semiconductors so it is less reliant on the volatile wireless business. They have also followed the pattern of consolidation that has infected the entire semiconductor market, with buyers seeking big targets with opportunities to cut operating expenses. To wit: Broadcom brings in about 30% more sales than Avago, but its profit margin is 14 percentage points lower. Avago wants to reach a 40% margin on the combined entity, which is higher than either company alone.

The 39.1x multiple of Broadcom’s enterprise value over trailing EBITDA is almost three times the median EBITDA multiple on billion-dollar chip deals, according to the KnowledgeBase. Broadcom’s continued growing revenue and the paucity of remaining large semiconductors targets are two main factors in that higher-than-usual valuation. The deal includes $17bn in cash and the rest in Avago stock, with Broadcom shareholders owning about 32% of the combined business. Avago will fund the acquisition with $8bn in cash, $9bn in new debt and 140 million Avago shares worth $20bn. The transaction is expected to close early next year. Both boards have approved the deal, but it’s still subject to approval by regulators and shareholders.

Contact: John Abbott Daniel Bizo

Amazon’s somewhat surprising acquisition of stealthy chip startup Annapurna Labs for a reported $350-375m isn’t perhaps as unexpected as it appears at first sight. One of the exit strategies of such startups nowadays is to be sold off to a larger company building or operating its own systems hardware that has reached the stage where it needs its own custom silicon. That means the startup abandons the aspirations it had to be a more broadly applicable company. The acquired personnel typically become an internal chip design team for their new parent.

The most obvious example is Apple’s $278m acquisition of P.A. Semi in April 2008 . Apple obtained a 150-strong team of engineers, including the lead designer of the DEC Alpha and StrongARM processors, that boosted the development of its A series chips used in the iPhone and iPad. Apple followed up that move by buying Intrinsity in April 2010 for a reported $121m and flash memory chip designer Anobit in December 2011 for a reported $500m. Three years later, Apple snared yet another silicon startup, Passif Semiconductor, for its wireless networking chips.

Another systems maker that has its eye on chipmakers is Oracle. “You could see us buy a chip company,” said Oracle chief Larry Ellison back in 2010 . It hasn’t yet, but Ellison continues to hint. At the recent launch of Oracle’s X5 Engineered Systems range, Ellison told the audience that the company was in the process of moving more software functionality into silicon: “We are doing a lot of it,” he said. The easiest way of doing this while keeping full control would be to buy a team with expertise in hardware acceleration.

Closer, perhaps, to what Amazon is doing is the example of Google, which bought early-stage chip startup Agnilux in April 2010. Agnilux had been formed by some of the P.A. Semi team that subsequently left Apple. Even before that (in June 2007), Google had acquired PeakStream, a company that had developed a layer to make the programming of multicore processors easier. And since the Agnilux buy, Google has hired several prominent chip designers, including HP’s Partha Ranganathan, who was involved in the development of Moonshot. Google has also been quite public about its interest in IBM’s OpenPower initiative and the possibilities of using OpenPower as the basis for bespoke chip development (although it’s fair to say that things have gone quiet recently in this area).

Facebook has also been investigating the possibilities of its own chip design. The company already designs its own servers and was the driving force behind the formation of the Open Compute Project, a means of opening out the specifications of system designs so that customers can modify servers to more closely fit their own workload requirements.

Which brings us back to Amazon, which began advertising for chip designers with ARM expertise last year, and hired former Calxeda chip designer Mark Davis as the manager of a new hardware engineering and silicon optimization team based in Austin, Texas. (As an aside, defunct chip startup Calxeda, which ran out of money at the end of 2013 while trying to develop an ARM server chip, may itself reemerge – its intellectual property assets were picked up at the start of 2015 by Silver Lining Systems, a division of Taiwan-based systems provider AtGames Cloud Holdings, which is working in conjunction with ARM and the server group of Taiwan-based ODM Foxconn Technology).

Little is currently known about Annapurna Labs, an Israel-based company founded in 2011, except that some of its systems-on-a-chip parts are already being deployed in some entry-level storage systems using standard ARM processing cores, integrated networking and IO controllers. Amazon Web Services will likely employ Annapurna’s silicon-tailoring expertise to gain an edge in storage cost performance over rival cloud providers by using unique chips in its storage systems and, over time, networking gears. We expect to see more chip M&A activity from both traditional systems vendors and giant scale-out datacenter operators.

For more real-time information on tech M&A, follow us on Twitter @451TechMnA.

Contact: Scott Denne

In what’s becoming a familiar refrain in semiconductor M&A, Cypress inks its largest-ever deal with the $1.6bn stock acquisition of Spansion. The transaction, which will leave Spansion shareholders with 50% of the combined company, is the sixth $1bn-plus takeout of a chipmaker this year. Like Cypress, all but one (Qualcomm) of the acquirers had never crossed that threshold before.

What’s more, most of the dealmakers driving this year’s big purchases hadn’t previously been very active. Cypress, for example, had only bought three companies for a combined $198m in the decade prior to yesterday’s announcement. Or take Analog Devices, which snagged Hittite Microwave for $2.4bn this summer. Prior to that deal, it had only done seven acquisitions since 2002, and never paid more than $150m. Only RF Micro Devices had come close to the $1bn mark, with a $900m purchase back in 2007.

That data points to two polarized trends in the semiconductor market: the exponentially growing cost of building a new chip vendor in the past 10 years has left a dearth of modestly sized targets, while better margins through bigger scale has become a key differentiator for many businesses. Cypress’ management points to a need to get its opex margins down to 30%, from about 37%, as a motivation for the Spansion buy. A few $1bn transactions aren’t an anomaly in the chip business. According to The 451 M&A KnowledgeBase, the median deal value has risen in each of the past five years to $92m in 2014 from just $33.9m in 2010.

For more real-time information on tech M&A, follow us on Twitter @451TechMnA.

Contact: Scott Denne

SiTime nabs a $200m exit in a sale to Japan’s MegaChips, navigating a market that’s crowded with failed competitors. Time, ironically, seems to be the element missing from earlier attempts by semiconductor companies to break into the timing market.

SiTime designs chips that take the place of the quartz crystals that for decades have provided the timing element in nearly all electronic devices – from high-end communications equipment to everyday consumer electronics. As silicon is smaller, cheaper and more energy efficient than quartz, it’s a large and promising market. You wouldn’t know it looking at SiTime’s peers, almost all of which ran low on capital while waiting for OEMs to get comfortable enough to replace quartz.

Earlier deals in this space have ranged from $6m-25m, typically at a loss to investors. While not the oldest of this group, SiTime had products in the market far longer, having generated revenue since 2008 (last year it posted $15.5m). It was also better capitalized: SiTime raised $79m from investors, including a $25m debt and equity round just last month.

Needham & Company advised SiTime on the transaction.

Acquisitions of timing chip vendors

Date announced Target Acquirer Deal value October 29, 2014 SiTime MegaChips $200m August 30, 2013 Discera Micrel $6.1m March 30, 2012 Multigig Analog Devices $24.2m April 28, 2010 Silicon Clocks Silicon Labs $21m January 14, 2010 Mobius Microsystems Integrated Device Technology $20.2m

Source: The 451 M&A KnowledgeBase

For more real-time information on tech M&A, follow us on Twitter @451TechMnA.

Contact: Scott Denne

Murata Manufacturing’s $465m acquisition of Peregrine Semiconductor marks a tranquil end to the target’s two years as a public company. Peregrine went public at $14 per share two years ago this month and has agreed to sell to Murata for $12.50 per share. Though the price is a loss for its initial shareholders, it could have been far worse and the 76% premium from Peregrine’s price 30 days ago helps soften the blow.

Peregrine came through 2012 with a gain for shareholders, then a key OEM (rumored to be Apple) diversified its radio frequency component supply chain, sending the company’s stock down and contributing to stagnated revenue since the start of 2013. Peregrine is not the only semiconductor vendor that’s been hurt by an overreliance on one of the two major smartphone OEMs (Samsung and Apple). Earlier this year, Cirrus Logic snagged Wolfson Microelectronics in a $488m deal aimed at spreading out the customer base for its audio chips, and Audience, a Cirrus competitor, picked up Sensor Platforms for $41m to try to push its products beyond the smartphone market.

According to The 451 M&A KnowledgeBase, the premium for Peregrine’s stock is the highest for any chipmaker in the past 12 months, but Peregrine shareholders aren’t the only ones to get bailed out of a poor-performing semiconductor stock with a hefty premium. Avago Technologies’ $6.6bn purchase of LSI at 36% brought that company’s stock up to levels it hadn’t seen since 2006, and the 75% premium that M/A-COM Technology handed to Mindspeed Technologies shareholders helped level off a 36% drop in the target’s stock in 2013.

For more real-time information on tech M&A, follow us on Twitter @451TechMnA.

Contact: Scott Denne

Analog Devices looks to the cellular infrastructure market to burst out of its growth funk, paying $2.4bn to acquire Hittite Microwave. Not only is the purchase Analog’s largest by a factor of 20, it also has an unusually high multiple.

In the company’s most recent quarter, sales of chips for the communications infrastructure market jumped 25% year over year, surpassing management’s expectations for the sector. (Communications accounted for 20% of its 2013 revenue.) That level of growth – beyond anything Analog typically sees in its various segments – points to a way out of the company’s growth slump. The signal-processing vendor’s annual revenue skidded down 12% between 2011 and 2013.

Hittite specializes in communications chips that utilize high frequencies, compared with Analog’s communications portfolio, which was built around the lower frequencies typically used in cellular communications. As frequencies become increasingly crowded, cellular signals are drifting into higher frequencies, and Hittite has benefitted, as its portfolio, which was initially geared toward government and military applications, has seen growth in the broadband and cellular markets.

While Analog expects there will be opportunities to cut costs when the deal closes, the chance to grow by cross-selling the higher- and lower-frequency products drove this acquisition – and pushed up the price to 7.1x Hittite’s trailing 12-month revenue. Since cost synergies, not new sales, drive most chip transactions these days, the multiple for Hittite is substantially larger than most. In fact, it is three times larger than the median for chip deals of more than $500m in the past 12 months and is the highest multiple paid in such a transaction since Broadcom’s reach for NetLogic Microsystems in September 2011.

For more real-time information on tech M&A, follow us on Twitter @451TechMnA.

Contact: Scott Denne

Microchip Technology prints its first major international purchase, picking up Taiwan’s ISSC Technologies for $328.5m. Despite averaging two acquisitions per year since 2009 and getting about two-thirds of its revenue from Asia, this is its first M&A foray into the region.

The appeal of avoiding US taxes by putting its foreign cash to work is apparent in the multiple Microchip’s paying. The 4.3x trailing 12-month revenue makes ISSC the highest-valued company Microchip has bought. Not all of the multiple can be attributed to ISSC’s location – its revenue doubled to $69.2m last year, giving it an atypical growth rate for a Microchip target.

ISSC likely won’t be Microchip’s last major foreign acquisition (it previously bought a European vendor, EqcoLogic, for $9m). Microchip’s stock price is up more than 200% since embarking on its recent M&A tear in 2009 (before that year, it only inked four purchases). Also, about $2bn of its roughly $2.1bn in cash at the end of last quarter is still overseas.

For more real-time information on tech M&A, follow us on Twitter @451TechMnA.