VCs piling up chip deals

This year’s surging M&A market has brought relief to a tech sector that’s long struggled to find exits. With today’s announcement that Cisco is acquiring venture-backed Luxtera for $660m, venture capitalists have realized more value from their semiconductor investments than any time since 2013. In a reflection of the broader market, the return of strategic acquirers has boosted the sales of chip startups.

As 2018 heads toward a close, buyers have spent a collective $1.4bn on purchasing semiconductor startups from venture portfolios, compared with just $1.8bn in the three previous years combined. The acquisition of Luxtera goes a long way toward boosting this total. According to 451 Research’s M&A KnowledgeBase, the deal is the largest purchase of a VC-backed chipmaker since the dot-com bubble burst. (Neither that record nor the annual totals include acquisitions of public companies that previously raised venture capital.)

That Luxtera’s exit marks a record speaks as much to the dearth of liquidity for chip startups as it does to the target’s accomplishments. In the face of rising development costs, VCs have long shied away from chip investments and acquirers have come to depend on product development more than M&A for incremental improvements. Most of the dealmaking among semiconductor vendors in recent years has been large consolidations, rather than midmarket acquisitions.

Still, there’s been some relief from the paucity of exits this year. Prior to Luxtera, Cisco, a frequent buyer of software vendors in recent months, hadn’t acquired a semiconductor business in nearly three years. The same goes for Skyworks, which provided VCs with the third-largest exit in the category when it shelled out $405m for Avnera in August – it was the acquirer’s first purchase in more than 30 months. And Intel picked up a pair of chip startups this year after a 16-month hiatus from semiconductor M&A.

Life after a breakup

by Brenon Daly

A lot has happened since Qualcomm announced its as-yet-unclosed $44bn acquisition of NXP Semiconductors in October 2016. The would-be buyer has successfully fended off an unwelcome suitor in what would have been the tech industry’s largest transaction. It has seen a former king at the company look to raise an army and reclaim the throne. Meanwhile, as Qualcomm’s deal has languished in limbo, more than 150 other chipmakers have been gobbled up, according to 451 Research’s M&A KnowledgeBase.

While those matters are mostly closed, Qualcomm’s bid for NXP remains open. At least it remains open for another day. The deadline for gaining the last remaining regulatory approval for the transaction is Wednesday night, just hours after it is scheduled to report fiscal third-quarter results.

Qualcomm and NXP have the green light from all of the necessary national and international bodies except one: China. Qualcomm extended its bid for NXP last April, when it refiled its application to China’s Ministry of Commerce. At the time, the company said it would walk away from the deal if it didn’t get approval on July 25 and pay a $2bn breakup fee to NXP the following day.

Right now, that appears likely to happen. Investors have put almost a 20% discount on NXP shares, compared with Qualcomm’s already raised offer of $127.50 in cash for each share of NXP. Over the past month, the discrepancy between the two prices has widened in almost every trading session. In mid-afternoon trading on Tuesday, NXP stock was changing hands at about $102.

Of course, that period has also seen a near continuous escalation in the trade war between the US and China. (We recently noted how tech acquisitions and investments have suffered pretty serious collateral damage in the ongoing spat between the two economic superpowers. China is currently on pace to purchase the fewest number of US tech providers since the country began shopping here about a half-decade ago, according to the M&A KnowledgeBase.)

Assuming Wall Street’s terminal view of the deal does indeed come to pass, what will happen to the two sides? For NXP, it’s pretty simple: deposit the $2bn termination fee into its treasury and go on with business. (It’s a pretty significant windfall for the company, which generated only $2.2bn in profit in all of last year.)

For Qualcomm, which would remain inexorably tied to the ever-maturing cellphone market, the options are a bit more limited. Nonetheless, one move we can probably rule out: Unlike Broadcom, the chipmaker that tried to buy Qualcomm, we don’t see the 33-year-old semiconductor provider announcing a multibillion-dollar purchase of a software vendor. Even two weeks on, Broadcom’s $19bn acquisition of CA Technologies is still a bit of a head-scratcher.

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

MIPS takes Wave to the edge

Eyeing a move from training to endpoints, Wave Computing has acquired MIPS Tech, a pioneer in the development of RISC processors. The target, recently spun off of Imagination Technologies, provides the buyer, a designer of artificial intelligence (AI) semiconductors, the chance to sell its wares more broadly as organizations look to run AI algorithms directly on the endpoint.

Founded in 2010, Wave is one of multiple startups producing accelerators for AI workloads, and one of a smaller select group (along with Cambrian Systems, Cerebras Systems, Graphcore and Horizon Robotics) that have already raised over $100m in venture funding. Wave emerged from stealth in 2016 and made its compute appliance for training neural networks available to early-access customers in 2017. In March, Wave said it would be using the MIPS core as the integrated CPU within its next-generation Dataflow Processing Unit.

We noted in a previous report that MIPS would likely prove a valuable asset for AI applications. It’s been in business since the 1980s and has a significant embedded systems user base and a range of extensible cores. Once owned by SGI, MIPS ended up as part of UK-based GPU maker Imagination Technologies, but when Imagination was sold to China-based investor Canyon Bridge Capital Partners, US-based MIPS had to be divested separately to Tallwood Venture Capital for $65m. Tallwood is also an investor in Wave.

It’s likely that the power-efficient MIPS cores will be useful for the development of inference processors within edge devices, giving Wave an end-to-end story beyond its initial training focus, increasing its potential total addressable market significantly. MIPS will continue to be run as an independent unit, and will continue licensing its cores to third parties. 451 Research’s recent VotE: Internet of Things report shows that many companies are already making advanced calculations right on the endpoint – 40% of respondents claimed to run data analysis, cognitive computing or AI at the network edge or perimeter.

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

Connected home is key to Sigma Designs’ valuation 

Contact:Scott Denne

The addition of products for emerging Internet of Things (IoT) markets has spurred higher valuations on many semiconductor deals amid a record amount of consolidation in the space over the past three years. Silicon Labs’ acquisition of Sigma Designs highlights how much higher those products are valued amid larger semiconductor portfolios. The terms of the $282m transaction revolve explicitly around the seller’s home automation hardware.

According to 451 Research’s M&A KnowledgeBase, the median valuation for a semiconductor vendor has hovered a bit below 2x trailing revenue for most of the past decade, although it surged to 2.9x amid a profusion of IoT-related purchases. Indeed, many chip deals with an IoT element have traded well above the standard valuation. Consider Qualcomm’s pending $39bn pickup of NXP Semiconductors (5.5x), SoftBank’s $32bn reach for ARM (20.9x) and Intel’s $15bn acquisition of Mobileye at an unheard of 41x trailing revenue.

The $282m price tag for Sigma Designs ostensibly values the company at about 1.5x trailing revenue, but only if certain conditions are met – the seller must have at least $40m in cash on its balance sheet and it must shutter or sell its Smart TV components business within a week. (That business accounts for the majority of its revenue, but isn’t related to IoT and is heading downhill, having lost 61% of its quarterly revenue since this time a year ago.)

If those conditions aren’t met, Silicon Labs will buy Sigma Designs’ home automation components business, Z-Wave, for $240m. In other words, Silicon Labs wants to buy Z-Wave for about 5-6x trailing revenue and doesn’t value the other business lines at all. There’s a logic to that premium valuation. The market for home automation is advancing with plenty of open space. According to a third-quarter survey by 451 Research’s VoCUL, 69.9% of consumers still don’t own any internet-connected home devices, although that’s down from 74.4% at the start of the year.

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

Marvell’s belated bid to be a chip consolidator 

Contact: John Abbott, Scott Denne

With its $6bn reach for Cavium, Marvell Technology Group proves that a few large targets remain as the semiconductor industry emerges from a record streak of consolidation – a streak that happened with Marvell on the sidelines. Although deal value in semiconductor M&A remains well below the record levels of 2015 and 2016, the transactions getting done are commanding higher amounts.

Cavium becomes the fourth chipmaker to be acquired for more than $1bn this year, compared with 10 in all of last year, according to 451 Research’s M&A KnowledgeBase. Yet only one of the $1bn-plus deals this year has been done for less than $5bn, whereas half of 2016’s 10-digit semi transactions fell below that threshold.

Marvell doubles its market opportunity by purchasing Cavium and enters the high-growth datacenter market. Its current portfolio spans storage controllers, networking PHYs and SOCs for enterprise switches, and Wi-Fi and Bluetooth SOCs for wireless connectivity. To that Cavium adds compute, networking, storage and security components for the datacenter, including multi-core and datacenter processors, Ethernet adapter and datacenter switches, Ethernet and fiber-channel storage connectivity, and FIPS and virtual offload security.

Benefits of scale and volume include a full portfolio that will enable cross-selling, as well as pooled R&D expenses, where there is currently a lot of duplication – moving up to 10nm and 7nm process technology is a huge burden that can now be consolidated. Diversification will also reduce Marvell’s exposure to low-growth sectors, such as hard disc drive controllers and notebooks, and Cavium’s to the dwindling fiber-channel business. The two companies are located close to each other, easing integration challenges.

The change marks a departure from Marvell’s past M&A strategy. Since the start of 2002, the most it had ever paid for any asset was $600m, with most of its deals falling well short of that mark. Moreover, this is its first acquisition since early 2012 and only the third time it has bought an entire company, rather than a business unit.

Subscribers to 451 Research’s Market Insight Service will have access to a detailed report on this transaction later today.

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

Piling up the chips

Contact: Brenon Daly

Unveiling what would be the largest tech transaction in history, Broadcom said it is prepared to hand over $103bn in cash and stock, and assume some $25bn in debt, for Qualcomm. The unprecedented 12-digit pairing represents a consolidation of the two consolidators behind the semiconductor industry’s two largest consolidations.

To get a sense of the sheer scale of Broadcom’s ambitions, consider that this single deal would roughly match spending on all acquisitions in the chip industry from 2008-14, according to 451 Research’s M&A KnowledgeBase. However, regulatory challenges mean this marriage of giants highly is unlikely to go through. And that assumes Qualcomm even picks up negotiations with Broadcom and its relatively low opening bid.

Thus far, Qualcomm has pretty much dismissed Broadcom’s offer. That doesn’t mean Broadcom, which is being banked by six separate firms, won’t push the deal.

Broadcom is negotiating from a position of strength, while Qualcomm is suffering through a well-publicized legal fight with major customer Apple and has still come up empty in its high-risk effort to buy its way into new growth markets. (Qualcomm originally hoped to close its $39.2bn purchase of NXP Semiconductors, which makes chips for cars and Internet of Things deployments, by the end of this year. That appears unlikely, and Broadcom has said it wants to acquire Qualcomm whether NXP closes or not.)

Broadcom’s relative strength also comes through in the pricing of the transaction as it is currently envisioned. At an enterprise value of $130bn, Broadcom is valuing Qualcomm at just 3.9x its pro forma 2017 sales of $33bn. (That assumes Qualcomm, which will put up about $24bn in sales in 2017, does buy NXP, which will generate $9bn in sales.) That’s substantially lower than the 5.5x sales Qualcomm plans to pay for NXP on its own, and a full three turns lower than the nearly 6.9x 2017 sales where Broadcom trades on its own.

When the chips are down 

Contact: Scott Denne

After two years of record consolidation among semiconductor companies, M&A has deteriorated as fewer sizeable targets remain. The two most recent significant sales in the chip industry show that suitable buyers are just as lacking as the sellers in both deals ink uncertain transactions. In the largest chip transaction of the year, Toshiba has agreed to a $17.9bn sale of its flash memory business to an unwieldy syndicate – meanwhile, Imagination Technologies sold itself off in two separate deals (worth $800m) aimed at avoiding objections from US regulators.

Toshiba finds itself forced into the sale of its flash business to raise capital following the bankruptcy of its nuclear subsidiary – certainly not ideal conditions for a sale. Yet the winning bidder for the second-largest maker of flash memory, a syndicate led by Bain Capital, raised legal objections from Western Digital, a flash storage vendor that has a joint venture (JV) with Toshiba.

There are questions about how well the group could manage the asset, even if the buyer manages to get the transaction over the finish line – an uncertain prospect given the spat with Western Digital. The investor group (Bain, Apple, Dell, Toshiba, Seagate, SK Hynix and three others) called and abruptly cancelled a press conference on the deal. A squabble over media strategy doesn’t bode well for setting a coherent course for a business with $7bn a year in revenue.

Earlier this week, UK-based Imagination Technologies announced that it will sell most of its business to Canyon Bridge Capital Partners, a China-funded private equity firm whose proposed takeover of Lattice was recently shot down by the Trump administration. Imagination is selling its US-based MIPS business to Tallwood Venture Capital in hopes of avoiding such a fate.

If the Toshiba deal stands up to multiple legal challenges from Western Digital – the company claims Toshiba has limited rights to transfer ownership of a JV between the two companies – it will nearly double the size of semiconductor M&A this year to $44.2bn, a pace that’s less than half of last year’s, according to 451 Research’s M&A KnowledgeBase.

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

Trump’s death blow to a deal

Contact: Brenon Daly

Respondents to the previous edition of the M&A Leaders’ Survey from 451 Research and Morrison & Foerster have once again delivered the wisdom of the crowds. When asked last spring about the outlook for US-China tech deal flow, respondents overwhelmingly predicted that President Trump’s policies would crimp M&A activity between the world’s two largest economies. Specifically, two-thirds (65%) of the 157 respondents from across the tech M&A landscape forecast a decline in purchases of US tech companies by Chinese buyers. That was more than four times the level (14%) that anticipated an increase.

In line with that April forecast, Trump has blocked the proposed $1.3bn acquisition of Lattice Semiconductor by a Beijing-based fund, citing national security concerns. Regulatory approval of the planned purchase by Canyon Bridge Capital Partners, which was announced last November, had been viewed as virtually impossible after The Committee on Foreign Investment in the US indicated that it would not sign off on the transaction. Trump delivered the death blow to the deal on Wednesday.

Trump’s move represents a rare bit of White House intercession in an acquisition. But it isn’t necessarily out of character for Trump, who has singled out China for some of his sharpest criticism as he has pursued a self-described ‘America First’ policy. Again, respondents to the M&A Leaders’ Survey last spring accurately predicted that Trump’s singularly unfriendly views toward China would disproportionately impact US-Sino deal flow. In the survey, fully one out of five respondents (20%) forecast that Chinese buyers of US tech companies, such as Lattice Semi, would ‘substantially’ cut their activity due to the Trump administration, compared with just 3% who said they expected overall cross-border M&A to drop off ‘substantially’ in the current regime.

451 Research and Morrison & Foerster are currently in market with the latest edition of the M&A Leaders’ Survey, and would appreciate your views on where the tech M&A market is and where it’s heading. In addition to broad market questions, we also revisit questions around Trump’s impact on cross-border M&A as well the specific outlook for China-based buyers. We would appreciate your time and thoughts. To participate, simply click here.

An M&A break for chip vendors

Contact: Scott Denne

Intel’s $15.3bn acquisition of Mobileye, which closed today, extended a wave of big-ticket semiconductor deals into 2017, but only barely. Since that transaction’s announcement, only one other $1bn-plus chip purchase has printed, putting 2017 on pace to have the fewest such deals since 2013. There’s little indication that the rate of big semi acquisitions will pick up through the rest of the year.

Toshiba is currently seeking to sell its flash business, which would easily fetch more than $1bn and bring this year’s total of 10-digit purchases to three, leaving it far below recent category totals. Last year’s largest chip transactions – Qualcomm’s $39.2bn reach for NXP Semiconductors (a deal that an activist investor is pushing to renegotiate) and SoftBank’s $32.4bn pickup of ARM – featured two among the 11 companies that fetched more than $1bn. The previous year saw nine such companies get bought.

Two consecutive record years of dealmaking in the category have left behind a dearth of targets for would-be buyers. According to 451 Research’s M&A KnowledgeBase, acquirers spent $116bn on chip vendors last year, breaking the previous year’s record of $90bn – a number that itself was more than double the previous record set in 2006. And since venture capitalists have been absent from the market for a decade, the pool of companies that could command such a price has shrunk notably.

For those potential targets that remain, a run-up in stock prices makes a surge of big deals seem unlikely. The 43% growth in the PHLX Semiconductor index in the past 12 months has outpaced the broader Nasdaq by 20 percentage points. Accelerating stock prices make companies less inclined to launch a sale process or divest large units and the rising multiples that come with a rising stock won’t appeal to buyout shops – the driving force behind this year’s tech M&A market.

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

Qualcomm buys NXP in biggest chip deal ever

Contact: Mark Fontecchio, John Abbott

All the chips get pushed to the middle as Qualcomm agrees to acquire NXP Semiconductors for $39.2bn. The massive deal is the largest ever in semiconductors, and the second-biggest non-telco tech transaction in the past decade, according to 451 Research’s M&A KnowledgeBase. It is also magnitudes larger than Qualcomm’s previous high-water mark, the $3.6bn purchase of Atheros in 2011.

Qualcomm and NXP’s combined portfolio looks unusually synergistic. The buyer is dominant in smartphones and wireless networking, while the target has leading positions in automotive (self-driving cars and in-car entertainment) and the Internet of Things (including mobile payments, security and sensors). Those are two major market segments where Qualcomm can apply its long-standing mobile expertise. A side benefit is that the addition of NXP will increase the number of components Qualcomm can sell to its existing smartphone customers. The recently introduced Snapdragon Neural Processing Engine and Zeroth Machine Intelligence Platform may also help extend NXP’s existing autonomous car activities. NXP has its own manufacturing fabs, while Qualcomm is mostly fabless, but there’s plenty of room for flexibility between the two models due to the range of devices in the catalog.

Today’s deal marks the fifth $15bn+ chip purchase since the start of 2015. NXP is valued at 5.5x trailing revenue. That’s well past the 2.5x median multiple on billion-dollar chip transactions in the past five years and right in the middle of the pack of the recent $15bn+ deals (see chart below).

According to the KnowledgeBase, semiconductor M&A since 2015 has seen four quarters of relatively modest activity ($28bn total) bookended by four quarters – two at the start of 2015, and two at the end of this year – of massive value totaling more than $165bn. The scale of recent consolidation in the semiconductor industry has been astounding. The nearly $90bn in M&A spending we saw last year was more than the previous four years combined. And now, remarkably, value in 2016 has already crested $100bn with two months to go.


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