Thor's Battle Against the Jötnar (1872) by Mårten Eskil Winge (Source: commons.wikimedia.org)
I have seen tempests, when the scolding winds
Have riv'd the knotty oaks, and I have seen
The ambitious ocean swell and rage and foam,
To be exalted with the threat'ning clouds
But never till to-night, never till now,
Did I go through a tempest dropping fire. - Shakespeare, "Julius Caesar"
This morning I awoke to a thunderous Bloomberg headline about the largest "tech" merger in history:
Further headlines poured out during the day from the WSJ, EETimes, Forbes and CNBC:
The embedded videos and articles were alternately amusing and genuinely interesting. For instance: by lumping together companies in the semiconductor sector with the likes of Snapchat and Pinterest under the "tech" banner, the MSM demonstrates a collective intelligence somewhere on the evolutionary scale between amoeba and granite. On the other hand (and to ameliorate the somewhat brutal critique above), there is the suggestion that the size of this and other mergers either announced or under discussion on Wall Street may indicate some sort of M&A frenzy period is underway.
This event bears investigation for what it may portend. Let's start the discussion by examining the two companies side by side in detail.
We've been looking at Broadcom on a quarterly basis for over a year now, but this sort of news suggests an additional review is in order. As a pioneer in SoC, Broadcom grew at a blistering pace during the dotcom boom years of the mid to late 1990's. Lab technicians with stock options became multimillionaires during that heady period, as Broadcom fed a networking systems segment, itself growing in spectacular fashion, with increasingly sophisticated Ethernet chips to help wire the world together and meet up on the internet.
At its inception, Broadcom was extraordinarily daring, innovative and adventurous. The success of the company depended on its determination to continually flout conventional wisdom, which stated that the ASIC business would continue to predominate in Communications because of the overwhelming diversity of custom chip design approaches in the market and the alleged inability of chip houses to engineer system software. With unique sagacity, Broadcom developed chips that captured 80% of the common requirements for a legion of ASICs and left further customization in the hands of customers by providing distributions of software stacks and tools.
Post dotcom bust, the company branched out into consumer electronics, building particular strength in STB and HDTV markets with SoCs incorporating a succession of MIPS and ARM CPUs, multimedia accelerators, embedded SRAM and high performance I/O. Some of this was developed internally and some came from a long string of acquisitions, of which some failed and others did not.
Today, Broadcom is either dominant or ferociously competitive in nearly every segment of the consumer and communications markets. Its products include physical interface and processing/controller chips for both home and enterprise networking, wireless products (including WiFi and Bluetooth), NPUs and multimedia processors. The amount of analog, M/S and digital IP in the company is staggering. Only Qualcomm could legitimately claim to be in the same league of IP richness.
Yet the first half of the 2010's has been a tail of woe for Broadcom. It has retreated from HDTV and BluRay and is walking away as well from cellular baseband, obviating almost $800M of technology acquisitions over the last decade. The company's growth rate has also been anemic for over 4 years running, as the chart below illustrates.
Perhaps this multi-year record of desultory growth reflects revenue lost from Broadcom abandoning the consumer and mobile computing segments mentioned above. Yet that invites another, more worrisome question: why did Broadcom choose to walk away from those markets in the first place? It's a fabless company and thus should have no significant cost disadvantages versus its competitors - particularly any of the vertically integrated ones such as Samsung or the major European semiconductor houses (NXP, STMicro and Infineon.)
It could be that Broadcom is a victim of its own success. It is a very hierarchically organized firm with an extremely aggressive and even hostile work environment. Much of the company's chip development work is governed by strict conformance to industry standard specifications for system applications as well as quick TTM requirements, in particular for its consumer-oriented clients. Such pressures tend to make hierarchies rigid and ossified over time, with departments building virtual sandbags around their portions of the cube farms and department heads driving their employees ruthlessly and relentlessly.
An environment such as this is not one that values or even tolerates personal initiative and creativity. As a consequence, employees will inevitably become taciturn, wary of commitment and innately defensive, doing their utmost to avoid trouble, sticking their necks out or drawing any kind of negative attention on themselves. This increases ossification and decreases communication and sharing, makes the flowering of genuine innovation from internal talent close to impossible and quickly destroys it in acquired firms.
Further complications originate from Broadcom's executive management, which is stained by a history of stock option irregularities that resulted in losing court battles and severely affected earnings. Ex-CEO Henry Nicholas was even indicted by federal prosecutors on narcotics charges (it was alleged he was inviting customer executives to lavish parties replete with prostitutes and drinks spiked with drugs.) The charges were eventually dropped, but memories of damaged reputations tend to linger amongst members of the rank and file - especially in such high pressure companies where so much personal sacrifice is expected and demanded of employees.
This, then, is the firm Avago is interested in buying. From an engineering perspective, it is a company that operates with a certain mechanical efficiency and which will continue to do the best that it is able to enhance and refine its offerings within the constraints of an all-but-dead Moore's Law. It's IP is deep and broad, but has a shelf life. From a managerial point of view it's in distinctly poor shape. Integration is likely to be a struggle and stimulating innovation to take advantage of potential synergies will be inordinately difficult.
A Singapore-based company, Avago is not among the 16 firms selected for the VF quarterly portfolio. This is because Avago is not an industry leader, but a strong niche player. Included by HP in 1999 as part of the Agilent test and medical equipment spinoff, Avago was bought by KKR (the notorious LBO pirate firm that made headlines on Wall Street in the 1980's) and Silver Lake Partners (a combined LBO and VC firm) in 2005 for $2.6B. Like a significant number of other semiconductor companies, Avago turned its back on the high volume, low margin consumer segment and focused its energies on developing highly defensible niches in industrial applications based on specialization in analog, mixed signal and optical chip technologies.
Applications in power conversion, factory automation and motor control are supported by LED, mixed signal, fiber optics and optocoupler semiconductors. This technical strength also translated to success in physical interface and RF products such as SerDes, amplifiers, filters, diodes and motor controllers for a wide range of applications in smartphones, base stations, mass storage and carrier/WAN.
There can be no serious reservations about the technical strengths of Avago. Their SerDes I/O were feared in the ASIC sector and gave ulcers to LSI's and IBM's custom chip groups. There can also be little doubt as to the intelligence and ability of management, as their strategic plan clearly reflected a long term vision and has been remarkably well executed. In a classic 'textbook' approach, Avago focused its efforts on building technical skills and market segment excellence in carefully selected niches that created very steep barriers to entry for competitors. Using these as a secure base, they were able to launch carefully considered expansions into complementary niches. Though revenue was clearly limited (as evidenced by the above chart), this made Avago's niches all the more unattractive to larger competitors and permitted the company a longer window to build up its cash reserves.
This began to pay dividends in 2013, when Avago acquired Cyoptics for $400M and LSI for $6.6B. Cyoptics offered optical components that further reinforced Avago's core business. The LSI purchase, however, was a branch into new technologies and applications. Having transformed itself into a storage chip company, LSI was a good marriage choice, with its digital capabilities complementing Avago's analog, mixed signal and optical strengths in storage. LSI's ASIC business was also in a resurgent phase after years of management neglect, so the SerDes strength of Avago was an excellent addition. Furthermore, a singularly awful LSI management team (originally recruited from Intel and suffering from the acute market myopia typical of internally developed Intel executives) had left the company's finances in a pathetic state, rendering it a rather vulnerable acquisition target.
Avago continued growing its digital technology and applications base in 2014, when it bought PLX Technology (known for its PCIe physical interface products) for $300M. This was followed in early 2015 by a $606M buyout of Emulex, gaining access to the company's Fibre Channel and Ethernet PHY chips and software.
Viewed in a technical light, then, the Broadcom acquisition is both subtly perceptive and strategically brilliant. The combined firm will offer a selection of optical and RF components, physical interfaces, network & application processing and software stacks of a depth and breadth that is completely unrivaled in the industry. The merged enterprise will have, at least in theory, a unified hardware and software solution from the smartphone and tablet RF block to the basestation, thru the backhaul, across metro and LAN into the datacenter.
There are those who might see such a technology juggernaut as a disruptive force across High Technology markets. Qualcomm could easily be among the concerned parties, since in addition to using its mobile computing dominance as leverage to build an early presence in the IoT, the San Diego titan has its own plans for networking and datacenter segments:
On The Scales
"Judicis officium est ut res ita tempora rerum Quærere."
The judge's duty is to inquire about the time, as well as the facts. - Ovid
Though we are not ancient Egyptian dieties, we must also divine the meaning of this M&A event and ascertain its meaning, both for the companies involved and what it foretells of the future. Will this merger work, and why is it happening?
Mergers are always difficult at the level of middle management and the rank & file. I can bear witness directly to LSI's acquisition of Symbios, an excellent company with tremendous strength in analog and mixed signal design for storage applications. Despite the tremendous business and technical advantages of the merger, it took a full year for the dust to settle and there remained a residual tension between the geographically separated cultures of the two companies (with LSI based in California and Symbios in Colorado.)
Avago should be no stranger to this. The LSI acquisition in 2013 took a good 18 months to sort out organizationally, with advantage and momentum swinging back and forth between the two camps several times. Many work cycles were wasted on politicking and maneuvering instead of business and growth.
The Avago-Broadcom merger promises to be much more difficult. Avago is half the size of its acquisition target. Furthermore, stimulating and managing organizational change will be especially hard in this case, since broadcom's managerial principles embody the antithesis of change. The unified company will be called Broadcom Ltd., but I strongly doubt Broadcom's management hierarchy will wind up being imposed upon Avago's departments. The internal struggle promises to be a bitter one, regardless.
There may be a hidden message in the purchaser being only half the size of the target. Perhaps Intel's stealthy attack on Broadcom's STB business (discussed at length in earlier quarterly 'State of the Union' reviews) reflects an unspoken perception in executive suites across the semiconductor industry that, of all the SoC companies, Broadcom is particularly vulnerable and may be least able to fight serious challenges - a 'paper tiger' of the semiconductor sector. There can be no question that Broadcom's strategic stance over the last 4 years has been one of retreat and defense, as shown by its abandonment of the business units discussed earlier.
What will this acquisition do to the financial profile of the merged companies, and how will this negatively affect future applied R&D? The purchase is structured as a stock-cash deal with $17B cash and $20B stock. Where is the cash coming from? Allegedly $8B is from Avago, and another $9B from a 'consortium of banks.' Yet per the 2014 Avago annual report, the company only has $1.6B of cash on hand - and this is before the Emulex purchase. Which banks? One would imagine Silver Lake will be involved because of their long-standing relationship with Avago, but I doubt they'll cough up the entire $9B on their own. Finally - where is the $20B in stock to come from? Again per the 2014 annual report, Avago has shareholder equity on hand of only $3.2B.
One would imagine all the details had already been worked out by Avago's finance department. Per the CNBC link at the beginning of the editorial, Avago has offered a general description of the financing of the deal:
"Avago said the combined company will have US$14.2 billion of net debt, and leverage of around 2.5x EBITDA. It will have US$15.5 billion of new term loans at closing, including US$6.5 billion to refinance existing debt facilities and US$9 billion of new debt.
Avago is rated BB+ by both S&P and Fitch. Broadcom is rated A2 by Moody's and A- by S&P."
EBITDA stands for "earnings before interest, taxes, depreciation and amortization." The 2.5x leverage is, from what I understand, considered to be somewhat high but not alarmingly so. Nonetheless, the consolidated firm will be carrying sizable debts and its credit rating will be poor.
There is some speculation that Avago plans to shed a number of business units in the same manner as the LSI acquisition, where the networking group was sold to Intel and the Flash business to Seagate. A few analysts are predicting that Broadcom's consumer oriented businesses will be disposed of in like manner.
NXP seems to have already decided that selling off pieces of its own business to pay for M&A activity is a viable route both tactically and strategically. The european giant is selling its power amplifier unit (and losing some of its discrete component play in mobile computing infrastructure) for $1.8B to help finance its $11.8B buyout of Freescale:
However, selling off pieces of Broadcom could be an awfully stupid move in the long term. When one establishes a dominant position in a low margin segment, that segment has both technical and financial barriers to entry that turn it into something of a cash cow for the corporation. To a significant extent, Avago put itself in a position to execute its serial acquisitions by following such a strategy. Furthermore, if the company plans to play any significant role in the emerging IoT, Robotics and AI markets, it cannot walk away from the multimedia IP, skillsets and market knowledge in its consumer product departments.
Satire on Tulip Mania Oil on Canvas, Jan Breughel the Younger, ca 1640 (source: artwis.com)
The Dutch tulip craze is an oft-cited tale of bubble-induced financial fraud from the past. As we've discussed before, the current NASDAQ and S&P 500 record highs, driven by artificially suppressed interest rates and executives using stock buyback programs financed by cheap debt to commit acts of financial chicanery against their own companies draws some legitimate comparisons with the tragically comical events of Holland in the early 17th century.
We're seeing a rather dramatic rise of LBO and merger activity across the private sector, and clearly High Tech is participating. Cypress bought Spansion in December for $4B. NXP is buying Freescale for $11.8B. Intel is allegedly still talking to Altera about a buyout, now priced at $16B. There have also been whispers of Renesas and Xilinx as potential acquisition targets.
KKR, one of the original Avago investors in 2006, is better known for its piratical LBO activity on Wall Street during the 1980's. It would not surprise me in the least if KKR became one of the members of the 'banking consortium' helping to finance Avago's acquisition of Broadcom.
I think we are about to see a wave of M&A activity across High Tech driven by Wall Street's brokers and banks, cashing in on the current equity bubble in the spirit of "get while the gettin's good." This also suggests that the merger wave is a last hurrah before things become more difficult both in High Tech and across the broader global economy.
The SIA/WSTS is scheduled to update its 2015 forecast on June 2nd. At the start of the year, they prognosticated 3.5% growth for the semiconductor industry. Several weeks later, after seeing everyone else forecasting significantly higher growth, the SIA/WSTS folded under pressure and pushed their forecast up to 4.9%. I very strongly suspect that they will ratchet it back down on June 2nd, as macroeconomic indicators across the world indicate that another recession is in the offing. As an example: the BEA (Bureau of Economic Analysis) today released a new estimate of America's Q1 2015 GDP. Originally tallied at 0.2% growth, it has been revised down to -0.7%.
Frankly, if semiconductors can achieve half of the SIA's growth projection, I think we'll have done pretty well. We shall have to wait and see how events unfold. ;-)