Immersion search and browsing

The landscape is littered with dead bodies, but is it not time for someone to take another shot at a richer search and browsing experience?  Exploration of new topics via search (movies, music, art, science, travel, sports, and even shopping) is often as much about the journey as the destination.  I'd love to be able to navigate a visual array like the Safari Top Sites display that would tile Wiki's, stack videos, and other pertinent sites for me to rapidly explore a single search term.  Google has emphasized speed and simplicity (I suggest that you read In The Plexif you have not already done so) and delivered a highly efficient but sterile solution.

Instead of trying to compete with Google in search (SeachMe, Viewzi, Grokker failed for many reasons), possibly the immersion experience comes from an application, maybe a browser plug-in, or an app like FlipBoard, that leverages the power of Google search API and the steadily ramping network and processor bandwidth.  Search-Cube heads in the right direction, but it is ugly and inefficient; wouldn't you always want the Wiki page, video stack, etc. in the same spot on the screen every time so you wouldn't need to visually decipher or manipulate the presentation?

Apple does a fine job at user interfaces and could leverage the building blocks inherent in Safari, but I cannot imagine them investing in a utility that has Google search as its underlying technology.  Beyond a brilliant 2D search environment, I look forward to the day we have a true immersion search, browsing and computing interface.  Take a good look at Oblong Industries if you want a glimpse of the future.

Thoughts on the Freescale IPO

Bloomberg Businessweek published a solid summary of this week's Freescale IPO, but there is more to the story.  I have great respect for Rich Beyer leaving the comfort of Intersil and walking into the burning building in 2008, albeit with a outstanding compensation package averaging $5M cash per year plus equity, to rescue Freescale on behalf of their private equity investors.  The industry dominating business practices, technology innovation and quality that was Motorola Semiconductor of the 1980's died a horrible death in the 1990's and spawned On Semiconductor and Freescale in 1999 and 2004 respectively.  Both companies have had to navigate the semiconductor industry cycles with suffocating debt, and in the case of Freescale, fleeing talent, fleeing customers and collapsing markets. I've actively recruited engineers and business leaders away from Freescale over the past decade, but the well is nearly dry.  Those who are left acknowledge that they may be out of a job any day due to the relentless cutting, but they are generally unemployable elsewhere.


There is money to be made in all segments of the semiconductor industry, both as an employee and shareholder, but the Freescale story will continue to be one of gains driven by business and debt restructuring not from innovation or market disruption.


Good luck Rich!

The Smell of Success

Over the past year I’ve been helping one of my clients build their engineering team and executive staff. One of their projects for a new market and new customers required a new engineering team, new software tools, new manufacturing technology, new technology partners, etc…….often a recipe for “Fail Fast” lessons. We went from zero to sampling a revolutionary new product in 15 months and I attribute this success to the new team’s leader, the great engineers he was fortunate to hire and the company’s culture.

Venture capital and entrepreneurial communities all sing the familiar song that says that failure is OK….it’s a learning experience…..that which does not kill you will make you stronger…etc. I completely agree that mistakes and failures are powerful learning opportunities if your ego allows you to be introspective, but don't not lose sight of the fact that you shouldn’t be rewarded for failure. You may collect a salary (and severance) and be able to raise future capital, but you haven’t done your job.

I have written previously about “What do I look for when building a team?” and would like to add one more quality:

“Knows what success smells like.”

You must build leadership teams with people who have known great technical and/or business success. There is an intangible element of decision-making, risk analysis and leadership that comes from knowing what success smells like and how to find it.  I’ve likely reviewed 10,000 resumes in my career and hired about 250 engineers and executives. If you have been the industry for 10 years and haven’t contributed a major technical or business success, you’re not in the running for my organization. If you think your failures can be blamed on a succession of companies collapsing around you, maybe you chose the wrong companies in which to build your career.


Fail Fast is a very valuable concept because it allows you to move quickly onto your next Success.

Understanding the Semiconductor IP Companies

ARM, MIPS, Tessera, Rambus, InterDigital, Patriot Scientific and Mosaid (Toronto) derive a majority of their revenue from licensing intellectual property (IP) to semiconductor manufacturers. IP revenue is also a significant fraction of the revenue stream for Qualcomm (33%), Silicon Image (20%) Synopsys (13%), Cadence and Mentor Graphics.

Think of their business model as software development, sales and support with no exposure to the supply chain and yield headaches that periodically strike semiconductor manufactures and shock their shareholders.

The customers of these companies often face make-versus-buy decisions….do I invest time and money to design my own widget, or do I buy the building block from an IP vendor? Both options present different technical and financial risks and the IP companies work to compel you, with both stick and carrot, to buy! 

Companies like ARM and MIPS invest heavily in R&D and deliver critical new technologies for highly competitive consumer and communications infrastructure markets, some (Mosaid, Patriot Scientific, InterDigital) are exploiting patent portfolios through licensing and litigation, and others (Tessera and Rambus) are a mixture of both models.

Quick comments on a few of these companies:

ARM Holdings (ARMH) – I’m very bullish about ARM’s continuing global technology contributions and revenue growth for embedded microprocessors. They are a well respected vendor, highly responsive to the niche performance demands of consumer products (smart phones, tablets, automotive, etc.) and may make inroads against giant Intel in the cloud computing server market.

MIPS (MIPS) – While they have failed to take significant market share in the highest volume consumer markets, MIPS’ processor designs are utilized by Broadcom, Cavium Networks, Cisco, Entropic, Juniper, Netlogic Microsystems, Renesas Electronics, Sony, and Toshiba in set top boxes (cable and satellite), video games and network communications. They lost focus on the microprocessor business with their 2007 acquisition of analog IP vendor Chipidea that was divested in 2009. MIPS will likely exist as a runner-up to ARM until a customer or competitor acquires them.

Tessera (TSRA) – Licenses their portfolio of semiconductor packaging and imaging technology patents. Their continued revenue growth requires them to acquire and author new patents to replace portfolio patents that expire or are determined to be invalid by the courts or government patent offices. Companies like Tessera, Rambus and Patriot Scientific are not so affectionately referred to as patent trolls in the industry. They don’t have much in the way of a sales force, but they are expert intimidators and litigators.

Synopsys (SNPS), Cadence (CDNS) and Mentor Graphics (MENT) all supply software design tools to the semiconductor industry and added IP design and licensing to their portfolio of products.

Intel - No longer ideas, just wasted tweets

My friend Howard Lindzon's Twitter steam is genuinely his stream of consciousness.  He's usually funny and occasionally profane, and there are daily messages of wisdom.  One of todays tweets was:

Nice, a joke and some wisdom.

In an earlier post I made the case that Intel should buy ARM to maintain a dominant position in the microprocessor business. I have little doubt that we will unknowingly buy Intel processors buried in consumer electronics for decades to come and that they will maintain acceptable margins and moderate growth......but the story ends there.

Indeed....no longer ideas, just wasted Tweets.

Warning about warnings!

Given Japan's March 11th earthquake and tsunami's likely impact on the semiconductor and consumer electronics supply chain, I'm truly puzzled by how few revenue warnings have been released.  We just haven't heard much from the cell phone, PC and tablet companies........and their suppliers.

For component suppliers, this is NOT just a supply side issue.

Here is one of the Rumsfeldian "know unknowns" that haunts the consumer electronics supply chain.  Your semiconductor company may have escaped a disaster unharmed, but if there is supply disruption of ANY component (capacitors, resistors, inductors, specialty adhesives, packaging resins, flex connectors, LCD displays, etc.) for your customer's product (PC, cell phone, TV, etc.), production stops and your sales are slammed too!

In good times, semiconductor suppliers have horrible visibility into their sales channel. They typically ship into a logistics and distribution channel which provides supply to an ODM (Foxconn Compal, etc.), who in-turn supplies the name brand electronics company (Dell, HP, Nokia, etc.), who in-turn supplies retailers who sell to real consumers. When something goes horribly wrong, it can take weeks for inventories to back-up into the semiconductor supply lines and get then you get hit with order de-bookings.

It's not obvious if these risks are being priced into the market's recent retrenchment. As we move into the semiconductor industry quarterly earnings calls and forecasts...you have been warned!

Initial thoughts on Texas Instruments acquisition of National Semiconductor

While I cannot explain the 78% premium that Texas Instruments has agreed to pay for National Semiconductor, we can reasonably assume that they knew what it would take to get the deal done. As recently as January, dissident shareholder Ralph Whitworth and Relational Investors held 20.6M National shares (8.61% of outstanding) at basis of $16.29. The fundamental value in National Semiconductor is their position in Analog integrated circuits which are the highest margin (>60% GM) and longest lifetime (often > 20 years) products in the semiconductor industry. The market leaders are shown in a recent Analog Devices SEC 8K filing.

The hindsight is easy.  ADI, Maxim and Linear Technology closed today a P/E premium 30-300% compared to National.  Over the past eight years, National refocused on margins over top-line growth and shed many their less profitable endeavors.  TI and National have highly compatible business and technology cultures, where the second generation analog companies like Maxim and Linear could be very difficult to integrate.

I'm a bit of a broken record on the point, but the semiconductor industry is now 50 years old and consolidation will be a strategic play throughout the supply chain, from wafer fabrication all the way through distribution and logistics.  In the analog space, I own Intersil due to their technology, team and chosen markets and have long advocated that someone buy Micrel and tug that company out of the 1980's and into the new millennium.  I don't bet on M&A plays, but for the under appreciated analog companies wallowing at P/E's below 15, this news will be a welcomed boost!  

Why Intel must buy ARM

Last week Intel, the largest and most profitable semiconductor manufacturer, announced record sales and profit, but their performance by product segment combined with cataclysmic shifts in both retail and enterprise computing markets foretell Intel's growth to be in serious jeopardy if they do not acquire ARM.

Intel’s legacy PC processor business will never again be a growth engine. Consumers have an insatiable demand for electronics that enhance their experience in social networking, personal communications, portable gaming, electronic media (music, video, books, etc.), shopping, and financial management. In these markets the PC is losing badly to smart phones and tablets where ARM based processors dominate.  The decline in the PC's value (processor speed and data storage) will also be fueled by commercial customers moving their business management solutions to the cloud

The Data Center Group, which provides technology for enterprise servers and cloud computing, is Intel’s only promising growth business and delivered 23% of the top line revenue and 28% of the operating profit last quarter. But industry powerhouses like Nvidia, Marvell, IBM and start-ups like Calxeda (backed by TI, ARM, and the government of Abu Dhabi which also owns two of the largest semiconductor wafer suppliers) are preparing to launch very high performance ARM based products targeted to eclipse Intel in servers. There is little doubt that Intel will attempt to leverage supply and pricing of PC processors to retain server business with companies like Dell and HP, but they will have little sway over cloud computing companies like Google who develop their own server solutions. 

In the past 30 years, Intel has only done one thing well; microprocessors. If they are to maintain their dominance in the microprocessor business and grow shareholder value, they must acquire ARM while they can still afford the valuation. Too bad they have recently inked deals to acquire McAfee for $7.7B (strategically puzzling) and Infineon’s Wireless business for $1.4B (returning to a market where they have repeatedly failed). Capital poorly invested.