Pierr Johnson's  Instablog

Much of great interest to Tech investors has coursed through the news in recent days. First, the earnings releases of Adobe and Oracle cast light from two directions on the adoption of Cloud Computing broadly and, more specifically, software demand. Tibco, set to release later today, waits in the wings. At yet another leading edge, Nvidia used its Graphics Technology Conference to articulate not only its Graphics leadership, but the opportunity it has to leverage its technology in Mobile, Enterprise and Technical Computing. Finally, the behemoths HP and Dell are advancing distinct responses to the significant shift wrought in personal computing by the rise of Tablets and Smartphones, even as we acknowledge the negative impact on the PC market is intensifying. And in the noise that accompanies this shift, the silence of Intel on its search to replace Paul Otellini is deafening.

Having dramatically missed aggressive guidance with last night's earnings release, Oracle sees its stock down 5-10% today. I have yet to dig into the call and have only sampled the chatter. I am eager to get a better read on the size and growth rate of Oracle's Cloud offering. Adobe's results were better than expected, fueled by the rapid adoption of its Creative Cloud offering. Subscriptions grew 153,000, well above analysts' expectations.

Regarding another aspect of Oracle's results: that hardware is down 23% is striking. Having lost its suit against HP over Itanium support last summer, Oracle should approach the damages review with trepidation and even contrition. The company brazenly displayed baser instincts using Itanium support as a cudgel to promote its Sun server business; the damage to HP's Enterprise server business has been significant. I will readily acknowledge that the damage was no doubt intensified by the fact that sufficient capacity exists in high-end, deep computing. Yes, weak demand is broadly affecting this market. That said, the applications that require high-end servers are not going away, they just aren't growing. The market remains and the economic damage to HP's business must be accounted for. Note that HP's consulting economist has sized the impact at $4.0 to $4.5 billion.

Of course, growth in computing is now squarely centered in the Internet, Big Data and the Cloud, each of which continues to redefine our individual relationships with computing. Much of the data we generate is aggregated over the web, whether it concern our online shopping preferences, our work research, or our social interactions. Key then are the hardware and software solutions that allow companies to derive economic value from the analysis of this data.

In this regard, Nvidia's GTC was particularly interesting for some of the new directions it is identifying in its offering. The roadmaps it offered for its Graphics and Application Processors showed great promise, assuming it meets some significant technology challenges as well. (Feel free to write with questions.) Three additional press releases suggest the company has strong moves planned in new markets as well, though they trace an origin to the technologies Nvidia has developed around the use of its GPUs as accelerators and co-processors in High Performance (i.e., Super) Computing.

In the first of these releases, Nvidia has announced that its GPUs are being used in Big Data Analytics to more efficiently analyze especially large data sets. Shazam and Cortexica are using GPU Accelerators to track song and image consumption, while Salesforce.com is using GPU CUDA to track Tweets. Less clear, though clearly of interest, are the platforms used for this purpose. In a similar vein, Nvidia announced the adoption of its GRID technology by leading Technology vendors, including (in hardware) HP, Dell and IBM, and (in software) Citrix, Microsoft, and VMWare. I find this intriguing and am eager to learn more about the potential for this apparently new platform. Perhaps telling is the third announcement on Nvdia's establishment of a Visual Computing Alliance that is based on a 4U appliance with 16 GPUs that runs Adobe, Dessault and Autodesk applications supporting up to 16 concurrent users. This appliance could be disruptive to the technical computing market.

In every way, these announcements are very ambitious. And they will require additional analysis as time permits.

And regarding time, the bell is ringing here and it is time to move from this blog to my other work. My comment on HP and Dell will have to wait, regrettably. I welcome questions, comments and insights.

(posted as a comment in response to Rolf Winkler's "Intel's Higher Spending May Not Compute," WSJ 01-18-13

Intel has always had a sizable contingent of detractors in the investment community. And with the dramatic changes underway in the consumer PC market, they are better positioned to gloat. That said, to judge from the commentary in the recent earnings call, management seems to have a clearer grasp of the issues involved than its skeptics. Capital investment next year will be essentially flat, once you back out the $2 billion being directed at building a 450 mm R&D facility that will be equipped and utilized years hence. Moreover, they must maintain their substantial competitive edge in process technology just to sustain the business model against the advance of the ARM-centric alternative ecosystem. The suggestion that they spend less would not only halt Moore's Law for an interval, it would expose them to significant market share loss. Finally, due to the laws of physics, if they sustain their lead as they progress down the roadmap, they will automatically overtake ARM's lead in power efficiency while maintaining their 18 to 24 month lead in manufacturing technology. So the suggestion that the company not invest seems unwise, to say the least.

One final comment: while many seem to see a market where the PC is dead and Tablets prevail, Intel is describing a market where a profusion of form factors will arise this year in the PC market, ranging from classic notebooks, through hybrids, convertibles and Windows 8 Pro Tablets and onto ARM-based Windows RT, Android and iPad Tablets. Clearly, Intel will need to supply what demand arises from this market in coming years. But based on how the market has developed in the iPad and Android segment alone, things will likely be very dynamic in the years ahead. Given where Apple's share price is these days, investors are clearly skeptical about their outcome as well

Additional Comments Posted in my Seeking Alpha Instablog:

The comments of Deutsche Bank Analyst Ross Seymore figure prominently in both the article of Winkler and the Q&A of the Intel Earnings Release conference call. In the latter, Ross noted that R&D and capital expense growth had exceeded revenue growth in recent quarters and that some investors were dubious about the likelihood of adequate returns. And in the former Ross noted that ROIC was likely to decline from a recent cyclical high of ~25%.

The trends Ross cites are an unavoidable reality at this juncture. Macroeconomic factors and strong tablet sales have reversed the growth of the PC industry just as major technology transitions loom that require significant investment. Yet technology advances in the semiconductor industry are defined by the Technology Roadmap that arises out of Moore's Law-the notion that the density of transistors on a silicon wafer can double nearly every 18 months. And due to its virtuous character, the cycle requires investment to sustain progress, whatever given industry trends obtain. Yes, invest we must, in good times or bad. The proceeds of this investment will only be harvested as the industry improves, over an as yet unknown interval of time.

It is indeed both a capital and technology intensive business. The transition to FinFET transistors, EUV and 450 nm wafers are surely drivers of higher R&D spending. And the transition from 32 nm and 28 nm to 22 nm and 14 nm process technology is surely driving capital spending higher. Higher spending is thus being driven by technological advancement and the need to migrate volume production to more advanced manufacturing nodes ahead of competing manufacturers.

And indeed, Intel marches on: management indicated that over half of volume production in the fourth quarter was at 22 nm and that "maintenance" capital spending was target at finishing up the 14 nm factories. The extent of Intel's technology lead is visible in the fact that TSMC spent much of 2012 ramping its 28 nm capacity to meet the volume needs of Qualcomm and other IDM working at the most advanced nodes (behind Intel). That said, TSMC, Samsung and handful of other manufacturers will maintain this lagging interval with their own capital and R&D spending, as they advance their own production down the roadmap as well.

Today a flurry of news descends regarding Intel and its command of core and emerging markets.  Our colleague in  Seeking Alpha (David Zeiler) writes of the promise offered in Intel’s Light Peak data transfer technology, which it hopes will ultimately supplant USB x.0 as the primary PC-centric connect technology.  And both he and others comment of Intel’s newly energized rivalry with ARM and its partners in the Tablet and Smart Phone markets.

Indeed, Light Peak, which offers initial transmission rates twenty-five times the existing USB rate, is today being re-branded “Thunderbolt” with its release by Intel and Apple, which is going to include it in its new line of MacBook Pro laptops.  With a rate of 10 Gbps, it permits the transfer of a full-length HD movie in about 30 seconds, surely a benefit to the power-users’ movie viewing.  And it speaks to Intel’s command of the full reach of technologies that enable PC- and Server-centric computing as testimony to its prowess.

It has always been a challenge to Intel to extend this leadership in semiconductor design and process capabilities beyond PC-centric computing. And this too is a theme of the day. With last week’s Mobile World Congress still fresh in our minds, we see Intel’s relative absence from burgeoning markets for the Smart Phone and super-sized counterpart, the Tablet. For in this exciting extension of the mobile phone market, ARM-based processors of Broadcom, TI, Qualcomm and others dominate. And Johnny-Come-Lately Intel is seen as an outcast, at least for now.  In fact, because Tablets are displacing Netbooks, the momentum of Intel’s Atom processor line has indeed diminished. Further muddying the waters was Nokia’s selection of Windows Phone 7 for future Smart Phone offering, which appears to strand its alliance with Intel around Meego.

What is this woe unto Intel, one might ask? In my view, it is early skirmishing at the very periphery of Intel’s PC-centric universe. Surely it is Apple’s iPad this is driving the Tablet market and its key attribute is that its design is so perfectly mated to its function.  To fully leverage the power efficiency of this new platform, Apple selected and application specific ARM-based processor not too far removed from that in its iPhone. For it is not truly a computer per se, but an Internet access device that can run custom apps and access the network via a Wireless LAN (for the in-home user) or a Wireless Carrier (for the mobile user). It is not really a computer at all, but an access device.

So there it is. In this game we are in the earliest inning and the teams appear to be aligned along their strongest axes.  Still, situated as it is between the mountain-like realms of the PC and Handset, this exciting new market for Mobile devices is likely to admit a variety of capabilities and variants. And we can be sure Intel will be there with a variety of offerings as it develops.