Next to Apple, Intel’s TV launch has been the most talked about digital media event that has yet to happen, with considerable similarities. Both know that the device itself must stand out from the competition and that content will be crucial for success. Of course while Apple already has its TV service providing content via a set top box, Intel has so far focused on the silicon side of TV. But in its desperation to atone for declining PC sales, compounded by being too late to develop a viable strategy for mobile and losing out big time to ARM in the process, Intel has decided it must gamble on a TV service itself. It certainly looks like being a play for high stakes given that Intel has decided to build its strategy around premium content. It is doing this to a greater extent than major competitors such as Apple and Google, but is also coming up against Amazon and Netflix, which are both rapidly expanding their investment in original content.
The problem for Intel is the classic market squeeze where a growing number of distributors are chasing an essentially static resource of premium content, whose price therefore continues to escalate, as we are witnessing with the inflation in the cost of both leading sports and blockbuster movie assets. This is exacerbated by the reluctance of existing rights holders to unlock their bundles that have proved lucrative for a long time, unless they obtain a price they cannot refuse. So while Intel is sitting on a cash mountain of $25 billion, which is still growing having expanded by $3.41 billion in 2012, TV premium content is a bottomless pit. The $2 billion it is said to be ready to spend now would merely buy just over a year’s rights to one country’s premier soccer.
We do know that Intel has been locked in negotiation with rights holders all year and is serious about amassing a comprehensive content package from most of the leading US players before unveiling its TV service. Intel wanted to have full control over a wide range of content in order to achieve one of its presumed differentiators, which is a wide range of bundles including smaller ones so that customers can specify precisely the channels they want. But it looks like the providers are holding out for a premium of at least 50% and up to 75% above what traditional operators pay, partly because Intel wants more flexible bundling, but mainly because it is a Greenfield as far as TV concerned. It therefore represents a bigger risk for rights holders because prospects of ad revenue are less certain. While details are sketchy it seems that Intel has got at least close to deals with CBS, News Corp, and Viacom, while being at earlier stages with NBC Universal.
Apart from paying much more than it had hoped, not helped by the knowledge of its cash reserves, Intel has had to make some concessions such as agreeing to block fast-forwarding through commercials during the first run of any show. Intel has been trading in the technological features it can bring to the table to boost ad revenue and deliver effective targeting for both ads and recommendations.
Its big idea here is itself controversial, to deploy a camera with in-built facial recognition technology in the set top box, not to invade privacy directly but to assess the age and gender of people viewing to get round the targeting problem of not knowing who exactly among a family and friends are watching at a given time. The theory is that even a rough and only partially accurate assessment of who is watching is a lot better than mere knowledge of the household demographic on its own.
The two can be combined, also taking account of the content being consumed at the time, to determine which ads and recommendations to target. It is not hard to conceive of even more contentious extensions to this, such as audio recognition to analyze the actual conversations being had around the box.
But with the camera seemingly failing to convince rights holders, Intel may well decide, like Netflix and Amazon, that the best way to deal with content owners is to join their ranks and get into original production. In a way this is the logical culmination of Intel’s unfolding TV strategy. Yet but for its miscalculation over mobile, which effectively cost the job of recently departed CEO Paul Otellini, it would have been in a much more relaxed and assertive position now over TV.
The roots of the problem lie in earlier unbridled success, in Intel’s case with the x86 architecture that drove the PC and server markets through the 1980s and 1990s and where its core IP (Intellectual Property) resided. This diverted Intel’s eye from the ball when RISC (Reduced Instruction Set Computing) architecture emerged in the mid-1980s, at that time adopted in preference to x86 for top end graphical workstations pioneered by Sun Microsystems, achieving higher speed by simplifying the chip design at the expense of requiring more complex software to exploit it effectively. The advantage for high end workstations of the time was that the greater simplicity reduced the number of transistors required for the core logic on the chip, freeing up space for an increased register set and internal parallelism.
RISC then enjoyed a greater reincarnation in embedded systems, above all in mobile devices under ARM, where the lower transistor count scored primarily by reducing power consumed to execute a given set of tasks, again providing the software was properly written. But Intel failed to learn its earlier RISC lesson and missed the boat again, still being wedded to the x86.
Intel clung on to the belief that x86 could be adapted for the mobile market while retaining its advantages as a programming platform with reduced time to market for software apps. Under Otellini’s reign it actually gave up its attempts to get into RISC and sold its XScale ARM family to Marvell. This led to development of the Atom x86 family for netbook PCs and other smaller portable devices and ultimately showed that x86 could match ARM for energy efficiency. But this looks like being far too late as ARM is now well established as the platform of choice for mobile computing, having been adopted by Apple for both the iPhone and iPad as well as nearly all Android devices, Nintendo and many others. Apart from market presence this has eroded the original argument for x86 of compatibility with the vast library of Windows software written for the platform.
So the only significant x86 presence in the mobile space outside PCs is on the Microsoft Surface Tablet and even that has an ARM version as well. Microsoft is hedging its bets, but Intel does not have that luxury.