The message has trickled down in speeches, earnings calls, and analyst presentations, but on Tuesday, Intel chief executive Brian Krzanich drew a line in the sand: Intel is not a PC company any more.
In what only can be called a manifesto of Intel’s new values, Krzanich described how Intel is transforming itself “from a PC company to a company that powers the cloud and billions of smart, connected computing devices.” To drive the point home, Krzanich noted that the PC is just one among many connected devices.
What might be called the “new” Intel will be built upon five pillars, Krzanich said:
- The Cloud—including servers, datacentre, and virtualisation
- Connected “things,” such as sensors, autonomous vehicles, or PCs
- An evolving memory business, from 3D XPoint memory to advances in server and datacentre infrastructure
- Connectivity, specifically 5G networking
- Manufacturing and the underlying fab technology.
About 40 per cent of Intel’s revenue and 60 per cent of its profit margin already come from outside the PC, Krzanich said last week, when the company began publicly signalling its new focus. “It’s time to make this transition and to push the company over all the way to that strategy and that strategic direction,” Krzanich said then. “That’s why we wanted to do it now.”
Cloud first - wait, we’ve heard this before
Historically, Intel has been built on a single foundation: the microprocessor, which powered the majority of the world’s PCs, then servers, then notebooks. Now, Intel’s evolution looks surprisingly like that of Microsoft: predicated upon the cloud and potentially billions of connected, mobile devices, with a broad, diversified product line to address a multitude of opportunities.
“We will also lead by becoming a company with a broader focus, and with sharper execution,” Krzanich wrote Tuesday. “ In doing so, we will create lasting value for our customers, partners and shareholders, and achieve our mission to lead in a smart, connected world.”
You had only to attend Intel’s Intel Developer Forum conference last August to see this coming: Krzanich barely mentioned the company’s Skylake PC processor, focusing on Intel’s push in the Internet of Things instead. Ditto for Krzanich’s CES 2016 keynote address, where he played ringmaster to a circus of devices showing off Intel’s embedded silicon. Finally, Krzanich reorganized Intel shortly before announcing its first-quarter earnings, where he divulged that each of its PC and IoT projects were being evaluated for possible cancellation
Now, like Microsoft, Intel sees the cloud as the driver of Intel’s business. The company can charge thousands of dollars for a Xeon processor that powers a server, but just a fraction of that for a standard Core chip.
Krzanich wrote that Intel plans to attack the data center on two key fronts: virtualization, which creates demand for pricey, high-end chips by using them to power many “virtual PCs” in the cloud; and analytics, which takes all the data being collected by the cloud from sensors and other devices, and extracts information from it. The latter capability, of course, requires even more server hardware.
Krzanich also pledged to drive “more and more of the footprint of the data center to Intel architecture.” Try as it might, AMD has continued to lose share in the enterprise—though it recently tried to steal some back with a licensing effort.
Putting the PC in its place
For months now, Intel executives have offered variations on the same line: “Everything’s connected, and everything that’s connected, computes.” Intel plans to lead in the technology around connected things, Kzanich pledged.
“’Things’ range from PCs to what we now call the Internet of Things,” Krzanich explained. “The Internet of Things encompasses all smart devices—every device, sensor, console and any other client device—that are connected to the cloud. The key phrase here is ‘connected to the cloud.’ It means that everything that a ‘thing’ does can be captured as a piece of data, measured real-time, and is accessible from anywhere.”
In Intel’s world, devices are simply means of producing (not consuming!) data. And the amount of data you produce, tapping away at your computer, is potentially minuscule compared to the data sampled by the LiDAR sensor of a self-driving car. That’s important to realize: At one time, PCs demanded so much data that they could bog down the available network bandwidth and compute power of a client device. Today, machines talking to machines generate those workloads.
“At Intel, we will focus on autonomous vehicles, industrial and retail as our primary growth drivers of the Internet of Things,” Krzanich wrote. “Similarly, we view our core client business of PCs and mobile as among the many variations of connected things, which is driving our strategy of differentiation and segmentation in the Internet of Things business.”
Intel doubles down on 5G connectivity
What’s unclear about that strategy is whether Intel intends to compete with ARM in powering smartphone processors. The recent, quiet departure of a key embedded exec, Aicha Evans, indicates that Intel has failed for the moment. But Krzanich and Intel clearly intend to double down on their investments in 5G connectivity, as Krzanich pledged that Intel would lead in this new sector.
“Threading all of this virtuous cycle together is connectivity – the fact that providing computing power to a device and connecting it to the cloud makes it more valuable,” Krzanich wrote.
It’s worth noting that Intel has invested heavily before in next-generation wireless solutions that failed miserably—WiMAX, anyone? This time around, however, Intel apparently is pursuing a more conventional route.
The wild cards: memory, programmable solutions
Krzanich also highlighted how recent innovations would drive the company's future. 3D XPoint or Optane memory technology could be game-changing: far faster than SSDs, Intel sees it as a possibility for either memory or storage within high-performance PCs and servers. Intel’s other moonshots include silicon photonics, which swaps electricity for light in connecting chips and boards; and integrated FPGAs and conventional silicon, which offer the possibility of actually reprogramming the chip to make it more efficient or perform specialized tasks.
All of these represent a luxury Intel has: By consistently pulling in profits using its current technology, Intel can continue to fund its moonshot innovations. Over time, Intel hopes, they’ll become the next big thing.
The foundation for it all: manufacturing
Intel, of course, has become synonymous with Moore’s Law, the axiom that semiconductor density will double about every two years. Intel grew to dominate its industry by pretty consistently hitting that target. By alternating manufacturing advances with processor advances, Intel has been able to offer a choice: Do you want higher performance from its processors, lower power, or something in between?
As far as device performance goes, however, that’s reaching a wall. Now, PC performance is often measured more by how many computing cores are within them then by the speed they run at. With the shift toward mobile devices, customers are demanding phones and tablets that remain powered on for longer. And you might think that’s what Intel’s building upon for the next generation.
Not quite. “Moore’s Law is fundamentally a law of economics, and Intel will confidently continue to harness its value,” Krzanich wrote. “The law says that we can shrink transistor dimensions by roughly 50% at a roughly fixed cost, thus driving twice the transistors for the same cost (or the same number of transistors for half the cost).”
What Krzanich is saying is that Intel believes it can take the same processor and make it physically smaller, but for the same manufacturing cost. Alternatively, it could take the same chip and make it more powerful, by doubling the transistor count. Finally, Krzanich is saying that Intel could also choose to manufacture the same chip at a significantly lower price point.
This dovetails nicely with Intel’s newly articulated strategy of embedding computational capabilities in as many devices as possible, shrinking down processors to a point where you’ll simply expect sensors and wearables to embed computational capabilities. Meanwhile, it can dial up computational horsepower where it needs to, while also adjusting its prices for an increasingly competitive PC market.
Intel’s new business model—embedded devices talking with one another and the cloud, and becoming more powerful with each new generation—is what Krzanich called a “virtuous cycle,” with each segment of its business adding momentum.
There are still a few rough edges—for one thing, where Intel’s McAfee security business plays into this. What’s clear, however, is that Krzanich has taken one of the most dramatic steps in Intel’s history. Intel simply isn’t a PC company any more.