The Implications of Microsoft and Apple’s ARMs Race
After years of iteration, are computers finally going to evolve?
Even the best Black Friday deals didn’t kick up enough dust to hide Apple’s new ARM-powered Macbooks. The Cupertino giant made a mockery of Intel’s x86 chips with its M1 processors, marrying both performance and battery life in a svelte package. Compatibility concerns aside, Apple’s looking pretty smug right now. Instead of having a single ARM variant like Microsoft did with the Surface Pro, Apple boldly launched three new Macs. Despite dealing with an unprecedented pandemic, I was stunned to see shoppers line up outside an Apple Store and walk out with new Macs as though they were getting their groceries.
It marks a new beginning not just for Apple, but for the tech industry as a whole. Intel grappled with its manufacturing process for years without being able to shink its transistors from the dated 14-nanometer process while Apple’s new Macs tout 5 nm chips. The advent of an ARM-powered future has ramifications that will ripple across several facets of the consumer tech business.
But if you think Apple’s the only horse in the race, I’m sorry to burst your bubble. Companies have been trying to replace ARM chips with x86 ones and vice versa for decades. Not all of them were big enough to absorb losses. And while some have celebrated varying degrees of success, Apple throwing its hat in the ring elevates the bar brawl into a full-scale siege. But before we start placing bets on who’s taking the crown, let’s examine the conditions that got us here.
Rapid advances in ARM processors
Designing processors is no walk in the park. Intel’s evident struggles with shrinking its transistors are more than enough proof. Over 35 years ago, ARM decided to make the lives of chip manufacturers a whole lot easier by designing the components that went into said chips. This gave chip designers a blueprint of sorts, letting them create chips that could fit in everything from pregnancy tests to supercomputers. Licensing these designs to partners proved instrumental in its success. 160 billion devices later, ARM’s processors are more relevant than ever. If you’re looking for a history lesson, this exhaustive list might help.
Manufacturers began integrating several components together to build SoCs (system-on-a-chip), boosting performance by incorporating several functions into the confines of a single microprocessor. Qualcomm partially owes its undisputable smartphone CPU reign to ARM’s licensed designs. And yes, even Apple’s new chips benefit from that very repository of knowledge. This lets their newly launched Mac lineup look at Intel’s power-hungry x86 chips in the eye and wink on their way to the bank. Smartphone-grade efficiency and desktop-level performance in the same sentence is no longer a pipe dream.
Other firms like Samsung and even NVIDIA (its Tegra chips are at the heart of the Nintendo Switch and Tesla’s earlier self-driving efforts) tinkered with ARM to realize their ambitions. Providing a chip architecture that would consistently improve with time helped ARM propel the consumer tech industry into the throne it occupies today. But ARM has foes too.
Intel’s x86 mobile attempts were disasters
While ARM had its eye on efficiency, Intel had other ideas. Both firms have entirely different visions for the future. And while Intel took a while to fabricate 10nm chips, they no longer hold their own against not only its x86 peer AMD but ARM’s drastically buffed chips. It tried its hand at the mobile market too, an arena that ARM is more than familiar with. In 2016, Qualcomm’s devious stranglehold on licensing and purchasing terms meant that Intel’s Medfield chips were dead in the water. Motorola’s slim Kevlar-toting RAZR i phone couldn’t take a bullet to the chest.
One cannot talk about Intel’s anti-ARM efforts without bringing up its Atom chips. While it is known today as the chip that made netbooks a thing, Atom processors were supposed to beat at the heart of Intel’s MIDs (Mobile Internet Devices). These keyboard-bearing phones were supposed to bridge the gap between mobile devices and laptops, a gap that ARM has recently teleported past. MIDs didn’t make it past the “bizarre experiment” stage. Instead of doubling down on Atom, Intel chose not to shift its focus from its primary source of revenue: regular computers. Its underestimation of cellular radios cost them the game. For a deep dive, I highly recommend ExtremeTech’s two-part piece that examines Intel’s wobbly missteps in the mobile market.
Microsoft might still share close ties with Intel but its recent shift in both the data center and consumer tech spaces mean that things are changing. With Apple and Microsoft betting big on ARM, Intel’s future is uncertain. And while it has the fortunes (and the brains) to plug the holes, a sinking ship can only sail for so long on uncharted waters. Laptops are exciting again in 2020, and Intel isn’t in the spotlight this time. But Intel isn’t the only one in for a rough ride.
Microsoft’s first ARM PC? It cost them $0.9 billion in losses
Remember the Surface RT? In 2012, it boldly imagined a future whether 2in1s could eclipse both traditional laptops and tablets. Alas, Microsoft’s first ARM attempt placed its bets on “Universal Apps” (sound familiar, Apple?) that had to be built with ARM and x86 compatibility in mind. While Apple had fortresses on both sides, Microsoft didn’t have a potent app ecosystem on the ARM end of things. Without a means of getting software developers on board, it was dead on arrival for most consumers. Its successors did little to change its Achilles’ heel while the Intel chip-equipped Surface Pro lineup kept getting thinner and lighter.
Today, the Surface Pro 7 strikes a fair balance between performance and portability. But Apple’s ARM laptops didn’t have to sacrifice either of these tenets on their path to market domination. And while the Qualcomm SQ1-toting Surface Pro X did show up first, it was a cautious dip in the pool while Apple went full tilt on ARM. Windows on ARM recently gained 64-bit emulation support but it has yet to achieve the level of polish that Apple managed to accomplish. True, reports indicate that not all is well at the Cupertino camp. But that doesn’t take away from the fact that Apple’s Rosetta 2 software remains the best Intel x86 application emulation service on the scene.
Microsoft’s ARM experiments didn’t start with Surface RT though. They had even bolder plans for an experience that I still remember fondly: Windows Phone. While it has been written off as a failure in the three-way horse race against iOS and Android, Microsoft had loftier ambitions. Apple put the heart of a smartphone in a Macbook. Intel tried to do the opposite and failed. Microsoft’s vision? Make those smartphone guts run desktop-grade applications. Yes, not entirely different from how Apple’s Macs did it.
Microsoft can still make good on Windows Phone’s promise
Picture this. The power to run regular desktop applications on your phone. Hook it up to a monitor setup and you effectively have a portable workstation in the palm of your hand. Don’t believe me? Someone got PhotoShop to work on a Windows Phone. But this wasn’t how Microsoft sold its vision to customers. Instead, it differentiated itself with an experience unlike anything iPhone or Android users had experienced before, starting with the app gap.
I’m not saying that Windows Phone didn’t have its share of advantages. The OS that launched a thousand budget phones did get a lot right. It even had offline maps and a solid Spotify equivalent (MixRadio) before streaming took over the music industry. But vivid Live Tiles and capable hardware could only get Microsoft so far. Its phones met the same fate as its Surface RT line, an untimely death. And while some may argue that they were on the cusp of brilliance, app support is a heavy anchor that Microsoft had to deal with.
But with emulation software getting better, Microsoft could shake things up with its new Android-powered foldable, the Surface Duo. With a form factor that already promises boosts in productivity, the Redmond behemoth could stage a coup with the help of x86 app support. It could give Windows Phone a second wind, defying its eulogy in the process. While it’s more of a pie in the sky at the moment, desperate times call for unconventional measures. But over at the desktop side of things, the app gap isn’t what’s holding Microsoft back. Apple proved that it possessed something that Microsoft doesn’t: courage. And a userbase willing to put up with temporary setbacks on the journey to a radical future.
Apple did what Microsoft was afraid to do
While Microsoft’s Surface Pro X serves as a cautionary tale for ARM cultists, Apple’s new lineup is a swashbuckling adventure in comparison. Not one but three of its Macs showed up with chart-topping ARM chips. Only Apple could pull off such an experiment and get away with it. And this isn’t the first time Apple was “courageous.”
The headphone jack debacle is one that continues to divide smartphone users but that hasn’t stopped manufacturers from getting away with removing a feature that was once ubiquitous. Apple’s decision to sell its iPhone 12 lineup without chargers stirred the pot once more. And while other smartphone vendors may gain some clout with witty tweets, it’s only a matter of time before they drop their chargers into the same rabbit hole.
But unlike those measures, Apple was prepared with countermeasures in place to ensure that the migration went smoothly. Robust x86 emulation and iOS app support are certain to ease Apple’s x86 to ARM transition in a way Microsoft could only dream of. The Windows camp has neither of these bullet sponges to make professional use-cases feasible. It’s no surprise that Microsoft’s latest phone (the Surface Duo) adopted Google’s Android operating system. App compatibility dictates where users transition to. And while macOS isn’t known for supporting every obscure application out there, Apple’s ARM-powered machines still wipe the floor simply because they nailed emulation. Traditional applications are no longer restricted to hulking laptops with mediocre battery life. But at what cost?
The tussle between freedom and freedom
Apple made a big deal of the way it approached building its latest M1 SoC for the Macs. Considering that they managed to leapfrog Intel’s performance plateau, some may consider the tradeoffs worth it. But while unified memory and integrated components let them work better together, this jump has its fair share of caveats. Soldered components come at the cost of freedom. You can’t open them up and add RAM chips or storage drives. Consumers are now at the mercy of the manufacturers since the latter now have full control over pricing. Apple charges $200 for a RAM upgrade on its latest Macbooks and there’s nothing you can do about it. But this loss enables a different kind of freedom.
iPads running those mighty M1 chips? It’s a possibility worth pondering over. Running macOS lets Apple’s tablets clear the last hurdle standing between them and real computers: desktop app compatibility. And with iPads costing as little as $329, a future where Apple devices are affordable continues to keep me up at night. But tablets aren’t the only devices that will see improvements. Smartphones could be tethered to monitors to run full-blown apps like PhotoShop or MATLAB. Digital education might just finally reach the masses, fuelled by smartphones that can perform double duty as traditional computers. And with foldables crawling their way out of infancy, the possibilities are exciting.
Integrating components together on a single chip lets them consume less power and occupy less space, letting these processors show up in smaller devices. True, companies like Lenovo managed to stuff an entire Intel CPU into a foldable device but think of the uptick in efficiency if they switch to an ARM alternative. Without fans and cooling systems, think of how small our devices could get. And with external GPUs slowly becoming relevant, consumers can leave the heavy-duty parts at home or at the office. Speaking of graphics cards, AMD and NVIDIA have a bigger role in an ARM future than you might think.
NVIDIA is more than relevant — it bought ARM for $40 billion
It’s no surprise that AMD has been making a killing. Its latest 7 nm processors put a serious dent in Intel’s hubris and Team Red’s Big Navi GPU lineup is no longer sidelined by NVIDIA’s efforts. AMD is adopting the “if it ain’t broke, don’t fix it” approach while NVIDIA decided to fill an ARM-shaped hole in its sleeve. $40 billion dollars later, the company literally bought a ticket to the future in one sweep. Regulatory scrutiny aside, Team Green really is doubling down on an ARM-driven future.
It’s no surprise given NVIDIA’s ambition to penetrate and dominate as many spheres of influence as possible. The fastest supercomputer today relies on an ARM processor, standing as a testament to the architecture’s ubiquity. From data centers to self-driving cars, NVIDIA’s hold on intensive applications is only going to get stronger. Its ARM acquisition does raise some valid concerns. Neutrality is something that can be imposed by anti-trust organizations but whether it is followed in practice is something that only time will tell. NVIDIA might even end up making ARM even more open to make the most of its opportunities. And with a little-known open-source competitor gaining prominence, NVIDIA can’t endanger its relationship with its partners.
While ARM has been the vendor-agnostic platform of choice for most chipmakers over the past 35 years, this juvenile rival is merely a decade old. But RISC-V promises something that NVIDIA doesn’t: an open-source approach that is fair to all. It’ll take significant monopolistic measures for established players like Qualcomm or Samsung to migrate from ARM’s stronghold but an extra option doesn’t hurt. Visionaries see RISC-V as the golden future of computing, potentially surpassing even ARM and x86 in prominence. And with reports suggesting that over 60 billion RISC-V chips will be out in the wild by 2025, Intel and ARM’s duopoly might be at risk.
With AI in the mix, optimization is key
It’s remarkable how effective artificial intelligence and machine learning networks have become in performing tasks that were once thought impossible. With an arsenal of servers at their disposal, rendering impressive visuals can now be done over the cloud. Chipmakers have even begun modifying their processors to support AI-intensive workloads. Be it Intel’s Gaussian Neural Accelerator or Apple’s Neural Engine, firms know that AI can serve as a walking stick for chips well into the future. Even graphics cards, already potent at number-crunching, now possess hardware designed specifically for AI and ML workloads.
NVIDIA’s Turing GPUs gain an edge over AMD’s new cards with Tensor cores that lend strength to AI applications. Team Green also currently has its hands on DLSS (Deep Learning Super Sampling), a piece of AI-rendering tech that upscales games without a drop in performance. Here’s an article that deciphers the voodoo magic on display. With AI doing the heavy lifting, our devices can shrink while retaining their capabilities. If you’ve been holding out for a true generational leap in computing, it’s almost here.
No one’s going to stop for the competition to show up for a three-car photo finish. Intel’s unsteady grip on the desktop market is contested both by AMD’s conventional assault and ARM’s radical onslaught. With Microsoft and Apple vying for OS supremacy, their adoption of ARM is a sign that traditional laptops might go the way of the headphone jack. Graphic cards showing up with AI-rendering hardware warrant an overhaul of software once intrinsic to the regular x86 experience. Even data centers are getting in on the AI action. What once required cumbersome desktops can now be performed on the go. A promising future awaits us beyond the horizon and the body count of obsolete gadgets shows no signs of stopping. There’s no better time to be on the lookout for bizarre experiments and bold gambles on the bleeding edge of technology.