2018 SPECTRUM AWARD WINNER
Jacob Sonenshine Business Insider
Wall Street is throwing around guesstimates of when Microsoft will become a $1 trillion company, but there’s one catalyst many aren’t considering that could help catapult the tech behemoth to that level: Microsoft’s promising gaming business.
“We see Microsoft building out the ‘Netflix of Gaming,'” Morgan Stanley analyst Keith Weiss wrote in a note out to clients. “We think that gaming has historically been largely ignored, misunderstood, and undervalued by analysts and investors.”
Jefferies analyst Timothy O’Shea also recognizes that Microsoft is on the verge of developing a new and explosive gaming businesses.
“Google, Amazon and Microsoft are all working on cloud-based, streaming video game platforms designed to compete with PlayStation and Xbox, and these could launch before the end of the decade,” he wrote. “These services would expand the addressable market by at least 4x and would be available on all screens including TVs, laptops, tablets and smartphones.”
And Weiss agrees with how Microsoft is set to capitalize on the shift in the industry.
“With the gaming ecosystem expanding from the console to the PC and Mobile device, a shift from hardware and one-time video game sales to subscription services, and a future which includes streaming, broadcasting, and mixed-reality, we see Microsoft well positioned for the future of Gaming,” he wrote.
Newzoo market-intelligence data shows that mobile gaming leads the entire gaming market. Mobile saw a 19.3% year-over-year first-quarter revenue increase in 2017, with consoles growing at a comparably weak 3.6%.
Microsoft has already been selling those software-based gaming services, and its Xbox Live subscription service has seen its monthly active users almost double over the past four years, Weiss said.
But here’s the rub: Microsoft has a unique advantage in gaming streaming, with Microsoft Azure. “Azure is a key differentiator for Microsoft in the technology required to achieve streaming,” Weiss said, adding that it “allows for developers to scale and customize gaming infrastructure on a reliable cloud.”
Not only can Azure boost gaming revenues, but there’s yet another layer of upside for Microsoft. Xbox GamePass, a service users download onto a computer and play using a console, could go completely digital soon.
“Over the next 3-5 years we expect Microsoft to extend the reach of Game Pass from the console and PCs, to eventually cover mobile devices – greatly expanding the potential base for monetization,” he said.
Moving Xbox GamePass completely to digital could greatly expanding the entire gaming market. The move “would enlarge the gaming market as users would not be required to purchase a ~$400 console as a base requirement for gaming,” Weiss wrote.
Gaming could be what gets Microsoft over the finish line in its quest to becoming a $1 trillion company. Weiss said his margin estimates look “increasingly conservative given our stance that the mix of Software and Services Gaming revenue should increase to >80% in the outer years of our forecast period.”
His projection of a 45% gross margin on the gaming unit by 2021 is a key piece to “support a path to $50 billion in earnings-before-interest-and-tax and a $1 trillion market cap for Microsoft.”
Writer Brian Barrett
Personal computers have not elicited many thrills of late. They’ve gotten cheaper, sure, and a little faster. But despite some wild promises the gap between the actual and the possible has remained expansive. This week, it narrowed significantly.
PCs that work on smartphone parts. Devices that run all day, but for real this time. A 32-core hellbeast processor. The first GPU shrunk down to a 7nm process. Always-connected 5G laptops. And while not a PC, the ASUS ROG put vapor cooling in gaming-focused smartphone. (Vapor cooling. In a smartphone.) This was the bounty of Computex, the Taiwan trade show that serves as the cradle of innovation for what goes inside your gear.
Not all of these innovations will wind up in consumer-facing devices, at least not for a while. Not all of them are strictly necessary for the majority of the computer-using masses. But if you’ve grown tired of waiting for the future to become the present, this week, the industry pressed the fast-forward button.
Let’s start with the near-term, and work outward from there. Qualcomm, for years the king of (non-iPhone) mobile processors, set up shop in Windows PCs last December, pushing the convergence Google has already pushed heavily in its Chromebook line. The premise: all the productivity of a laptop, with the battery life and connectivity of a smartphone.
That first effort used a Snapdragon 835 chip, the same you’ll find top-tier smartphones like the Samsung Galaxy S8. Its successor, the Snapdragon 850, is strictly for laptops. You still won’t want to edit video on a so-called Always Connected PC, and it’s not going to do you any favors playing PUBG. But it’s a new enough platform that the gains between each generation are potentially transformative, rather than iterative. The Snapdragon 850 promises a 30 percent systemwide performance jump over the 835, triple the AI performance, and up to 25 hours of continuous-use battery life.
Keep your grains of salt handy, especially over battery life claims, which are notoriously juiced industry-wide. But by optimizing for Windows 10 specifically, the 850 potentially gives PC makers the key to a viable everything machine, feather-light laptops that work anywhere, any time, for as long as you need it to. And they’re coming this year.
“It’s a big deal. The future of all notebooks is going to be like this,” says Patrick Moorhead, CEO of Moor Insights & Strategy. “They’re always going to be connected, and you’re truly going to be able to use them the entire day, regardless of what you’re doing, without having to bring a power cord with you.”
If connectivity and efficiency don’t rev your engine, fear not: AMD has you covered. Its Threadripper CPU line may sound like it was named after a dainty Victorian serial killer, but its next-generation release includes a variant that packs in 32 cores and 64 threads. For the uninitiated: That is very, very many cores and threads. Contextually, Intel made a huge splash last May—with a chipset that had half as many.
“Particularly in workloads like video editing, doubling the amount of cores is almost linear. What would have taken you a half an hour to edit or recode a 4K or 8K video now takes 15 minutes,” says Moorhead. “Workstations, creative-type applications. Any type of rendering is where you’ll see the real-world benefits. Wherever you see the hourglass today.”
In fact, think of Threadripper 2 the way you might a tricked-out auto-show reveal, there more to demonstrate what’s possible than for you to actually buy. (To be clear, you can buy it, sometime in the next few months, for a yet-unspecified price.) But it’s got plenty of appeal, even on possibility alone.
“In the past, I’ve built my own PCs from various components. I would seek out the best processor, the best GPU I could get. A 32-core PC processor sounds like really cool stuff to me,” says Shane Rau, who leads semiconductor research at IDC. Rau cautions that his enthusiasm isn’t an endorsement, especially given that Threadripper 2’s actual performance remains to be seen.
Still, it’s progress! Real, near-term progress. And that’s before you even get to what’s coming a little further out on the horizon.
The most directly impactful development out of Computex this year may ultimately come from Intel. It had chip announcements both whimsical—the limited edition Core i7-8086K is Intel’s fastest yet, in celebration of the 40th anniversary of the x86 chip architecture that redefined personal computing—and aggressive, teasing a CPU with 28 cores. But look, instead, at Intel’s purported display breakthrough, which claims to halve the amount of battery used by a laptop screen.
“Being able to tackle what is one of the big power consumption aspects on your average laptop or mobile device, the technology pieces that Intel’s put into doing some very dramatic power reduction in display is I think going to have the biggest near-term impact,” says Eric Hanselman, chief analyst at 451 Research.
Intel Low Power Display Technology is classic trade show sizzle, promise of performance without much detail to back it up. It comes in the form of a 1-watt display, manufactured in partnership with Sharp and Innolux, and Intel pegs the overall battery life gains at up to eight hours. That’s about as much information as you’ll get for now; an Intel spokesperson described it as more of a prototype than a ready product. But let that temper your expectations, not deplete them. Your screen is a battery-sucking vampire. How wonderful that someone’s even attempting to fashion such a sharp stake.
(Let it be noted here that Intel also showed off a dual-display, clamshell computer this week, in case you’re into that sort of thing. It gets demoted to a parenthetical because of how many manufacturers have tried and spectacularly failed to make that work over the last five or six years. But godspeed!)
And then there’s arguably the real star of the show, an AMD prototype of the first-ever GPU built on a 7nm process. Unlike some of the other blockbuster announcements out of Taipei this week, this one won’t make its way to your computer any time soon. It’ll find a home in data centers, helping AI and machine learning compute at blistering speeds, using a lot less energy to do so.
That means the gains won’t be as flashy or as visible as those from other corners. It’s still a big honking deal. “The jump from 12nm to 7nm is going to be significant,” says Hanselman. “7nm means that they ought to be able to save on significant power savings in terms of compute capacity. It’s a big potential step.”
It’s big, too, that AMD is the the company that took it. After years of lagging behind, it’s now throwing down some gauntlets of its own.
“We’re back to this familiar sense of competition,” says IDC’s Rau. “These companies are producing better and faster, more cost-competitive products that I think bode well for the PC.”
And perhaps more importantly, for those who buy them, whether they need an all-day, always connected device or a hair-singeing powerhouse. It’s too early to know what they next generation of PCs will look like, or how they’ll incorporate all these new toys. But at least on the inside, they’re already brimming with potential.
Intel used a bit of misdirection and what it’s calling an accidental omission to combat new product announcements from rival Advanced Micro Devices an important trade show this week.
Intel showed off a new super-chip that, on the surface, seemed like a huge breakthrough for video gamers and other consumers who push their PCs hard. But in fact, after the applause died down, later reports showed that the chip wouldn’t be within reach of most PC users, even those consumers who want extremely high performance hardware.
The dustup started when Intel, the world’s leading chipmaker, gave a keynote address at the Computex trade show in Taiwan Tuesday, just a few hours before AMD took the stage. Leaks and hints had already indicated that AMD was planning to update its high performing line up of Ryzen chips introduced last year, including its 16-core Ryzen Threadripper.
So the pressure was on Intel to match the excitement. But in end the presentation only created confusion and controversy.
The episode comes as Intel is locked in an increasingly competitive battle with AMD, which revamped its entire chip line up last year to better match Intel’s offerings. And it also follows scrutiny of a demo Google did at its I/O developer conference last month showing off an AI assistant making appointments over the phone.
At Tuesday’s keynote, Intel started with one new offering that was straight ahead and clear. The company said it would make a “limited edition” of its Core i7 desktop processors dubbed the i7-8086K in honor of the 40th anniversary of Intel’s groundbreaking 8086 chip that helped launch the PC revolution. The new $425, six-core chip’s claim to fame was a maximum top speed of 5 GHz. To be made available in limited quantities and based on existing designs, it wasn’t exactly a headline grabber.
So Intel finished up its keynote with what it described as a “preview” of a new “incredibly high performing” chip–a “beast” according to Intel’s Greg Bryant–that would be available by year end.
Intel demonstrates a new chip with 28 cores at the Computex computer show on June 5, 2018.
[/caption]A picture displayed above the stage showed a typical teenage video gamer wearing a bulky headset and pumping his fist. Without much explanation, Intel said it had created a monster 28 core CPU that could run all of those internal processing units at a speed of 5 GHz.
Most multi-core chips must slow their clock speed to much slower rates when they use many cores at the same time to avoid overheating. And Intel also name-checked Asus and Gigabyte as making compatible hardware, two popular suppliers for video gaming PCs. Intel then ran one of the most strenuous software benchmarks, called Cinebench, which measures performance for 3D scene rendering. It showed a phenomenal result about double that of the fastest consumer desktop PC chips.
All of that led many in the audience, including many journalists, to conclude that the new super chip would be for the millions of customers who play high-end video games or do a lot of video editing or 3D rendering. PC Worldcompared the chip to AMD’s Threadripper line, as did The Verge and ExtremeTech. When, as rumored, AMD (amd, +2.38%) introduced its second generation Threadripper, now with 32 cores, a few hours later nearly every publication also mentioned Intel’s upcoming 28 core model as well.
The problem is that the two chips aren’t remotely comparable. The new Threadripper chip fits in the same socket as current AMD consumer chips and is expected to cost around $1,000. That puts it at the high end of the consumer PC gaming market.
By contrast, when a few tech news sites got to go behind the scenes of Intel’s demo later, they discovered that the mysterious 28-core chip only fit in a PC with a high-end socket called LGA 3647 that is reserved for server computers and very high-end workstations. That meant it was likely a variant of Intel’s Xeon Scalable line for corporations that costs up to $10,000, or five times what its most expensive consumer and “prosumer” chips cost.
It also turned out, as Intel discreetly admitted later, that the 5 GHz clock speed for all the cores on the new chip was only achievable because the computer was connected to a hidden 68-pound water chiller meant for large home aquariums to keeps its temperature down. Intel said the presenter accidentally forgot to mention that the system was boosted above its normal speed, or overclocked. Intel wouldn’t say what speed the new chip will run out of the box, or under overclocking using more typical cooling set ups.
In a statement, Intel (intc, -1.50%) emphasized that the new chip was not a phony demo, but a real product, and that it cited a commercial benchmark, Cinebench, not a video gaming test. “The 28C demo at the keynote is a real product in development targeted at the high end prosumer and enthusiast audience,” Intel said. “Intel continues to optimize design and process across its products and the demo showcased an upcoming Intel product having the capability of 5.0 GHz overclocking across all 28 cores.”
The issue stems from an app Apple introduced at its Worldwide Developers Conference (WWDC) that will be become available in the fall with Apple’s iOS 12. The app helps create shortcuts for Apple’s AI assistant Siri that are tailored to the user.
At WWDC, Apple gave the example of users saying “Hey Siri, I lost my keys,” and Siri would connect them to a dongle on their keychain which would make sound to help find it.
Shift alleges that Apple’s Shortcuts app logo — two rounded corners rhomboids fading into each other to vaguely look like an “S” — was copied from the Shift logo — two squares layered on top of each other (like a pointed figure eight) that also resembles an “S.”
“It’s mind-blowing that Apple, the firm with the biggest cash pile in history, the firm that is so design oriented, had to copy our logo,” a spokesperson for Shift told The Sun.
In a statement to Fortune, Shift reiterated that Apple “infringed on our logo which we have relied on to identify us in the community,” but said it wanted to resolve the issue amicably and not distract from the company’s blockchain-oriented website goals.
According to The Sun, Shift sent Apple a cease and desist letter, accusing Apple of “unauthorized use” of its logo. The legal representative said the company spent “substantial time and effort in advertising, and promoting” the logo as the icon for its downloadable app.
“As a result, the Shift trademark has become an asset of susbtantial [sic] value and a symbol of our client’s goodwill,” the representative said, according to The Sun, which saw a copy of the letter.
Shift asked Apple to either change the Shortcuts the logo, or pay it $200,000 so it can “rebrand” and hire a designer to create a new one.
At the time of publication, Apple had not responded to a request for comment from Fortune.
As a major player in the cloud infrastructure market, Microsoft is keen on figuring out better ways to store and power its machinery. To that end, it’s now deployed a data center on the seafloor off the Orkney Islands in Scotland.
The idea is part of Project Natick, an initiative to identify environmentally sustainable ways of running large-scale data management units. The first phase, which kicked off in 2014, saw Microsoft build a small submersible data center that housed the equivalent of 300 desktop PCs and drop it into the ocean back in February 2016 to test whether it’d actually work.
That vessel operated for nearly four months. In phase two, which kicks off today, the company has fit 864 servers on 12 racks in a sealed submersible that’s about the size of a standard 40-foot-long shipping container; it’s designed to function for five years without requiring any maintenance.
Why take all this effort to dunk data centers underwater? According to Microsoft, it’s about energy efficiency and fast data transfer. The company notes that roughly half of the world’s population lives within 120 miles of the shore, so being able to house servers near them could allow for quicker access to online services.
Next, these can be powered by wind mills located near the water, so you’d never have to worry about outages or energy costs. Lastly, the water in the ocean is perenially cold, and can thus take care of cooling the data centers without incurring additional costs for thermal balance.
This particular data center will be monitored for the next 12 months; researchers at Microsoft will keep an eye on factors like power consumption, internal humidity levels, and temperature. The company hopes that it’ll eventually perfect this system of deploying data centers that are cheaper to run for years on end.
This isn’t the first time Microsoft has taken its infrastructure below sea level. Last September, it partnered with Facebook and Spanish telecom firm Telxius on the Marea Project to lay a 6,600 kilometer-long (~4,000 miles) undersea cable between the Virginia Beach in the US and Bilbao, Spain, to transmit data at a speed of up to 160 terabits per second.
A kind of sanity returned to Windows’ status in May as the outgoing Windows 7 finally dropped some significant user share while Windows 10 padded its tally.
According to U.S.-based analytics vendor Net Applications, Windows 7 sloughed off 1.8 percentage points last month, accounting for 41.8% of the user share of all personal computers and 47.3% of all those running Windows in May. (The second number is larger than the first because Windows powered 88.4% of all PCs, not 100%.)
May’s decline was the largest in nearly two years, excepting a late-2017 reset when Net Applications purged its data of criminal bot traffic.
The change from past months was dramatic: In both March and April, Windows 7 gained ground, exactly the opposite of what Microsoft wants to see as it pushes customers to adopt Windows 10 and rid themselves of the older Windows 7.
Meanwhile, Windows 10 increased its user share last month by nine-tenths of a percentage point to reach 34.7% of all PCs, and 39.3% of all Windows personal computers. That, too, was in contrast to the two months prior, during which Windows 10 backtracked by a total of eight-tenths of a point.
The news for Microsoft may have been good in May, but long term, the outlook is still unsettled. Basing a forecast on Windows’ past 12 months movement, Computerworld estimates that Windows 7 will account for nearly 35% of all active Windows editions in January 2020. At that time, Windows 10 should power 57% of all Windows laptop and desktop PCs.
The latest numbers advanced the cross-over point for Windows 10 – when the newer OS will run a larger percentage of all Windows PCs than the older edition – from February 2019 (using April’s numbers) to November 2018. The latest trend line for the two operating systems shows that at the end of this year, just 12 months from Windows 7’s retirement, Windows 10 will run 45.6% of all Windows systems, while Windows 7 will power 43.2%.
Elsewhere in Net Applications’ data, the user share of Windows 8 and Windows 8.1 – long combined by Computerworld – stayed steady in May at 6.5% of all personal computers and 7.3% of those PCs running a Windows flavor. The 2012 (and for Windows 8.1, 2013) operating system accounted for about the same fraction of last month’s Windows-based online activity as the ancient Windows XP, which still held a 5% user share.
Within a year, Windows 8/8.1 should drop below 5%, another confirmation that the OS was a Microsoft flops. The dramatically-different Windows 8/8.1 never caught on, peaking in November 2014 at around 18%, or several percentage points lower than the top mark of 2007’s much-more-derided Windows Vista.
Also, in May, user share for Apple’s macOS fell by two-tenths of a percentage point, to 9%, after climbing almost that same amount in April. Linux, whose loyalists have perennially touted the open-source OS as the cure to the world’s ills, recorded a one-tenth percentage point boost in May, returning to the plus side of 2% after two months below that bar.
Net Applications calculates user share by detecting the agent strings of the browsers people use to visit its clients’ websites. It then tallies the visitor sessions – which are effectively visits to the site, with multiple sessions possible daily – rather than count only users, as it once did. Net Applications thus measures activity most of all, although differently than rival metrics sources which focus on page views.