Google software and AI too good, but its hardware can do great?

According to media reports, a long article published recently by the famous American technology magazine “Connection” elaborated that Google is relying on Rick Osterloh to reshape the hardware strategy in order to integrate the Google Assistant Google Assistant People’s lives.

A few seconds later, on the morning of October 4, 2016, Rick Osterloh showed the world’s public the latest Google product portfolio. At this point less than 6 months before he founded Google’s new hardware division.

In April, Sundar Pichai, Google’s chief executive, was appointed to Ostrow, hoping he could turn Google’s software giant into a hardware device maker that could compete with Apple. Ostro has not had enough time to understand all of Google’s products, let alone to conceive and release a series of new products.

However, he was here, tall, broad shoulders, wearing a gray short-sleeved Henry blouse. When he stepped onto the left side of the stage, greeting reporters and analysts at Full House, it was clear that he was in a little restless.

A few minutes ago Pizarre stood on the stage did not relieve Ostrow’s tension. Picazzan pointed out the historical significance of the day and brought indestructible reasons to the audience.

“We are at a groundbreaking time in computer technology,” Pizjik told the audience, explaining how artificial intelligence will lead to a revolution on the Internet or smartphones. And Google’s focus is on the Google Assistant, a virtual digital assistant for months. Google’s vision is to create a “personal Google” for everyone on the planet through virtual digital assistants, helping users search for information, get things done, work more easily and more enjoyably.

Pizjaya made it clear that Google Assistant is a re-noted product under Google, and that Google has invested heavily in developing electronics that integrate Google Assistant. Then he introduced Osterlo, the newcomer who will make this vision a reality.

Within the next hour, Osterlo and his new colleagues launched six products at a stretch, including the Pixel smartphone, the Home smart speaker and the Daydream View virtual reality headset. These products are not dominated by Ostrow, before it came to Google’s headquarters in Mountain View, Google has already started to develop hardware devices, but most are not very successful.

Google no longer wants to develop gadgets that make people feel bored. Alphabet, the parent company of Google, has become the world’s second-largest company, with software developed by Google that serves all the world through apps and websites.

But the nature of computing is changing, and the next phase of development will no longer be around app stores and smartphones. It will focus on the development of artificial intelligence devices that fit seamlessly into the daily lives of its owners. This includes voice assistants, simple wearable devices, home smart devices, and augmented reality devices.

In other words, the future requires more hardware, which Google represents as a threat to itself. Users will no longer search for what they want by logging in to Google’s site; they will only ask for Echo smart speakers within their hearing range and will not care what algorithm Echo will use to answer that question.

Or they’ll use Siri because it’s included with the button on the iPhone. Google needs to figure out exactly how to compete with the electronics manufactured by Amazon, Apple and other technology companies, especially with products from Apple’s headquarters in Cupertino.

Google does have some huge advantages – its software and artificial intelligence features are unparalleled. However, companies have repeatedly tried to develop the hardware in a way that builds the software, but every time it goes missing.

Its innovative streaming media device, Nexus Q, suffered a dramatic failure. In a matter of months, the so-called “best of breed” Nexus handsets have outpaced rivals and even Google’s own hardware partners. As Google Glass Google Glass, we all know what is going on.

Ostro’s job is not to develop new hardware products. His mission is to teach the software company how to overcome the drying up, messiness and lengthy time necessary to develop hardware products and change the corporate culture from the inside. It is not enough just to have great software and the best artificial intelligence researchers in the industry. In order to compete with Apple, Google had to learn how to develop good hardware products.

Ostrow, who is responsible for Google’s hardware renaissance, is fascinated by electronics. Ostrow, who grew up in Los Angeles, once disassembled junk computers in his father’s office trying to assemble a supercomputer. Although eventually ended in failure, but still good memories of Ostrow.

Ostrow initially obsessed with sports, from an early age he was tall and athletic ability, both a volleyball player, but also a basketball player, he was admitted to Stanford University not because of its reputation in Silicon Valley, and Because you can keep moving in the best school in California.

However, in the freshman year, his knees injured twice, may end sports career. Ostrow’s mood was very low. He said: “Much of my identity is about track and field, and I have to completely change myself.” He started looking for other ways to get his positive feeling in sports: teams working toward common goals, Pleasure and usual pleasure.

He found the same feeling plan in a project and began working hard to make up for the shortcomings in his professionalism. Ostero is immersed in strategic thinking about computing as well as problem-oriented solutions.

Still, Ostrow is still a sports fan – as he can easily see at his Google office – with a huge poster about Warriors star Stephen Curry – but technology The industry soon became his destination.

After graduating from Stanford University in 1994, Ostro won a consulting job, but he did not like he was dealing only with documents and slides. So he went to Stanford University Business School to rebuild. After some internship with Amazon, he found a job at Kleiner Perkins Caufield Byers, a venture capital firm, who looked at possible investments in mobile technology.

At that time, the BlackBerry started to attract public interest, and Ostrow also proceeded to study it in depth. He built Blackberry’s first device, the internet pager, Inter @ ctive Pager, and was surprised by the productivity of this small form factor communication device. He has been thinking about this thing.

In Kleiner’s portfolio, there is a company called Good Technology, which has assigned Osterro to help shape its business model.

Initially, Good’s plan was to develop a module for the Handspring Visor (a modular PDA) that many would consider a PDA as the next big computing platform. Good’s first device was an MP3 player module called SoundsGood. However, Visor never took off, and SoundsGood also sold very badly.

Ostrow came up with a new idea: let’s compete with the BlackBerry. He thinks Good can develop simple synchronization and messaging software, and the BlackBerry is already very powerful and valuable in this space, with any competing idea appealing to investors. Good raised millions of dollars.

Good is a software company, but it requires hardware to run the code. The team met with the BlackBerry who just started making smartphones. Blackberry executives looked at what Good was developing, “They exclude it because Good’s software is better than BlackBerry,” Osterlo said. “They realize we’re enemies, not friends.”

At that time, Palm, Danger and Nokia are developing smart phones, but none of them are comparable to the BlackBerry. Both Ostrow and Good are well aware that the only way to find a working device for their software is to develop their own hardware and as such they are starting to develop a G100 device similar to a BlackBerry.

Ostrow has benefited from recalling his development of the G100. “It is very interesting to go through the entire design process and test with the user,” he said. Everything about hardware development is new and complex: how to get the keyboard right, how to adjust the trackball until it feels perfect, and how to make sure the battery lasts for days.

Ostro said: “The development of this product is very difficult when we ship, I think ‘This is what I want to do forever’.” He not only found his mission, but also clearly from the software In the most profitable, the only way is to develop matching hardware products.

Unfortunately, this is only Ostrow’s wishful thinking, Good does not want to develop the hardware forever. The G100, which went public in 2002, received wide acclaim in the industry, but others at the company saw it as a reference device and a blueprint that other companies could follow and adjust. They think the mobile phone industry will look like a personal computer: Many companies produce hardware but run the same software.

As a result, there is no need to develop a good mobile phone product. Ostrow said: “After that we seem to have walked into the desert and tried one new device after another, but none of them matched our software.” Good worked on software for every cell phone at the time, and even And to inject HTC and other mobile phone manufacturers to improve the user experience, but no longer find the same products as the G100.

“Companies come to us and say ‘we like your software, but we hate Treos’,” Ostero refers to Palm’s smartphone line. He never forgot it.

In 2006, Good was acquired by Motorola, the giants of this feature machine was already in the siege of the smartphone maker. Motorola has no software expertise and no plan to develop smartphones. Good joins like a savior. But the timing can no longer be worse. Just days after the acquisition was completed, Motorola’s Razr was once an incredible cash cow and stopped selling almost overnight.

Soon after, Apple released the iPhone. Osterre clearly the trend has been irreversible: Prior to the deal with Motorola, he and Good worked with Apple to develop software on the new device iPhone. He told bosses who did not like Apple’s touchscreen phones that they were laughing at the future.

While dealing with Apple, Ostrow and Good have also been trying to integrate their software with a smartphone operating system called Android.

Now, as a Motorola employee, he thinks Android is the only way companies can fight the iPhone. Ostrow is convinced that Motorola’s only hope is to develop as soon as possible an equal level of smart phones, which means to use the Android operating system.

Motorola eventually went into effect, largely because of the efforts of the new chief executive, Sanjay Jha, who closed almost all of his divisions except Android after taking office in 2008. Ostrow helped develop and release the Cliq and later Droid, the first Android phone to be great, and the device that pulled Motorola back from the cliff.

Pictured: Before Ostro joined Google, the company’s hardware product development team was up against each other. Ostrow has focused hardware products, including Daydream View virtual reality headsets, Chromecast laptops, home smart speakers Google Home and Pixel smartphones.

Shortly after, Ostrow left to join Skype as product manager for two years. However, Ostrow breakthrough time in the hardware field is very short. Google is buying Motorola for $ 12.5 billion and is looking for new leadership for the company. Dennis Woodside, Google’s senior vice president, was named chief executive of Motorola. Google senior vice president Jonathan Rosenberg called Ostrow and asked if he was willing to come back to lead Motorola’s product management team.

The offer from Google seems to be a perfect match. This gives Osterlo the opportunity to develop hardware inside of Google and work with the successful Android team. With Google’s hardware and software support, they finally have the capital to compete with the iPhone.

but it is not the truth. Because of fear of alienating other Android partners such as Samsung and LG, Google makes every effort to maintain Motorola’s independence. “There’s virtually no technology consolidation,” Ostrow said. “That’s not what I expected.” He had thought he would eventually combine software and hardware, but Motorola was seen as a completely separate company . He said: “It’s really close to what I’ve been dreaming of, but it never got there.”

The relationship between Google and hardware has always been awkward. Most of the company’s hardware products are born in the same way: Someone has a good idea of ​​the software, but they can not find the right device to run the software. Then, the man started to make the device himself, and the team did little to help it.

Google tends to see these products as a reference device or source of inspiration, confirms the viability of an idea, and hopes that the hardware manufacturer’s ecosystem can benefit from it. So from Chromebox, Nexus Q to Nexus Player, there are so many similar Google products and abandoned ideas.

This is not surprising: manufacturing hardware runs counter to Google’s entire corporate culture. Companies exclude standardized processes and management, but these are two things that hardware manufacturers can not do without. In fact, Google encourages and promotes chaos in its software development and invariably invites anyone in the company unexpectedly to develop something, even just to see if it works. (Google even once tried a corporate structure without any manager.)

Google’s most successful products are constantly improving. Eric Schmidt, a former chief executive, calls the system “posted and iterated.” His Proposal in “How Google Works” is a consistent concept that does not try to get things done right once.

Schmidt writes: “Create a product, post it, see how it works, then design and implement improvements, then push it out … Post and iterate, and the fastest companies in the process will win “When Google changed its name to Alphabet, all of the company’s long-term projects seemed to be halting and it was being spun off from Google’s product line. These projects are called “moonshots,” and it seems like anything that takes more than a year is impossible.

But release and iteration are not suitable for hardware development at all. A fine-tuning of a hardware product can take weeks and cost millions of dollars at the same time. Every minute change in the hardware product spills over the entire supply chain, changing vendor time schedules, changing existing equipment, and slowing down all progress.

If some of the time is postponed, you will miss the original delivery date and will not reverse your Black Friday. You want to set 50% more goods? Well, even if you’re lucky, you’ll have to wait six months. The hardware world will not change because of your whim.

Even as hardware development goes well, Google’s corporate culture does not support these teams to get the software they need. “We had to beg for all the software teams,” said Rishi Chandra, Google’s senior staff. He was responsible for developing Google TV platform in 2010, but eventually failed. He was then responsible for developing Google Home.

Engineers working on Chrome or Android are accustomed to developing products that can reach millions or even billions of people. They will ask Chandra, how do you attract so many users? Why do we care about your product until it can land?

Google’s corporate culture is exactly the opposite of Apple. Apple’s software executives are always concerned with specific products, such as Craig Federighi, senior vice president of software engineering, whose goal is to make the iPhone a better place. At Google, however, the priorities for various issues do not seem to be one of the top priorities because they support the products they own, as well as the products they use for partners and across the Internet.

Google’s corporate culture revolves around software. This is its strength and its billions of dollars in profits. This approach is no different and no exception in promoting virtual assistants. Of course, Google’s determination seems to be much higher this time.

Pizzer is convinced that such a useful chat machine will be the way for billions of people to interact with Google in the future. If right, virtual assistants can be ubiquitous artificial intelligence to help users do all their everyday tasks – whether through mobile phones, Google Home smart speakers, smart home-bound smart bulbs, dishwashers and thermostats, and more .

In addition to the keyboard and screen, it connects information and services to users in a more natural and contextual manner. It may even encourage consumers to use Google services more often.

So what if Google did wrong? No doubt Amazon Alexa, Apple’s Siri and Microsoft’s Cortana are ready.

In the early days of speech technology, users were still struggling to find out what their virtual assistant had other than setting timers and playing music. However, the only way to improve virtual assistants is to make Google believe that it is now worthwhile to interact with virtual assistants. Google needs more data to understand more of the sound, to do more, to persuade third-party developers to expand its capabilities and to integrate virtual assistants into their hardware products. The key first step is to ensure that users always have easy access to Google’s virtual assistant wherever they are.

In the past, when Google developed hardware products, it was mainly through partnerships with experienced manufacturers. However, its relationship with Android partners is on the decline. As the most important partner of the Android operating system, Samsung is developing its own virtual assistant Bixby, and deliberately keep a distance from Google. Google is also working with its traditional Nexus project to work on new devices with hardware manufacturers such as LG or Huawei, but its own hardware development can not rely on it.

In fact, apart from being able to voice their opinions on the aesthetic, these partnerships leave Google with nothing to control, and partners often retain the best for themselves. Ostero pointed out: “Last year, HTC helped us to develop Pixel smartphones, but in a few months they released the U11, which has the industry’s best smartphone camera. In fact, every Nexus device Replaced by the handset released by Google hardware partners just after it. ”

Even more unfortunate is that Google has given up its own hardware skills, in 2014 for about 3 billion US dollars to Motorola sold to Lenovo. Nest, the other hardware vendor acquired by Google, has lost its founders and is in turmoil with management and product storms. If Google wants to develop its own hardware, then it needs to start again – this time to be done within Google.

Ostrow is also planning to resign from Motorola as Google prepares to prototype its hardware products by early 2016. He did not want to go to Lenovo’s headquarters in China, and he himself had been offered an offer to go to the CEO of DocuSign, an electronic document signing company. He called Jonathan Rosenberg, his long-time adviser and advisor, to thank him for all his help during Google.

Rosenberg interrupted him: “You said I was helpful, right?” Yes, absolutely, Osterlo answered. “So, would you help me a favor? Talk to Sandal?” Rosenberg asked. He told Ostrow that Pizjay is planning to set up a hardware team, hoping for some advice. Rosenberg also emphasized that it was just a suggestion, with no other ideas.

The meeting between Ostrow and Pizzer soon turned into an interview, and both sides raised many questions. After several days of exchange, Ostrow realized that Pizjay was talking about his dream job. He also began to believe that Google eventually began to take hardware manufacturing seriously.

But Ostrow had a lesson. So he found Hiroshi Lockheimer, head of Google’s Android team, who had worked with Lockheimer while working at Google, who happened to be Osterlo’s closest friend. They spent the whole day talking about how they became co-workers again. Ostro raised many questions about how to integrate hardware and software, and how internally-built hardware coexists with other Android ecosystems.

Ostro said: “If it would be as difficult and tense as Motorola, I would not want to join the company.” However, he found the opposite: Google is ready to make hardware a priority and take it very seriously. Eventually Ostrow called DocuSign, declining to be chief executive officer. This time he became a real Google employee again.

After joining Google, Osterlo immediately with Rosenberg to look for each of Google’s hardware projects, regardless of size. They found more than a dozen projects involving more than 1,000 people. Some work on Nexus phones while others work on a new product line called Pixel. There are also many people in the widespread promotion of such long-term projects such as Google Glass and Project Ara modular smart phones. Some Google employees are developing Chromebooks while others are researching a new Wi-Fi router.

There is no centralized management of these teams and no overall plan. Ostro called it a loose federation, probably similar to the “hardware of the European Union”, anyway, is not good words.

Ostrow has all his hardware teams under their own leadership, making 55% of these 1,000 employees new managers. Instead of letting the head of each product take responsibility, Ostrot opted to implement a “functional” architecture that allows its leaders to monitor Google’s entire hardware organization to the maximum.

Ivy Ross, who was the head of Google Glass, was responsible for the hardware design. Mario Queiroz is responsible for product management. Ana Corrales, Nest’s chief financial officer and chief operating officer, has long been a manufacturing executive who oversees all business operations and supply chains. The team started to focus on planning and forecasting and simplified communication with suppliers. They set out a five-year plan that was once a curse for Google.

And Ostrow worked with people who described him as a benevolent, a good boss. Chandra said: “What I most admire Rick is that he really is talking about patience.” In conversation, he was very excited and tended to spend 45 minutes answering all the simple questions. Based on past and present colleagues’ opinions, he is the perfect candidate for the job: high attention to detail, unresponsive panic, and quick decision-making. Most importantly, he is a well-deserved product geek.

“He always changed his cellphone, and he wants us to keep changing his cellphone,” said Claras. “I do not want to change my phone!”

Part of the motivation behind Ostro’s new structure is ensuring that no one feels that his job is tied to a product, so they do not panic if the product is killed. Because Ostrow needs to kick off some of the products.

He traverses every piece of hardware in Google’s plan, choosing which one to keep and which one to stop. Ostrow said every decision is not easy, but two of them are particularly difficult.

As early as working in Motorola, Ostrow worked for Ara’s modular cellphone project and is fully committed to the project’s mission: to develop a $ 50 cellphone with upgradable parts that can last longer, and More environmentally friendly than any other device. However, the device is less modular than anyone wants.

“So it’s just like any other cellphone, except that it adds 6 modules to the back,” said Ostrow. He just wanted to make a phone, not a lot of phones, so he closed the Ara project.

Pictured: A year and a half after taking part in Google, Oslo at the conference described how Google applied artificial intelligence to the latest electronic products.

Google Glass is also true. Ostrow understood the project’s vision, but could not figure out how to get there quickly. He screened out something that, although indispensable for the development of an excellent facial augmentation device, was not yet achievable: batteries that last longer in smaller packages, produce less heat but are faster Processor, and potential user groups that are prepared to use such devices.

“It will be a key part of our job in the long run, but timing is a key uncertainty.” Meanwhile, Osterre re-released the Google Glass for Business Edition, which includes factory workers And warehouse employees, including niche markets.

Ostrow has also been working with top Google executives to figure out what Google’s hardware strategy should be in reshaping the organizational structure and screening product lines. These executives shape clichés such as “radical help” and seek ways to communicate humanity and proximity to the public, but their focus is on three words, specifically artificial intelligence, software, and hardware.

He had to accept the fact that even if Google started taking electronics seriously, the focus of the company was always elsewhere. Osterlo likes to say that Moore’s Law that predicts computing power is about to die out, making it more and more difficult to make a fundamental leap in hardware capabilities. He said Google’s strength lies in its algorithms and neural networks. Austro’s job is to Google’s AI more deeply into people’s lives.

The task for the new hardware team is clear: find more ways to get Google Assistant Assistant to emerge and build a sustainable business around it.

This must be fast, because Google is already behind, Siri and Alexa have been deeply rooted in the minds of consumers. Osteros put a lot of resources into the Pixel mobile project, which is a new project launched between several Google employees and HTC, Google for the first time assumed full responsibility for design and engineering, HTC is just a product manufacturer.

Ostrow hopes through this phone, Google will eventually be able to provide their own software the required physical form. Ostrow said: “We have a great ecosystem on the Android operating system, but I do not think anyone can actually deliver a complete Google product experience.”

Integrated hardware and software design can make a more detailed point of decision-making, people can fall in love with the phone. Seang Chau, vice president of engineering at Google’s Pixel mobile team, cites an example: Swiping should be smooth and responsive, requiring in-depth understanding of things like when to start the GPU, how to adjust the processor, how to manage the power, and even which core of the chip is at Run at any given time.

“You pick up another cellphone that has no detailed decision and no optimized cellphone,” he said, noting the difference. Apple has been advocating its products for many years, precisely because it is still in the process of developing software and building hardware, and Google is now following suit.

In addition to Pixel, Ostrov decided to build the other devices that match with Assistant in the same way.

Another team within Google has in the past released two very good laptops, called the Chromebook Pixel, but with limited performance. Ostrow told the team to develop a lighter, thinner, and better laptop and integrate Assistant. They decided to name it Pixelbook and set out to do it. Another group started researching headsets called Pixel Buds, which allowed users to use Assistant without a phone. Google’s Home team and the Chromecast team are also involved in the overall work.

Ostrow said: “Ultimately, users may have a series of devices to complete the task.” Google will eventually think Tablet PC, will certainly consider augmented reality glasses, we will certainly consider wearable devices. But Osterho talked about “fighting for rights” by chasing device consumers to prove his team’s viability in the existing market.

As hardware staff gradually adapt to new roles, Ostrow and his team began to study their production needs. He and Claras visited Asian manufacturers, told them what Google is doing, how they would communicate, and reached a new deal with the supplier.

In November 2017, Ostrow led Google’s $ 1.1bn acquisition of HTC, which brought over 2,000 engineers to Google, many of whom have been developing Nexus and Pixel devices for the past decade people. Ostro said the deal “is very important to helping Google accelerate its expansion, and it takes a lot of time to hire one person and we want to speed things up.” In early 2018, Alphabet began letting Ostrow lead the entire Nest team, Control Google’s future in the smart home market.

Google has too many reasons to catch up. Apple and Samsung continue to introduce competitive software on their hardware products, and new-generation devices are getting better and better. However, Ostrow has stressed again and again (and reminds himself) that it is a good thing to develop a hardware product, which is a good thing, and patience is a virtue. This time, he had the opportunity to prove himself through a work on the combination of hardware and software.

Ivy Ross, head of hardware design at Google and one of Osterlo’s chief deputies, said: “It has a huge stake in him and the company. When you personally have reason to be motivated, you It will be better. ”

October 4, 2017, a year after Osterlo first demonstrated a new generation of Google hardware.

Everything is deja vu. Ostrow wore the same gray Henry blouse and stood on the side of the stage, while Pizjay also explained that artificial intelligence was the future. But this time they are in the Jazz Center in San Francisco, a bigger and more impressive place. They have been rehearsing for several weeks, adjusting the content and order of speeches to better explain Google’s goals.

Ostrow said the biggest difference is that this time he knows the story. In 2016, he tried to elaborate a grand story around many unrelated products. Although these products are very popular, the successes achieved are limited. (Amazon Echo is still killing Google Home, Pixel did not completely affect the iPhone’s bottom line.) 18 months after joining Google, Ostrow is to show the world the true face of Google’s hardware.

When he started to tell when he stepped onto the stage, he was obviously more confident than a year ago. He reminded viewers of the 2016 conference and mentioned the recent acquisition of HTC.

“By working more closely we will be able to better integrate Google’s hardware and software.” “And our products,” he continued, smiling, with joy, reassurance and ambition, “he said. In the second year the momentum has been more and more fierce. “He then turned to promotional video, users from YouTubers to Modern Family, as always, like Google Home smart speakers and Pixel mobile phone.

For the next 90 minutes, Ostrow and his team introduced a host of new products that incorporated Assistant. Ostrow explains how artificial intelligence can give users an excellent experience with ordinary hardware. When introducing Pixel 2, he mentioned that although the phone has only one camera, its software includes a face-training algorithm that makes standard photos even more beautiful. Google tweaked the audio processor inside Pixel Buds to simplify the experience of using the Assistant Assistant via headphones and to achieve real-time translation.

Home Max can adjust its audio output to match any space, improving sound quality. A new camera called “Clips” recognizes moments worth taking pictures and takes photos and videos on their own. Ostrow and his team show everyone how smarter their products are from their competitors, one by one, to everyone.

The press conference is very smooth, but not perfect. Some users find the OLED screen of Pixel 2 XL has some problems, Google chose this screen to show its cool context software. Others found problems with the Home Mini touch panel, accidentally turning it on and recording hours of audio. Commentators praised Pixel Buds product philosophy, but not every feature.

All of these issues have sickened Ostrow – “I suffer from insomnia every time my customers are upset,” he said, but it seems to inspire him. He knows how to deal with these challenges: more rigorous processes and more rigorous management. This is a typical hardware thinking, which he learned a long time ago. Different from the previous year, this time Osterllo has a clear direction of development. He leads the product is part of the entire company. After completing its mission, Google needs to do more than just launch the device, it needs to learn how to win.

2018 chip industry is full of uncertainty

2017 is an impressive year for the semiconductor industry due to the fact that the three end markets: PC, smartphones and data centers outperform 2016.

At the same time, we have noticed that in 2017, the driving force for rapid development of the semiconductor market around the world is also on the rise. Artificial Intelligence and the Internet of Things are becoming the driving forces behind the growth of the market.

Specifically, markets such as autopilot and industrial internet of things are on the rise, with unparalleled demand for storage solutions, while artificial intelligence places high demands on high-performance processors.

Memory market

In 2017, the storage market is growing faster than the entire semiconductor market. In the past year, the storage market increased by 80.5%, while the semiconductor market grew by only 35%. At the same time, memory prices have risen dramatically in 2017 due to demand for DRAM out of supply by mobile and server equipment.


In 2017, Samsung, SK Hynix and Micron are the three largest players in the DRAM market. According to statistics from DRAMeXchange, the overall revenue of the DRAM market in 2017 increased by 16.2%, reaching a record high of US $ 19.2 billion in the third quarter of 2017.

Among them, Samsung’s DRAM business revenue increased 15.2%, reaching 8.8 billion step. DRAM business revenue of SK hynix increased by 22.5, reaching 5.5 billion U.S. dollars. Micron’s revenue increased 21% to 4 billion U.S. dollars.

Analysts ICInsights said Samsung, SK hynix and Micron, the top three companies in the DRAM market, accounting for more than 95% market share, but also makes these companies in the 2017 semiconductor company revenue rankings in the top five.

Mika Kitagawa, principal analyst at Gartner, predicts that DRAM out-of-stock will continue until the end of 2018, meaning that Samsung will become a semiconductor company with revenue over Intel’s revenue in 2017 and 2018, respectively.


DRAMeXchange data show that NAND market revenue increased by 14.3% in the third quarter of 2017 reached a record 15 billion US dollars.

Although the price of NAND is stable, the demand of NAND and SSD has been strong due to seasonal factors, setting a record high.

In the third quarter, Samsung, Toshiba and SK hynos increased by 19.5%, 18.1% and 15.4% respectively.

Walter Coon, director of NAND market research at IHS, said the NAND market will see weak and seasonal demand from the end of 2017 to early 2018, but the demand in the 3D NAND market is still strong.

As the market gradually realizes a balance between supply and demand, the rise in memory prices will not be the norm, especially as vendors continue to increase capacity.

Because suppliers want to earn long-term profits, they must comply with market rules.

As the IoT market continues to evolve, memory needs will also grow. Expected, the memory price will be moderate correction in 2018, but not a substantial decline.

Micron and Samsung in 2018 will continue to see high DRAM prices?

The biggest difference between DRAM and NAND is that DRAM is inelastic in demand, which means that demand is not affected by price changes, and it is difficult to limit its demand even now that DRAM prices are high.

The advantage is that PCs and mobile devices must use a certain amount of DRAM to achieve the expected performance, making DRAM a necessity in electronic devices.

Therefore, the shortage of DRAM products in PC and the increase of DRAM usage in mobile devices are also one of the reasons that DRAM continues to rise since June 2016. In turn, Samsung, SK Hynix and Micron are among the top five companies in revenue.

As more and more graphics-intensive applications in smartphones emerge and more and more adopt new technologies such as machine learning in data centers, there is a growing demand for memory. In particular, the graphics cards used by AMD and the memory used in some smartphones have caught up with mid-range computers, all of which indicate that the demand for DRAM will not decline in 2018.

DRAMeXchange predicts that by 2018, global DRAM supply will increase by 19.6%

The stock situation will be accompanied by phase.

The reason for this is that DRAM suppliers have always followed the rule that most of the money should be invested in technology rather than for capacity expansion.

Although these companies are now building a new DRAM manufacturing plant, but from the current progress point of view, by 2019 on the line is unlikely, so all three companies in 2018 will maintain a high revenue.

In particular, Samsung is currently focusing primarily on high-value products such as 10nm DRAM and DDR4X.

2018 NAND market will happen?

We mentioned before, DRAM market demand is inelastic, but the demand for NAND is flexible. So naturally, as NAND prices rise, market demand will decrease.

What’s more, the shortage of NAND market in 2017 is mainly due to the transfer of technology.

Another reason is that with the exception of Samsung and SK Hynix, NAND players are more involved and more competitive, and Intel, Western Digital and Toshiba are both involved in this market. Therefore also makes this market competition more fiercely than the DRAM market.

According to data from DRAMeXchange, last year, Samsung, SK Hynix and Toshiba saw growth of around 15% in this market.

NAND market 2018 outlook

For the 2018 NAND market, analysts predict not exactly the same.
Some believe the NAND market will be in balance by 2018, while others believe there will be oversupply in this market by 2018. The reason for this concern is that many suppliers are mass-producing 64-layer or 72 Layer NAND way to expand production capacity.

DRAMeXchange expects NAND production capacity to grow 43% by 2018 while market demand will only increase by 38%. Obviously, this will create oversupply which may result in a price drop.

At present, some signs have even been seen on the market. In the first quarter of 2018, the average sales price of Micron’s NAND products dropped by 3%, which is obviously not a good sign.

However, the impact of this indeed for Samsung will be much smaller, because Samsung has great technical advantages in 3D NAND, but still does not rule out future competition from Micron pressure.

Micron, 2017 earnings milestone

Like Samsung, in 2017, Micron’s revenue also hit a new high due to the memory market.

Specifically, Micron’s financial reports have experienced double-digit increases for six consecutive quarters.

Micron’s 2018 fiscal first-quarter report, ending in November 2017, shows that Micron’s revenues increased 11% to 6.8 billion, up from analysts’ expectations of 6.3 billion.

The main drivers of this growth are the strong demand in the DRAM market, the steady growth of NVIDIA graphics cards, and the impact of data centers and other factors.

In addition, Micron contributed record financial performance to the SSD products used in the data center.

After entering 2018, many analysts are predicting the future memory market.

Many believe the consumer demand will be met as suppliers’ availability increases. However, it is worth noting that as mentioned earlier, since the new capacity required in 2018 will require at least two Year’s time to climb, memory prices will remain high in the short term.

What’s more, there will not be a cliff-like decline in any recession, especially with the advent of artificial intelligence and more IoT products in 2018, which will further expand the consumer-grade memory market. For example, we have seen NVIDIA graphics cards, Tesla cars have made a lot of memory requirements.

Analysts believe that these will further promote the growth of Micron’s revenue, the automotive market, the car memory is a dedicated memory, the price will be higher.

With the growth of market segments such as automobiles and data centers, analysts believe Micron will continue to grow in these markets.

What will happen with the 2018 Micron?

As we said earlier, the DRAM market is a market with inelastic demand. In 2018, the DRAM market will continue to grow. Micron, the forerunner in the market, is also pushing the next generation of new technology.

In the first quarter of fiscal 2018, Micron’s 1X DRAMs have achieved relatively mature production volume and are expected to surpass the 20nm of the previous generation by the end of 2018.

However, due to new technology inputs, Micron said that this year Micron DRAM market growth rate will be lower than the market average.

On the other hand, at CES 2018, we have already seen some clues about DDR5 technology in the future. Although this product will be hard to be commercialized by 2020, it will consume more than its predecessor Low power.

In terms of NAND products, Micron expects to launch the second generation of 64-layer 3D NAND products in the second half of FY2018 and trial production of third-generation 3D NAND products in the second half of 2018. Although Micron did not disclose the details of its third-generation 3D NAND products, but know that Samsung’s third-generation 3D NAND products are 72 layers.

Micron believes that its NAND product growth rate will be higher than the market average growth rate of 50%.

From the above situation, we can see that Micron has been continuously investing and updating products in both DRAM and NAND products.

NVIDIA and the semiconductor industry trends

At the industry summit held in January 2017, there have been opinions that the Internet of Things, autonomous driving, 5G, AR / VR and artificial intelligence are the future directions of development. These technologies will bring about 30 billion U.S. dollars by 2025.

In addition to 5G, NVIDIA has long been concerned about the above several other technical trends. In fact, the GPU it produces currently has great advantages in the areas of artificial intelligence and autonomous driving, and NVIDIA is also investigating the possibility of GPU applications in the Internet of Things, and in particular the Industrial Internet of Things.

A report released by Jon Peddie Research released a detailed analysis of GPU applications on different terminal platforms. From the figure above we can see that mobile devices and computers are still the largest GPU application market.

At present, NVIDIA, AMD and Intel are the three main GPU suppliers. Among them, Intel’s shipments accounted for 68% to 70% of the overall market, but the GPU it provides are mainly integrated in the processor used by the PC.

NVIDIA GPU is a separate GPU, accounting for more than 70% market share.

Independent GPUs provide consumers with better graphics performance, whereas Intel GPUs can only handle relatively simple graphics.

In the past few years, NVIDIA has been widely applied to its independent GPU in addition to the PC market, artificial intelligence, intelligent driving, Internet of Things and industry and other market segments.

Many analysts believe that now NVIDIA is at the cusp of the fourth technology transfer. There are many possibilities for development, and even compared it to Apple’s ten years ago.

In the past year of 2017, we have seen some areas such as artificial intelligence and e-sports propelled NVIDIA revenue growth. These growth points will continue in 2018, meanwhile, the new growth point of autonomous driving may also break out.

It can be said that 2017 Nvidia is invincible, and 2018 also has a lot of competition.

In the past three years, NVIDIA’s revenue more than doubled from $ 4.7 billion in 2015 to $ 9.7 billion in FY 2018. This growth is also the NVIDIA into the world’s top ten semiconductor companies ranks.

In addition, the strong growth momentum of NVIDIA also comes from the data center. In this market, NVIDIA has also enjoyed a three-digit increase in the past six quarters. In addition, NVIDIA has witnessed the increasing popularity of electronic sports. NVIDIA also needs a better discrete graphics card to provide higher performance. The trend is also for NVIDIA to provide more and more customers with more detailed products.

NVIDIA’s latest Volta GPU architecture has received strong customer reaction. This product also contributed to GPU revenue in the data center. In 2018, NVIDIA revenue in this market reached 20%.

NVIDIA will continue to grow in 2018, but due to pressure from competitors and growth rate.

For now, the biggest competitor of NVIDIA is Intel, which is seeking cooperation with AMD in an attempt to enter the independent GPU market.

In addition, Intel’s processor is also used by Google for autonomous vehicles.

However, for NVIDIA, it is delightful that in 2018 we will see the launch of NVIDIA-equipped vehicles on the stage, and NVIDIA will launch its own car platform to help drive the development of autonomous vehicles.

In 2018, how the PC market will affect semiconductor companies

From the above, we can see that, whether it is Micron or NVIDIA, the PC market in its revenue still occupy not a small proportion. Intel and AMD also occupy a significant share of these two markets.

2017 is a good year for the PC market, where the slowdown in PC market sales has finally slowed down. Although the traditional desktop market is still declining, sales of laptops and gaming computers have increased rapidly. Although the forecasts for the PC market from Gartner and IDC are different, they both expressed similar views.

For 2017, Gartner expects global PC shipments to drop by 3.2% year-over-year to 425 million units. However, the data given by IDC dropped 2.7% to 423 million units.

IDC said that although the memory out of stock, but the performance of the traditional computer market has more than expected. In 2018 the situation will be better.

Gartner expects PC shipments to increase by 0.8% to 426 million units by 2018, while IDC believes the PC market will continue to decline, down 4% from a year ago to 400 million units.

AMD 2018 can maintain growth

When it comes to GPU, we have to mention AMD.

As Intel in the field of CPU, NVIDIA’s largest competitor in the field of GPU, AMD has seen significant growth in the past two years.

Finishing AMD The first quarter of 2016 to the latest quarter can be found, in fact, AMD from the second quarter of 2016 has been the continuous growth of performance, turnover is all the way to sing. In terms of profit, though, the company lost a huge loss of 109 million U.S. dollars in 2016, but by 2017, the company started to turn around again.

In terms of products, AMD has released various new products and consumer terminals frequently in the past two years. Not only AMD Rui Long 3 Series, 5 Series and 7 Series, but also high-end Threadripper processors are firmly in the hands of different consumers The demand. In terms of graphics cards, it launched Vega, focusing on new areas such as AI, game design and visualization. In the server market, released EPYC, return to the data center market; also introduced Rui Long mobile processor.

AMD expects fiscal 2017 revenue growth of 22.5% to 5.2 billion US dollars, while analysts expect the chip maker in fiscal year 2017 revenue up 23%. However, analysts expect this growth to slow to 12.5% ​​in FY 2018.

In December 2017, AMD partnered with Intel to provide a discrete GPU for the latter’s gaming mobile PC platform. This partnership may increase Intel’s PC processor market revenue. By combining its GPUs with Intel’s processors, AMD can easily increase its GPU market share and compete with NVIDIA.

In 2018, AMD also plans to launch a lot of new products, so AMD can really survive in 2018 Intel’s competition?

2017 is a very busy year for AMD.

In the just-concluded global technology event CES, AMD introduced a variety of heavyweight products, and announced a substantial price plan. Ryzen 7 1800X high-end price cuts by 40%. Significant price cuts over Intel processor vulnerabilities, comprehensively enhance the company Ryzen and Threadripper chips and Intel chips of similar competitiveness, will help grab a certain market share in the future.

Also at CES, Dr. Su Tzufeng, AMD’s CEO, mentioned that Ryzen products provide the data processing power that rivals the Intel Core i5-8400 processor and the NVIDIA GT 1030 graphics card, but at a cheaper price of $ 120. This will be the good news for those who want to save money but are not willing to sacrifice their computing power.

AMD is operating in a duopoly market. Its goal is not to surpass its competitors by introducing state-of-the-art technology, but to provide competitive products that offer consumers cheaper options and help them gain market share.

In 2017, AMD foundry partners announced a 12-nanometer technology node. So, AMD may upgrade its Ryzen series from 12nm to 7nm.

At present, AMD has provided a 12nm upgrade version of the Ryzen desktop processor sample, which was released last year, was using 14nm process technology. The chip will be put into production from April, using Globalfoundries upgrade process announced last year. At the same time, Ryzen 2 design has also been completed, but AMD did not give details of internal or shipping date.

AMD’s first 7nm GPU will be a version of its Vega design aimed at machine learning. AMD will catch up with its rival Nvidia on the GPU side, which dominates this nascent field after the launch of Volta in 2017, with Volta becoming its first chip with machine-learning hardware.

Why 2017 is important to Intel

Finally, let’s take a look at Intel.

Although, in 2017, Samsung took the No. 1 spot on storage surpassing Intel, Intel did not sit still.

It is working hard to expand its business area to capture several other markets: the Internet of Things, 5G, AR / VR and Artificial Intelligence.

After the acquisition of Altera, due to strong growth in the data center, Internet of Things, etc., Intel’s revenue in 2017 relative to 2016, is expected to grow 4.4%.

As the world moves toward intelligence and the Internet of Things, Intel is working with its rivals in areas it believes are mutually beneficial.

In addition to the previously mentioned cooperation with AMD, Intel also announced several technical aspects of the car and 5G.

Intel is currently transitioning from a PC-centric to a data-centric company. 2016 and 2017, Intel’s corporate structure has undergone significant changes. 2018 More changes can be seen in the company’s strategy.

Wall Street analysts expect Intel’s fiscal year 2018 revenue rose 3% to $ 63.8 billion. However, this growth estimate is based on the prevailing market conditions and will vary with changing trends.

Intel currently offers GSM (Global System for Mobile communications) to provide modems for Apple’s iPhone handsets. The company is developing its own CDMA modem and expects to offer CDMA modems for the iPhone by 2018.

Intel is also competing with Qualcomm in the 5G competition. Both companies launched their 5G modems in 2017.

In addition, Intel is working with more and more automotive manufacturers on the development of autopilot programs that are dedicated to auto-emblem and state-of-the-art AI (Artificial Intelligence) technology.

Intel 2018 will be what kind of development

In addition to the new market, Intel can continue to grow in the PC market.

2018 You can see the long-awaited launch of the 10-nm node, which will make Intel a leading technology provider.

Intel demonstrated the performance of the 10-nanometer Cannon Lake desktop processor at CES 2017, but did not reveal information about the 10-nanometer node. In fact, Intel introduced the 14-nanometer eighth generation Coffee Lake processor by the end of 2017.

Now, Intel has changed its product roadmap, the node transition to become longer.

According to Intel’s original plan, the production of the old node will no longer be made when the conversion to the new node is completed. However, in 2018 we may see two nodes coexist.

At the end of 2018, Intel is likely to launch the Ninth Generation Cannon Lake or Ice Lake CPU at 10nm, which will co-exist with the 14nm Coffee Lake CPU.

If Intel further postponed its launch of 10nm, it will lose its technical advantages.

Intel’s Artificial Intelligence Innovation

As the global leader in the PC industry, Intel is committed to promoting the development of the traditional PC industry through technological advances on the one hand, and has gradually identified a number of key players including PCC Smart, autonomous driving, precision medical, 5G and other force direction.

In these new business areas, Intel is most likely to achieve increased chip business capabilities with artificial intelligence and autonomous driving.

Focus on autonomous driving, March 13, 2017, Intel announced the acquisition of Israeli company Mobileye for 15.3 billion US dollars. Mobileye’s algorithms and data are likely to become the monopoly market giants in the more subtle, core technology-based automotive market. Data shows that Mobileye’s technology has infiltrated almost into the mainstream automobile factory, and Mobileye products are used by automobile manufacturers such as Tesla, BMW, GM, Volvo and Ford.

In the area of ​​artificial intelligence, Intel’s $ 350 million acquisition of Nervana Systems, a deep learning company, in August 2017 was seen as a catalyst for Intel’s AI strategy. In order to promote artificial intelligence performance bottlenecks and technological popularization, Intel has established alliances, including Google and other giants, and at the same time, it cooperates with global organizations to provide developer training courses so as to build an ecosystem involving artificial intelligence technology upgrading, education and training, application optimization and the like .