With Software Eating Hardware, Silicon Valley Enters “Hard” Times

Next Story

Community Will Get Its Sixth Season… On Yahoo.

Software’s inevitable dominance is something of an axiom in Silicon Valley, where Marc Andreessen once famously wrote that it was “eating the world.” Software companies like Microsoft, Google and Facebook are among the world’s most iconic and valuable, and new startups like Airbnb and Uber aim to transform traditional industries like hotels and taxis.

For some 30 years, software has been the one iron (and golden!) rule of building a startup. These sorts of business models have been unmatched as vehicles for founders and venture capitalists, since companies can rapidly iterate on their products and massively scale when product-market fit has been found. Most importantly though, software is profitable like almost no other industry, with margins for some companies reaching well beyond 50 percent of revenue.

In such an environment, hardware startups became something of a backwater for Silicon Valley. Venture capital investment into hardware startups has traditionally hovered around the low single digits according to the National Venture Capital Association. Hardware startups are considered “hard,” requiring huge gambles on design, production quality, and supply chain management that software startups simply don’t need.

That rule has been entirely thrown out. Hardware is now in vogue at software companies, a strategy once considered anathema. The last 10 months have been a record period for exits from the hardware startup space, many through acquisitions by software companies. While it was once a commonplace that hardware is a drain on profits, margins and resources, today the leading software companies are throwing their weight behind their physical brethren.

What explains this sudden interest in hardware after so many years? There are a number of factors, but three would seem to be crucial in the change in heart. First, the culture of hardware design has become more like software coding, making integration post-acquisition easier. Second, companies are under pressure to provide the best user experience, and as a result, they are becoming more vertically integrated. Finally, competition for customers is becoming more intense, and hardware is one way to ensure consumers stay with one company.

A Heck Of A Year For Hardware

This is a banner year for startups in the hardware space, which has already seen several multi-billion-dollar exits. GoPro, the maker of a popular line of highly versatile cameras, had a sterling first day of trading on the NASDAQ, with its price jumping even after pricing at the top of its target range. Beats Electronics, the maker of a popular line of headphones, was purchased for $3 billion by Apple, the top hardware manufacturer in the world.

But the truly interesting pattern has been the increasing engagement of software companies in the space, with Google being the prime example. While the company has produced a yellow-boxed hardware version of its Google search engine for many years, today it’s key products are all software: Google Web Search, Gmail, Android, and Google Docs, while its revenues come predominantly from advertisements. In short, it is the very epitome of the Internet software business.

Yet, the company shelled out $3.2 billion for thermostat-maker Nest Labs, its second-largest acquisition. Just this month, the company bought two hardware businesses – Dropcam (through Nest Labs) and Skybox Imaging – for about $500 million each. And while the company has since sold off Motorola Mobility to Lenovo, that acquisition in 2011 remains its largest in history at $12.5 billion.

Internally, the company seems just as focused on building up its competency in hardware. The “moonshots” that are coming out of its GoogleX laboratory include such hardware products as Google Glass, Loon balloons, and self-driving cars.

Google is hardly an exception among top software companies. Microsoft’s $7.2 billion acquisition of Nokia last year is one of the company’s largest acquisitions in history. Facebook’s second largest acquisition to date was for virtual-reality startup Oculus VR earlier this year, a startup less than two years old that received a cool $2 billion check. Amazon continues its long investment in hardware, recently launching its Amazon Fire phone.

With billions of dollars spent acquiring hardware businesses, it is obvious that software companies are more engaged with hardware than ever before.

Hardware Is Becoming More Like Software

One challenge for software companies looking to acquire hardware businesses has been the dramatically different culture between coding and designing physical goods. Hardware products have had long gestation periods, requiring multiple stages from conception using computer drafting tools to prototyping and readying a product for mass production. If an issue cropped up, it would take weeks or even months to repeat the entire process again.

The huge upfront investment in a new hardware project also meant that companies generally needed to conduct market research before they began building, an approach completely contrarian to the Silicon Valley notion of launching software first and then using feedback from the market to evolve a product.

Thankfully over the last decade, hardware design has become easier and more democratized. 3D printing and better software tools have accelerated the ability of many startups to build their prototypes, allowing them to speed up their product iterations. Early hardware startups now often feel like lean software businesses, with much more interactive product development, thus allowing for more rigorous analysis of product-market fit.

In the past, when it came time to manufacture a product, only the largest companies used to have access to assembly lines in places like China. Minimum order sizes were often completely unaccommodating for startups. But as competition has intensified, custom manufacturing has become easier, and labor costs have risen, manufacturers appear to be more willing to engage with smaller orders, and they have also decreased their average turnaround time, as well.

Beyond just design, hardware startups also have a greater role in software than in the past. From what I can tell talking to hardware startups, it is much more common for product managers, marketers, and engineers to be cross-functional these days, working across the traditional divide between hardware and software. It’s the only way to create a seamless experience for the user, but along the way, it also creates a culture that is more compatible with potential software acquirers.

Software Is Feeling a Bit More Vertical

Verticalization could also be called the Apple-ification of software companies. Apple’s traditional strategy has been to sell vertically integrated technology, owning the complete user experience from hardware to software and services. By using its control, the company can ensure that all of its components work smoothly together, guaranteeing a level of quality that is difficult to emulate. Apple’s strategy has been unique in the industry, although this status is changing as Microsoft and others increasingly use this approach, especially in mobile.

While there are certainly economic reasons why software companies desire more control over their products, I think one of the lesser-discussed issues is simply that the sorts of engineering problems encountered by companies today are becoming more complex.

Consider the case of home automation. There have been many attempts in the space over the years, such as products around the X10 and KNX protocols. But solutions always faced the same vexing challenges – how to make all the sensors and control switches part of one comprehensive system, with easy setup and equally well-designed software. In addition, home automation is not currently a well-developed market, and so a company has to actually convince users on the merit of setting up one of these systems in the first place. It seems obvious that a verticalized company with a complete and well-developed product line on both hardware and software and excellent marketing is the only way this sort of product will be successful.

There is a class of problems marked by deep integration between hardware and software that is simply tough to do with multiple companies involved, especially in the timeframes required by the market today. There can also be conflicts over incentives in such collaborations as well, since a hardware manufacturer may prefer a smaller market share but higher margins while a software company wants greater user adoption to generate traffic. It’s much easier if everyone is paid from the same employer.


Perhaps the most important reason that software companies are increasingly interested in hardware is that owning the device of a user is much better lock-in strategy than being their first destination on the web.

Think of a product like Google’s Web Search. Users can switch to a competing site in just a few keystrokes, whether it be Bing or DuckDuckGo (or by switching their default search engine in their browser preferences). The same concerns are visible in Facebook’s competition with Snapchat over mobile messaging. While brand and habits are tough to break, software companies are always vulnerable to a new incumbent taking their market share.

Now take a hardware product like Nest. If you bought a new thermostat from the company, what is the likelihood of replacing that thermostat anytime in the near future? Even in my most progressive thinking, I can’t imagine that a thermostat is going to be regularly upgraded like a smartphone. Furthermore, it is highly unlikely that anyone is going to use two thermostats at the same time. Essentially, once Nest has made a sale to a customer, it owns that customer for an extended period of time.

The integration of hardware, software, and services in one complete package through verticalization not only creates a better user experience, but also means that customers who stray too far from one company’s products are likely to find it difficult to get everything working properly. By building this sort of sticky lock-in into its products, software companies can ensure more consistent revenues and also raise the switching costs for customers.

The Pendulum Swings Back

In addition to culture, verticalization and competition, there is one more important trend working deeply in favor of more hardware startups. Consumers are adopting hardware technology at incredibly rapid rates. Mary Meeker’s Internet Trends Report has a beautiful slide on this phenomenon, showing the acceleration of each new wave of technology, starting from the personal computer and continuing through tablets. With the expanding universe of customers around the world, companies can now scale their revenues up nearly as fast as software has done in the past.

So, hardware is cool again. In some ways, this is simply the pendulum moving back again. Hardware was once the bread and butter of Silicon Valley, the very name of which continues to evince the importance of this lineage. All of Silicon Valley’s early successes were physical, first in radar and klystron tubes, then later in the integrated circuit, semiconductor and personal computer. And of course, some of the venture capital industry’s most important early bets were on hardware companies like Apple, Tandem Computers, Atari and Intel. But then the software industry took off, and no one looked back.

There is probably still more risk in hardware than software today, but there is also less competition, and far more low-hanging fruit from the lack of startups over the past 30 years. For founders willing to engage, 2014 might just be “hard” times.