Daily Tech Snippet: Friday January 23

• Amazon buys an Israeli chip startup for $350-370M; Annapurna Labs makes chips optimized for the cloud business: (More coverage here and here) Amazon agreed to acquire Israeli semiconductor company Annapurna Labs, seeking to improve performance within its Amazon Web Services cloud unit. A deal has been reached but hasn’t yet closed, said Mary Camarata, a spokeswoman at Amazon Web Services. She declined to comment on the terms. Acquisition talks were reported earlier by Israeli newspaper Calcalist, which said the purchase price could be as much as $370 million. Annapurna Labs develops microprocessors that allow fast data traffic for power-efficient computing and storage servers, according to the newspaper. Seattle-based Amazon, the world’s largest online retailer, is constantly seeking to improve the cost and performance of the equipment in its data centers and has been hiring more semiconductor engineers to add to its capabilities, James Hamilton, a vice president of Amazon Web Services, said in a November interview. The unit rents out computing power on its servers to businesses for access through the Web.

• Google is gearing up to sell wireless service directly to customers as a mobile virtual network operator (MVNO), by acquiring excess network capacity from Sprint and T-Mobile and reselling it to customers under its own brand. This is the same approach used by Cricket Wireless, MetroPCS, Pure Talk, Republic Wireless and many others in the U.S., but Google’s arrangement apparently required special consideration, according to The Wall Street Journal, given the potential threat network providers perceived in giving the search giant and Android maker too much control. Sprint built a volume clause into their agreement with Google, per WSJ, which triggers if the Google wireless service acquires a large number of users and lets Sprint renegotiate the terms of the deal. There’s no word on how T-Mobile’s arrangement works, but given comments about the U.S. carrier by its parent company Deutsche Telekom, the provider is probably looking for ways to shore up the sustainability of its prolonged “Uncarrier” campaign. More analysis here: What is Google planning? The company hasn't confirmed anything, but reports suggest that Google wants to offer its own brand of cellular service by partnering with Sprint and T-Mobile, the nation's third- and fourth-largest carriers, respectively. This means you'd buy minutes and data from Google, but it would all ride over the other two companies' pipes. Is this like Google Fiber for cellphones? Not quite. With Google Fiber, Google is connecting homes to high-speed fiber-optic infrastructure that, in many places, the search giant laid down itself. But in this case, it doesn't appear that Google is building its own cell towers; it would simply resell Sprint and T-Mobile under the Google banner. Would this shake up the wireless industry? Google probably hopes so, but unlike the wired broadband industry that Google Fiber has been so successful at disrupting, the wireless industry is actually fairly competitive already. You have four major national carriers that are engaged in a major battle over pricing and customers right now. You also have dozens if not hundreds of smaller carriers who operate on the same basis as Google's rumored plan — paying the national carriers a wholesale rate and then repackaging the service into a different product. These piggyback carriers are called MVNOs, short for mobile virtual network operators. Why would Google want to become a wireless provider? If there's one thing driving Google, it's a thirst for your commercial data. Think about all the data generated by your Google searches that gets scooped up and used for advertising. Now think about all the data you generate when you place a phone call: whom you're calling, for how long, what time of day and so on. This information is incredibly personal and can be used to help build a profile of you — which, if you'll recall, is partly why everyone was so outraged when they found out about the National Security Agency's snooping into phone records. Then there's the business information Google can collect about which data plans people find most attractive and how they use those plans. And naturally, all of Google's handsets would likely run on Android, so there's that software integration.

• Box prices IPO - implied valuation of $1.6B, 33% discount to last funding as Box struggles to rise from storage play to platform play: (More here) Box Inc. is set to price its initial public offering at a discount to its latest financing round, as stiff competition overshadows the cloud-storage company’s plans to expand into new areas. The IPO comes almost a year after Box, which lets businesses manage, store and have access to data over the Web instead of through onsite computers, first filed to go public. The company’s financials in its March prospectus underwhelmed some investors just as deep-pocketed competitors including Microsoft Corp. (MSFT) were entering the fray, and the IPO was delayed. Ten months later, Box’s losses have narrowed as revenue surged 70 percent in the most recent quarter, and the company is offering new services that could help it retain and increase its customer base. The company, led by 29-year-old Aaron Levie, is still valuing itself below the level it fetched in a July private-financing round, betting the combination of its turnaround plan and cheaper price will lure buyers. “Investors are either going to have to buy into that story that they can diversify their business, or that this is coming cheap and that there will be consolidation,” said James Gellert, the president and chief executive officer of Rapid Ratings International Inc., which uses quantitative models to grade securities. Box, based in Los Altos, California, is seeking as much as $163 million from the IPO, scheduled to price Jan. 22. It is offering 12.5 million shares for $11 to $13 apiece, according to a Jan. 9 filing with the U.S. Securities and Exchange Commission. Those terms indicate a valuation as high as $1.6 billion, reflecting a 33 percent discount to the $2.4 billion Box was valued at during a private round by investors including TPG Capital and Coatue Management LLC. The market for file sharing and synchronization software is forecast to grow 23 percent a year on average to $2.3 billion in 2018, market researcher IDC said in a report looking at 2013. The enterprise side of the market, where Box fits, is growing slightly faster at 27 percent a year. Box had 14 percent share in the overall market, behind Dropbox Inc. with 27 percent and Microsoft with 17 percent. The challenge for Box, said IDC’s Maureen Fleming, is that Microsoft has been dropping prices for its OneDrive and including it free for customers who are signing up to use its Office apps online. That means many corporate customers decide to at least try OneDrive, curbing Box’s growth rates. Fleming regards Box as superior to OneDrive for companies looking for cloud-based systems but said product quality doesn’t always matter to chief information officers in the market for software. In response to those price wars, Box has created ancillary services on top of storage to differentiate its offerings. They include programs to help companies to build custom applications as well as security features that can be customized. The company also has a consulting arm, which helps customers navigate its services. Said Eli Lilly CIO Mike Meadows: “They are doing the right things and saying the right things,But we haven’t come across the use case that would drive us using them in more of platform mode rather than just a storage mode.”

• Reddit's Pick for New Ad Leader Shows the Site Is Getting Serious About Mobile Hires Zubair Jandali away from Google. The company hired Zubair Jandali from Google, who started this month as Reddit's first vp of sales. Jandali had been with Google since 2009 when the company bought ad network AdMob, where he also worked. His hiring is yet another sign Reddit is serious about developing a mobile advertising presence. Reddit recently bought Alien Blue, the app that lets readers access Reddit on mobile, to hone a more sophisticated approach to smartphones and tablets. Alien Blue is Reddit's first in-house app. Now, it has a vp of sales with experience in mobile advertising to go along with the product. "Mobile is definitely one big area of investment for the company, so my background certainly plays into that," Jandali said in a phone interview this week. Reddit offers brands two ways to advertise on its site, which reaches 150 million people a month: banners and sponsored headlines. Headlines are the items that users and brands post to Reddit, and users can vote up or down on whether they like them. The user-submitted headlines that accumulate the most positive votes rise to a higher position on the page, but brands pay for top placement. There are over 8,000 "active" user-generated groups on the site called subreddits dedicated to different categories; they can attract fans of brands, movies, science fiction or just about any interest you can imagine. Advertisers can target a subreddit or buy ads on the front page of the site, which gets 50 million views a day, Jandali said. For instance, on Wednesday TBS promoted a headline on reddit.com at the top of the page for the show King of the Nerds. Jandali said visitors who click on sponsored headlines can spend up to five minutes in the posts. Jandali reports to CEO Ellen Pao, who took the role after a shake-up late last year that saw the exit of the former CEO and the return of co-founder Alexis Ohanian to board chairman. The ad sales teams are based in New York, Los Angeles and San Francisco.

• Twitter pleads with power users not to use Instagram: Twitter appears to be sending out a message to a group of very high-profile users suggesting that these users post photos directly to Twitter instead of sharing through Instagram. Mashable secured a screenshot of the prompt, which shows the aesthetic differences between sharing an Instagram link and posting a photo directly through Twitter. In 2012, Instagram shut off Twitter Cards integration, meaning that the images would no longer appear in-line on the Twitter feed. Instead, Instagram photos shared out to Twitter would simply show as an Instagram.com link and push the user to the website for viewing. At the time, many people were upset that Instagram would interfere with the user experience for Twitter. It’s also worth noting that parent company Facebook has always displayed and still proudly shows Instagram images right in the feed. The original backlash over Instagram shutting down Twitter integration has long been forgotten, and in many ways, Instagram has won. In December, Instagram announced that it had surpassed 300 million active users, surging past Twitter’s 284 million active users. Twitter, which is older than Instagram, has struggled to keep up with the young gun in the media department. Twitter has released countless updates to the web product and the mobile apps to try to bring more attention to images and videos. It added a tab under user profiles to view media only, and put more focus on images that are used on profiles, like the addition of the Facebook-style header image and the now-larger profile photos. The company also added photo filters and editing tools long ago to deal with Instagram’s exodus from the feed in 2012. This latest message from Twitter to its mega-users, asking them to post photos from the app itself, only shows the severity of that struggle as consumers hungrily consume more and more multimedia content.

• Button, a mobile start-up, looks to deep link apps with commerce: It took years for the web to become what it is today: a sprawling, interconnected network of sites endlessly linking back and forth to one another. It will likely take years for the same thing to happen to smartphone applications, which are mostly stuck in silos, disconnected from one another. One New York-based start-up believes it can help speed that process. “The mobile app world is so very fragmented that for users, it makes interactions between apps much more difficult than they should be,” said Michael Jaconi, chief executive of Button, which announced a $12 million round of venture financing on Thursday morning. “Our thesis is, can we connect the mobile economy in a smarter way?” Button’s value proposition is largely technical: The company wants to create the plumbing behind what’s known as deep-linking between apps that make sense to be connected. So for example, if a customer books a table at a restaurant using Resy, a restaurant reservation app, Button could suggest you book a ride to the restaurant using Uber, the ride-hailing service, and link the customer directly to the Uber app to request a ride. That connection may come with a commerce-based incentive — say, a $30 credit toward an Uber ride. Button makes its money based solely on the amount of traffic it drives to its app partners. The idea, Mr. Jaconi said, is similar to what Google did for search more than a decade ago. When a user wants to find something specific, they type in a search query and Google serves up a list of suggestions, along with a series of advertisements. Google has made billions capitalizing on what technologists call the moment of intent — that is, the exact time when a person wants to see a paid advertisement. Button believes it can pinpoint that moment of intent inside apps, a problem that Google has yet to crack. “The future, we think, is one where we’re no longer living in a world of typing, but rather in a world of taps,” Mr. Jaconi said. “Using your phone, we’ll be able to infer what you want to do next based on your location, the time of day, the context of the other apps you’re using.” “It’s increasingly difficult for app developers to rise above the noise and acquire users, and then re-engage these users over time,” said Chris Moore, a partner at Redpoint Ventures, a venture capital firm that led Button’s recent funding round. “Button enables a new channel for user acquisition and engagement.” Button is not the only company taking on deep app links. Facebook wants its deep-linking solution, announced last year, to become the standard for developers. And start-ups like Branch Metrics, URX and Quixey offer different approaches to the same problem as well.