Description

Investment Thesis:

Xunlei Limited (XNET) is a relatively undiscovered gem in the coveted Chinese internet arena that offers U.S. investors a rare opportunity to invest alongside the “Steve Jobs of China” (Lei Jun) at a very attractive price.  Jun is the founder, CEO and 78% owner of privately held Xiaomi, the world’s fastest growing smartphone maker with an estimated value of $45 billion based on the latest funding round at the end of 2014.  Xunlei, of which Lei Jun indirectly controls 41% through Xiaomi and Kingsoft, is a turnaround story, and Jun is just the man for the job (see section below to find out why).  But what makes this stock especially interesting is that even in a muddle-through scenario, there is significant upside potential.  With nearly two thirds of the market cap in net cash ($7.00 per share in cash including proceeds from the pending sale of the online video business), the investor is paying under 5X what we think the company will earn in 2016 ($0.75 per share).  Given the growth of Xunlei’s mobile installed base, the valuable Xiaomi partnership, and the cash generative (FCF positive in each of the last 6 years), asset light business model, we think the stock could justifiably trade at a P/E of 12X which equates to a fair value of about $15 per share (40% upside).  But that’s not the real story; this is Lei Jun’s latest project, and to attempt to grasp the implications, we need to understand why he has been crowned the Steve Jobs of China.

The Legendary Lei Jun:

Unlike Jack Ma, the high profile founder of Alibaba who is now a household name in Silicon Valley, Lei Jun rarely travels outside China and is a relative unknown to most Westerners.  Lei Jun actually made his name earlier in his career at a firm he co-founded in 1992, Kingsoft.  Jun was instrumental in transforming Kingsoft from a struggling software company to a mobile pioneer, rolling out mobile internet applications in all of its mature businesses including online games, antivirus and office software.  From the time he rejoined the firm in 2011 (after a 4 year hiatus, which furthers the Steve Jobs comparison) to the present, the stock is up about 8X and now has a $5 billion market cap.  Jun helped orchestrate the spinoff of Cheetah Mobil (CMCM) in May 2014 (7 weeks before Xunlei’s IPO), and had a hand in its rise to one of the largest internet security software providers in China.  In just over a year post IPO the stock is up nearly 2.5X and now sports a $4 billion market cap.  Just to prove he is not a “one trick pony” (or two or three for that matter), Jun also co-founded the Chinese version of Facebook, YY Inc (YY).  YY is up nearly 6X since its IPO in 2012, and Jun’s $1M angel investment in 2005 is worth over $600M today with the current market cap at $3.5 billion. 

The Xiaomi Ecosystem:

So we have established that Lei Jun is the real deal, but what is the deal with Xiaomi?  The more we peel back the complex layers of Xiaomi, the more we begin to appreciate the genius of Lei Jun.  To the casual observer, Xiaomi is simply a low-cost smartphone company, and one that arrived late to the game (their first phone hitting the scene around the time of the iPhone 4S).  But as we look closer, we sense that their ambition appears to be much “grander”.  Like Google, the company values open source concepts: its Android operating system, MiUI, is available for other manufacturers’ Android devices.  Xiaomi has made TVs (MiTV), routers, air and water purifiers that all tie into MiUI.  Unlike Apple (Homekit), Samsung (SmartThings), and Google (Nest) who are offering software development kits so their customers can tie their own appliances into Android or iOS, Xiaomi is integrating everything itself and selling everything a customer needs on Mi.com.  So in some ways, it is the world’s first vertically integrated provider of the “Internet of Things”.  The level of devotion that Lei Jun has inspired in his customers was captured well in a recent The New York Times profile:

“Li Nan, vice president of the rival Meizu, which began in the early 2000s by making digital music players and aims at customers slightly older and wealthier than Xiaomi’s, likens the devotion of Xiaomi supporters to a religion.  “Xiaomi fans have a high level of organization,” he said. “They love Xiaomi. It’s a form of idolatry.”

So in a sense, Xiaomi is taking a page out of Starbucks’ book: they are selling an experience.  Xiaomi wants to sell everything a person needs for their first house. 

Xiaomi Ecosystem Ownership Table

Xunlei’s Business Model:

The next logical question is where does Xunlei fit within this powerful ecosystem?  Let’s start with the core business.  Xunlei is a top 10 internet service provider in China with over 300 million monthly unique visitors.  The company operates a powerful internet platform based on cloud computing that provides users with quick and easy access to digital media content.  Its core acceleration services, including the top accelerator product in China with 84% market share, help users accelerate digital downloads of media files or file transmission over the internet (think of this as the Akamai of China).  China has the largest internet user base in the world with 650 million users as of year-end 2014 (growing at about a 10% CAGR over the past 5 yrs), yet the average peak connection speed was just 17.8 megabits per second, which ranked 95^th globally in 4Q14 (Sources: CNNIC, Akamai).  China still has a long way to go, with an internet penetration rate at just 48% of the population in 2014 vs 95% in Sweden, 90% in the UK, 90% in the UK and 84% in the US.  Xunlei is well positioned to capitalize on China’s growing appetite for high speed internet access, and it has expanded into other online digital content access and consumption services including online video (Kankan) and other internet value added services (IVAS, mainly online games). 

Xunlei Accelerator is generally free of charge, while Xunlei also offers monthly or annual premium cloud acceleration subscription services, which serve as the major revenue source.  Since the introduction of paid subscription service in 2011, the number of paid users has leveled off at about 5 million, pressured by the Chinese government’s scrutiny on internet content (ie its anti-pornography campaign).

Portfolio Shuffling: In September 2014, Xunlei purchased a cloud-based storage business called Kuaipan from Kingsoft for $33 million.  Similar to Microsoft’s OneDrive or Google Drive, the targeted users are individual PC users and/or smartphone users, and this is a natural extension of Xunlei’s core accelerator product.  

On the other side of the ledger, Xunlei announced in late March 2015 that it entered a legally binding framework with Beijing Nesound International Media Corp to sell its 100% stake in Xunlei Kankan, its online video arm, for RMB 130M ($21M).  Xunlei’s Board approved it, and there is a high breakup fee of RMB 52M, so it’s likely to close in 2Q15.  Not only does this drastically improve profitability (Kankan is running annual losses of an estimated $20M), but it is a major step towards refocusing managements efforts towards mobility. 

Project Crystal: Xunlei launched an innovative plan in early 2014 that aimed to utilize the idle bandwidth and computing capacity of user devices to achieve better bandwidth economics and share those economics with users.  Essentially, participant users would be paid to contribute the upload bandwidth of their devices to the system when their devices were turned on.  The system would then identify all these devices and pre-load content to them for potential downloads.  When another user requested a download of specific content, the system would be able to locate and connect to an optimized set of data locations (other users’ devices) for downloading.  As of the last update (4Q14), 19% of the bandwidth that Xunlei used for its own acceleration services was crowd sourced (50% more than in the previous quarter).  So instead of buying this bandwidth from traditional carriers, Xunlei was able to supply 19% using Crystal technology.  Once Xunlei scales this up and adds third party sales (ie other internet content providers with bandwidth demand willing to pay Xunlei for the bandwidth crowd sourced through Crystal), Crystal could become a meaningful profit contributor. 

Mobility: Xunlei’s primary focus in 2015 is to transform from a primary PC-based company to a mobile internet-based company (hmmm, didn’t Lei Jun do exactly that with Kingsoft?).  The anchor of that mobility strategy is its relationship with Xiaomi.  Towards the end of 2014, Xunlei’s mobile acceleration software was officially adopted by Xiaomi’s latest operating system MiUI6.  Xunlei’s installed base was thus tied to Xiaomi’s installed base, which hit 55 million in 1Q15.  Xiaomi expects that number to reach 100 million by the end of 2015, quite an impressive growth trajectory.  One key metric that Xunlei is tracking is daily active users (DAU) which are running at about 20% of the installed user base.  While we’re in early innings here, we have the seeds of a turnaround in place.  The big question is what’s next in the Xunlei/Xiaomi partnership?  What will they do with the cash hoard?  Will they take advantage of the massive disparity in valuations of A shares vs US ADRs by going private and relisting in the A share market?  There are many possibilities here, and while the uncertainty level is high during this transition, and the many hypothetical portfolio/restructuring possibilities do require a leap of faith, we wanted to share this story with VIC and bring Lei Jun into the discussion Boards.  At a minimum, this is a pretty fascinating case study. 

Valuation:

The pending sale of Xunlei’s money-losing online video business Kankan (to close this summer for US $21M) should lift EPS from $0.48 to at least $0.75 in 2016.  The cash hoard should grow by about $80M to $515M by the end of 2016 with proceeds from the asset sale as well as internal cash generation (this is an asset light business model).  Discounting the cash by 25% to account for repatriation taxes (for a net $6 per share in cash) and applying a more reasonable multiple of 12X 2016 EPS, we get a fair value of $15 per share for about 40% upside.  Any strategic deals or successful internal restructuring efforts should lead to additional upside. 

The chart below is meant to shine a light on the value discrepancy between Xunlei and Lei Jun’s other major investments.  While there’s much work to do here, this project appears to be a great fit, especially in light of his success with Kingsoft. 

Xiaomi-Related Valuation Metrics

One other comp to keep in mind is US-listed company Box Inc (BOX), which is similar to the Dropbox model of online storage and content sharing, but the target customers are Fortune 500 companies looking for enterprise-wide storage and sharing solutions.  It has a $2.0 billion market cap ($1.7 billion enterprise value) and its revenue in the fiscal year ended 1/31/15 was $216M (not profitable at the EBITDA level), up 74% y/y. 

Risks:

• Improvement in China’s broadband speed: Clearly a speedier national broadband network could reduce the advantages the company can provide to its customers, jeopardizing growth prospects for the firm.  This appears more of a distant risk, but it’s real. 
• Internet purification: In April 2014, the Chinese government embarked on several initiatives designed to limit pornographic and non-authorized publications on the internet.  Since then, growth in subscriptions at Xunlei reversed amid rising complaints regarding less content and restricted functions.  This is an ongoing headwind, and Xunlei will need to adapt its content accordingly!
• M&A risk: Xunlei is sitting on quite a cash hoard.  While Lei Jun’s involvement should be helpful in crafting an M&A strategy, there’s always risk that management will do a value-destructive deal. 

Catalyst

• M&A: There are multiple possibilities here, but Lei Jun has got to be looking at Xunlei as a possible acquisition vehicle, with nearly two thirds of the market cap in unrestricted cash.  Assets could be folded in from Xiaomi or Kingsoft pretty easily.  Conversely, Kingsoft Cloud could use Xunlei as a shell to do a reverse takeover and get a U.S. listing. 
• A share listing: Xunlei’s competitor Beijing Baofeng Technology (offers online audio and video entertainment) prepared for an IPO at the same time as Xunlei.  But instead of listing in the U.S., it pulled the IPO and more recently (3/26/15) listed in the A share market (300431 CH).  So while Xunlei’s stock is down 11% from its IPO, its market cap languishing under $700 million, Beijing Baofeng’s stock is up 41X and now sports a nearly $6 billion market cap.  Management of Xunlei must be thinking hard about taking the company private and re-listing in the A share market. 
• Further portfolio/restructuring actions: Within the Xiaomi ecosystem, there are multiple synergistic relationships that have not yet been tapped.  For instance, Kingsoft and YY have their own online, web-based games with their own captive user bases – Xunlei could benefit from cross-selling opportunities.  We may also see more pruning – the online gaming division of Xunlei is off to a poor start and may be viewed as non-core. 
• Earnings 5/20/15 after the close: I’d expect another messy quarter with Internet purification continuing to pressure subs and ad rates, but we may see more portfolio moves.