CIO’s Quarterly Commentary – Felix Narhi – Q1 2017 – Part 2
All we want in life is no competition or an unfair advantage.
At Pender we sometimes joke that all we want in life is either no competition or an unfair advantage. This is only partly in jest. In truth, virtually all great businesses have unfair advantages – Warren Buffett calls such advantages economic “moats”. Moats are very valuable if they provide a business with the ability to generate sustainably high returns on capital, which creates wealth for its owners. The ideal business is one that uses very little capital, has a defensible business model and can continue to grow at a decent clip. The growth of such a capital-light business generates lots of extra free cash flow every year which is not required in the core business. As an investor, you get a double-barreled compounding effect from this earnings growth – first from the value of the growing business and second from the value of the excess cash which can be used to fund other promising opportunities, repurchase shares or pay out dividends. There are relatively few businesses like that, but this approach describes the business models of many of today’s tech giants including Apple, Google, Facebook, Amazon and Microsoft. As these enterprises illustrate, management teams that have the ability to look around the bend and make smart early bets on tomorrow’s promising trends and achieve critical mass have the potential to create large fortunes. Collectively, these companies have grown so rapidly that extra cash continues to pile up faster than many of these companies can deploy it. A high class problem!
We believe that Artificial Intelligence (AI) represents the next great technological leap for society and a once-in-a-generation opportunity to develop new commercial frontiers while transforming a wide range of industries such as automotive, health care, education, financial services and industrial automation. During the first quarter we bought stakes in Baidu (NASD: BIDU) in part to tap into the firm’s AI optionality. We believe it makes sense to value Baidu on two key pillars. First on a “what do we have now?” sum-of-the-parts basis, to assess if the valuation is sensible relative to the opportunity set today and second, on a “what are they going to do with it?” basis to assess future optionality.
What do we have now?
Led by its co-founder Robin Li, Baidu is often referred to as “China’s Google” because it is the dominant search engine in China. Similar to Google’s core business, Baidu’s search functionality continues to serve a very fundamental need for society as more and more information is digitized. The business model is very scalable and has historically generated consistently high margins. Baidu’s core search/advertising business fits in the ideal business profile noted above because it requires virtually no invested capital to continue to operate and grow.
Taking a page out of Google’s playbook, Baidu has used its core search engine’s cash cow to fund other start up business initiatives, albeit with mixed results. Successes include its fast growing on-line video service in China, iQiyi (comparable to Netflix) and investments that led to a 20% stake in China’s leading online travel agency, Ctrip (think Priceline and Expedia). But some missteps have overshadowed these successes. Last year the company faced headwinds following the death of a Chinese student who found a fraudulent cancer treatment searching on Baidu. The subsequent reputational damage and government crackdown on the advertising of its medical advertising partners caused Search Services revenues to decline for the first time since its 2005 IPO. In addition, some key investments have been disappointing, while losses in other major units continue to drag down company margins. Even iQiyi, which is considered a success, is not yet profitable. Not surprisingly, the stock has been pressured by these developments. However, we believe there are some early signs that a number of key business units could be nearing an inflection point, which could unlock hidden value in the shares.
Baidu is widely acknowledged as one of the pioneers of AI research and its early bets are already breathing new life into the search business with new products like “newsfeed” posting impressive adoption rates over a short period of time. Moreover, the drivers of the losses in the non-core divisions are largely discretionary as management seeks scale to reach critical mass in large new markets. Management now anticipates the losses in these non-core businesses to moderate as those units reach scale over the next few years (iQiyi), are sold (Baidu Mobile Games), or as underlying business models change (Nuomi). In short, it appears operating profits are set to rebound.
We believe a sum-of-the-parts (SOTP) valuation provides a good starting point for the fair value of Baidu today. Some assets like net cash and its 20% stake in publicly-traded Ctrip (NASD:CTRP) are straightforward to value, while others require conservative “guestimates” to derive sensible intrinsic value ranges. Excluding net cash and other non-core investments, we estimate BIDU is trading at less than 12x the normalized pre-tax profits of its core search business, a compelling discount to other well-known tech peers who operate other ideal businesses with deep moats. We believe a SOTP analysis provides foundational downside support for BIDU at current price levels, but Baidu’s value over the medium-to-longer terms will largely depend if it can parlay its early AI leadership in China into commercial success.
What are they going to do with it?
Baidu recently made an all-in bet on Artificial Intelligence which we believe gives optionality to the stock that is not fully appreciated by investors focused on today’s opportunity set. There are three key ingredients for AI success: 1) ownership of massive hard-to-replicate data sets; 2) high performance computing to mine insights from the data and; 3) ability to harness a deep bench of AI talent. Given these gating factors, it’s not surprising that the early AI leaders are all familiar tech names that have collected massive amounts of user data over the normal course of their business operations for many years. As a result, a relatively small handful of high-profile tech giants own the most promising data sets which can be applied to some of the most interesting business opportunities. It has been said that data in 21st Century is like oil in the 18th Century – an immense, untapped valuable asset. However, unlike oil, the amount of data keeps on growing exponentially and is not depleted after use. Those who can tap into data’s fundamental value, learn to extract and monetize it should continue to see huge rewards. These include the “Super 7” hyper scale tech giants in the US and China: Facebook, Amazon, Google, Microsoft, Alibaba, Tencent and Baidu. It is virtually impossible to replicate the data treasure trove that has emerged from Facebook’s ubiquitous social media presence, Google’s global search data/mapping domination and Amazon’s online retail empire/ cloud computing expertise. However, these global giants are largely absent in China and as a result, the most valuable datasets are now owned by domestic tech equivalents.
Like the other tech giants, Baidu enjoys huge barriers to entry which would be very difficult for others to replicate anytime soon (if not impossible without a time machine!). As the leading search engine in China, Baidu has amassed vast amounts of diverse and proprietary data in a market with over 700 million internet users who speak the same language, have the same culture and abide by the same laws. Baidu also owns one of the leading map services in China, which is increasingly important as autonomous driving takes off and online/offline worlds continue to blur. These ever-growing data sets are becoming increasingly valuable as traditional learning algorithms are supplanted by large artificial neural networks that are driven by advances in high performance computing. New business models are being created and many old models will be displaced as disruptors uncover new insights and possibilities from this data. Better to own the disruptor than the disrupted!
Beyond the moat created by the ownership of big proprietary data sets and advances in high performance computing, arguably the most unfair advantage in China is the support of the government. The government views AI as strategically important and as a once-in-many-decades chance for China to lead the charge in technology. Although Chinese peers Tencent and Alibaba are very strong players with huge accessible user databases of their own, earlier this year the government granted special permission to Baidu to lead the development of a national deep-learning AI research lab, further solidifying Baidu’s “moat” and status as the nation’s AI champion. Although details are still scarce and the benefits are hard to quantify, we believe this is a big deal because it will help nurture its homegrown leader and keep global competition at bay. Baidu can now access the enormous amounts of data held by Chinese companies and universities, be aided by the large number of Chinese engineers being trained on both sides of the Pacific and obtain government backing (including funding and assistance with special permits required to test self-driving cars, for example). Baidu will work with leading AI researchers in Tsinghua and Beihang universities, as well as other Chinese research institutions and focus on computer vision, biometric identification, intellectual property rights and human-computer interaction. Interestingly, the Chinese government is accelerating its AI efforts just as the Trump-led White House appears increasingly distracted by internal issues and looking backwards to revive yesteryear industries like coal mining. In addition, Trump’s restrictive immigration policy is giving Chinese firms an unprecedented opportunity to attract top AI talent. Time will tell whether the leading Chinese AI firm or a leading US coal firm turns out to be the better investment opportunity.
Worth mentioning, many of existing Baidu business units have already been enhanced by AI. For example, search by voice or by image, instead of just touch/typing, is fundamentally changing the way people interact with technology. For Baidu, it extends the utility of its search capability to every device and in every scenario including in homes and cars, and moves society away from today’s ubiquitous head-down interactions with smartphones. Baidu has launched new AI-driven business units to maximise its world-class capabilities in autonomous driving (think Tesla’s Autopilot) and with its DuerOS Conversational Computing platform to harness voice-activated technologies (think Amazon Alexa/Echo). Both have the potential to be disruptive to their respective industries as consumers embrace new technologies to improve their lives, especially in China where Baidu enjoys essentially “no-competition or an unfair advantage” relative to their global peers. This is just the beginning, as we anticipate AI will open up many value creation opportunities which are hard-to-imagine today, but will no doubt be commonplace in the future.
Baidu – an emerging Chinese AI giant nearing an inflection point
We believe the stock was pricey five years ago, but the fundamentals have since backfilled in nicely. Baidu’s revenue is up four-fold over the last five years, but the stock has been range bound because operating profits have flat lined due to the drag from start-up losses in promising non-core business segments. As noted, the company could be near an inflection point as losses from non-core businesses start to fade and the core search business picks up momentum again following last year’s setback. Moreover, we believe the promise of AI is very real, with little to no optionality from potentially game changing opportunities yet baked into the stock price (autonomous driving and conversational computing being only the start). As Amazon founder Jeff Bezos might say, it is just “Day 1” in the AI sector which has finally reached a tipping point and stands at the cusp of its golden-age. The early promise of AI has not impacted the shares of all AI leaders (yet). Of note, Baidu’s hyper scale “Super 7” peers have market capitalizations that are 5-10 times larger than BIDU’s current market valuation. In our view, BIDU appears somewhat akin to a compressed spring, poised for significant upside as it starts to monetize its AI lead in China. Hence the opportunity. We are long BIDU.