2023-12-15 15:14:06 ET
Summary
- Baidu is the tip of the spear in the coming LLM/AI wave in China, offering a large upside for investors with aggressive risk profiles.
- The company's AI efforts are funded by a cash cow search engine business and a fortress balance sheet.
- China's geopolitical risk is a legitimate concern, but the risk has softened over the past few years.
- Baidu stands to gain from greater operating efficiency: management has signaled intent to become leaner and more focused with capital allocation.
- Price target of $170 for Baidu even after applying a 50% China discount, the stock rates a Strong Buy for risk-hungry investors willing to take a flier on China.
Following Baidu's ( BIDU ) 3Q earn ings announced on November 21st, the market is beginning to wake up to the stock as China's premier AI play, with an upgrade from Nomura, Barron's publishing an article entitled "Forget Alibaba. Baidu Is The Best Way to Play Chinese Tech Stocks," and The Motley Fool comparing Baidu to Nvidia ( NVDA ). Baidu ran up +20% off its lows on the back of a strong earnings report and rising investor sentiment, before the usual China pessimism kicked in, pulling the stock back -10%.
Despite (legitimate) geopolitical concerns, Baidu offers large upside for investors with aggressive risk profiles; it is the tip of the spear in the coming LLM/AI wave for the second-largest economy in the world, funded by a search engine cash cow and $20B (half its market cap) in actual cash. Baidu is no Nvidia yet, but one day it might be.
The elephant in the room when it comes to any Chinese stock is China, as foreign capital continues to see a net outflow from the country after Beijing's 2021 crackdown on Big Tech. The risks are real, but I think significantly reduced in the past couple of years, especially for Baidu.
Since 2021, China has offered positive regulatory signal regarding the VIE structure , given the PCAOB full access to audit Chinese firms, and President Biden's executive order to block outbound tech investment to China in sensitive sectors notably excluded publicly traded securities. President Xi Jinping's visit to San Francisco last month is another signal that China is willing to play ball to try to regain investor confidence. President Biden's EO may be as much a tailwind for Baidu as a headwind, since the restrictions targeting greenfield investments/joint ventures/sales of frontier technologies will stifle Chinese startups far more than China's largest hi-tech firms, locking in Baidu's lead over the competition.
Speaking of leads, Baidu is definitely the current frontrunner in China's AI race, owing to years of heavy investment in the technology, long before ChatGPT made AI the new global hotness in November 2022. It was the first publicly traded company to release an LLM, beating Google's ( GOOGL ) Bard to the punch by 5 days with ERNIE Bot in March 16th, 2023. Alibaba ( BABA ) released Tongyi Qianwen nearly a month later; Tencent ( TCEHY ) didn't unveil Hunyuan until September.
This was not a happy accident, Baidu was able to bring its LLM offering to market so much faster than rivals because it's been heavily investing in AI for years. As of year-end 2022, Baidu holds more machine learning + artificial intelligence patents than Tencent, a company 10x the size, and more than Alibaba and Huawei combined !
Largest patent owners in AI/ML (statista.com)
Baidu also recruited Andrew Ng, cofounder of Google Brain and one of the pioneers of GPU-based deep learning, to head its Artificial Intelligence Group as Chief Scientist all the way back in 2014 (Ng left Baidu three years later).
PaddlePaddle, Baidu's open source deep learning framework, is the leading ML framework in China, the equivalent of Google's TensorFlow or Meta's ( META ) PyTorch. A CB Insights report showed Baidu has been discussing AI strategy on earnings calls long before Alibaba, Tencent, and even Google:
Risks
Other than the standard China risk, the main risk to the Baidu thesis is the still-unproven monetization pathway for LLMs. Given the crowded competitive field, including with open source models like Meta's LLaMa, it's highly possible that models will become commoditized, which means profits will likely accrue to the bottom of the stack (the cloud providers) and the top of the stack (the app developers). Baidu has a strong AI Cloud offering which achieved operating profitability Q1 of this year , and management indicated that GAAP profitability will be sustained moving forward in the Q3 conference call . Its biggest competitor in this area is Alibaba, which has a staggering 36% market share:
As a search engine, Baidu also faces the same risk as Google that its CPC ads business may be cannibalized by large language models. This is one area where ironically it may be beneficial, in a sense that Baidu's search business is far less profitable than Google's. China's "walled garden" style Internet, with large amounts of content being locked behind proprietary platforms (such as WeChat Official Accounts being searchable only through WeChat), means traditional search is both less useful and less lucrative...however, this means the potential upside from being a leading AI disruptor may actually be worth its core business being one of those disrupted for a company like Baidu, to an extent that isn't the case for Google.
Baidu also faces execution risk, as its cost management over the years has not been the greatest. As seen in the chart below, although the company has increased revenues significantly over the past few years, this has not translated into operating profit growth:
Baidu operating results (quickfs.net)
It's not all bad, as Baidu owes a large part of its lackluster single digit returns-on-equity to its heavy investments in R&D: over the past 5 years, it spent an average of 17.8% of annual revenues on R&D, versus only 7% for Alibaba. Baidu's combined R&D + SGA expenses were 35.8% of revenues for the trailing 5 years, versus 27.5% for Alphabet/Google. Management recognizes the need to trim the fat, noting a plan to prioritize efficiency moving forward:
"In the future, we will realign resources to invest in this growth opportunity and shift away from lower priority efforts and improve efficiency for existing businesses, thus balancing investment and margins ." - Baidu CEO Robin Li, Q3 2023 Earnings Call
Outlook
Although Baidu's revenues took a dip in 2022 from China's coronavirus restrictions (which dragged the whole sector), the abrupt end of the country's zero-COVID policy in December of last year has resulted in an earnings rebound in 2023. In Q1, Q2, and Q3, Baidu posted +10%, +15%, and +6% revenue growth YOY, and +91%, +53%, and +18% operating income growth YOY, respectively. We are looking at $8.74 diluted EPS TTM, for a trailing P/E of 13.15. Consensus EPS estimates for FY2024 land on $10.48, for a forward P/E of 10.75. This should be achievable moving ahead as we haven't lapped 2022's 4Q yet, so next quarter should still have easy comps.
Baidu's cash generation has been healthy as well, with $3.4B free cash flow TTM. It continues to have a fortress balance sheet, with $27.7B in cash equivalents, against $8B in long term debt, for a nearly $20B net cash position. The company also has $3.4B in long term time deposits (most with maturities within 3 years), and $7.2B in long term investments. I would credit the time deposits/held-to-maturity securities to cash, while discounting the long term investments to err on the side of caution, as fair market values tend to fluctuate.
There's no question Baidu has been putting up strong financial numbers this year as China emerges from lockdown, which it should be able to extend into next year. We are looking at a company that's a leader in its industry, with respectable historical profitability, and a massive war chest to fund its AI ambitions, trading at low teens price-to-earnings ratio, relative to the S&P's 26 P/E. Moving forward, I expect Baidu to expand earnings not only through revenue growth (especially in its cloud business), but also through achieving greater operating leverage by improving expense management, prioritizing return on invested capital.
It definitely has levers to pull in this domain. For example, Baidu attempted to acquire JOYY's live streaming service YY Live in 2020 for over $3B, a deal that's currently still stuck in regulatory limbo. In my opinion, this deal was an overpay for a product that adds nothing to Baidu's core businesses. There are many areas like this where Baidu can become leaner and meaner. Shareholders may also benefit from a capital return either from buybacks (great way to unlock value given Baidu's currently cheap shares) or a dividend (as Alibaba has recently done).
Takeaway
At the $115 price level, Baidu trades at a forward earnings multiple of 10x. If it were a US stock, this multiple would be well in excess of 40x for a company with such promising embedded AI potential. The Magnificent 7 currently trades for an average forward P/E of 32.6 at much larger market caps, with many of its constituent companies already having enacted severe cost-cutting measures and headcount reduction to improve operating efficiency, a lever Baidu still has available to pull. Palantir ( PLTR ), a company of similar size to Baidu, also with a mature legacy business that's proven to be a natural fit for expanding into AI, trades at a forward P/E of 66. Unlike Palantir, Baidu doesn't have to compete against the Mag 7 either.
If Baidu were a US company: normalizing earnings at $2.5B (achievable if it sustains the mid-range of its historical RoE), applying a 40x multiplier, and adding $20B net cash on top, we get a valuation of $120B or $340/share. Slash this in half for the geopolitical risk factors (call it the China discount), and we still have a company worth $60B or $170/share, +50% above its current price.
That's not even taking into account Baidu's other businesses like Apollo Go, the leading autonomous driving platform in China, or iQIYI, Baidu's Netflix-style streaming service. Before the regulatory crackdown on China's tech sector, Baidu's stock surpassed $300/share on its autonomous vehicle/EV ventures: while the market has become much better acquainted with the risks of investing in China since then, at the same time, Baidu's AI potential in LLMs is far larger than in self-driving cars. As such, Baidu's stock currently rates a Strong Buy in the sub-$125 price range for risk-hungry investors willing to take a flier on China.
For further details see:
Baidu Would Be $340 Per Share If It Were A US Company