Latest / Deep Values: Value Investing Deep Dives / Duan Yongping‘s $25 Billion Pivot to Software
Transcript
- 0:00Imagine managing $25 billion. That is a lot of money to keep
- 0:05track of, right? But instead of staring at
- 0:08blinking screens or, you know, shouting on a trading floor and
- 0:11firing off high frequency algorithmic trades, your entire
- 0:14strategy is based on Wu Wei. The ancient Eastern philosophy
- 0:19of effortless action. Exactly or essentially doing
- 0:22absolutely nothing. Today we are looking at the
- 0:25exact portfolio of a man who does just that.
- 0:28Yeah. And it is a profound departure
- 0:30from Wall Street culture. I mean, Wall Street usually
- 0:32demands constant frenetic activity just to justify
- 0:35management fees, you know? Well, completely.
- 0:37But today we're looking at the Q1202613-F SEC filings and the
- 0:43strategic analysis of H&H International Investment LLC,
- 0:47right. And the master allocator behind
- 0:48this is Duong Yongping, the the reclusive billionaire founder of
- 0:52BBK Electronics. He has somehow managed to take
- 0:55the foundational value investing principles of Warren Buffett and
- 0:58Charlie Munger and just fuse them with this meditative
- 1:01Eastern discipline. OK, let's unpack this, because
- 1:04our mission for this deep dive is to decode this seemingly
- 1:07contradictory strategy. Yeah, it really does seem
- 1:09contradictory at first glance. It does.
- 1:12We want to figure out why he just massively reshuffled his
- 1:15technology holdings in this latest quarter and learn how to
- 1:18apply his extreme, almost ruthless discipline to our own
- 1:22decision making. I love that because to
- 1:24understand how someone builds a $25 billion empire by actively
- 1:29trying not to act, we really have to start with his
- 1:32foundational definition of an asset, right?
- 1:35Because he doesn't look at it like most traders do.
- 1:37Not at all. For Duan buying a stock is never
- 1:40about trading a piece of paper or, you know, guessing where a
- 1:43line on a chart will go next week.
- 1:46He views it strictly as purchasing an undivided
- 1:49fractional share of a living business enterprise.
- 1:52He just discards short term macroeconomic forecasting,
- 1:55interest rate modeling and technical analysis as purely
- 1:59speculative noise. So he just ignores all of it.
- 2:02Completely right. His entire framework rests on a
- 2:04single mathematical concept, the discounted cash flow formula.
- 2:08Let's break that down real quick for anyone who hasn't spent time
- 2:10modeling spreadsheets. The discounted cash flow formula
- 2:14basically states that an asset is only worth the total amount
- 2:17of cash it is going to generate from today until the end of its
- 2:20life, discounted back to today's value, exactly to account for
- 2:24the fact that a dollar today is worth more than $1.10 years from
- 2:27now. Right, and the math itself is
- 2:29completely trivial. I mean, a middle schooler can
- 2:31calculate a discount rate. Yeah, the formula isn't the
- 2:34secret sauce. No, the profound difficulty lies
- 2:37in the inputs. Like how do you predict what a
- 2:39company's cash flow will be in the year 2036?
- 2:43Which is almost impossible for most companies.
- 2:46Exactly to do that accurately, a business must possess a
- 2:49competitive advantage, an economic Moat that is so durable
- 2:53and unassailable that its future revenues are practically
- 2:57guaranteed. Wow, so he's looking for near
- 2:59absolute certainty. He is, And because finding those
- 3:03types of businesses is incredibly rare, Duan forces
- 3:06himself into patients. And what's fascinating here is
- 3:09how he does it uses what he calls a stop doing list.
- 3:12Which is such a brilliant mental model.
- 3:14I mean, it isn't a checklist of goals to achieve, right?
- 3:17It's a list of absolute unbreakable negatives designed
- 3:21to prevent catastrophic, unrecoverable errors.
- 3:24Yeah, and his list is anchored by three structural rules #1 do
- 3:29not short equities #2 do not employ leverage or borrow on
- 3:33margin, OK and #3 do not invest a single dollar in businesses
- 3:38outside your own circle of competence.
- 3:41The sources highlight this really visceral story about that
- 3:44first rule, too. Duan actually lost $200 million
- 3:49shorting the Chinese search engine Baidu.
- 3:51Yeah, 200 million. That's a painful lesson.
- 3:54Right, because when you buy a stock, your maximum loss is
- 3:57capped at 100%, but when you short a stock betting it will go
- 4:00down, your potential loss is theoretically infinite if the
- 4:04stock keeps climbing. It's a mathematically
- 4:06asymmetrical risk profile that is completely incompatible with
- 4:10wealth preservation. It's like bowling with the
- 4:12bumpers up. You're structurally eliminating
- 4:14the gutter balls so you can focus entirely on the pins.
- 4:17That is a great way to put it, and that ties directly into his
- 4:19ban on leverage too. I mean, even the most rational,
- 4:22brilliant investor in the world can be completely wiped out if
- 4:25they use borrowed money. Because markets can stay
- 4:28irrational longer, then you can stay solvent.
- 4:30Precisely. If the market takes a temporary
- 4:33illogical nosedive and you are heavily leveraged, you will get
- 4:37a margin call. Your broker will just force you
- 4:39to sell your assets at the absolute bottom, locking in
- 4:42permanent destruction of capital.
- 4:43But wait, I need to push back on this a little bit.
- 4:46You're telling me this guy is obsessed with capital
- 4:48preservation and avoiding catastrophic losses?
- 4:52But the sources say he operates with a fully invested mandate.
- 4:56He does. So in a massive like 2008 style
- 4:59liquidity crisis, holding 0 cash sounds incredibly reckless.
- 5:04How does he navigate market downturns without pulling out
- 5:07into cash? Well, it sounds reckless until
- 5:09you look at how he redefines the concept of cash.
- 5:12Duan views holding actual physical Fiat currency as a
- 5:16guaranteed mathematically certain loss.
- 5:19Because of inflation, right? Inflation acts as a silent tax,
- 5:23constantly eroding the purchasing power of the dollar,
- 5:26So to him, parking billions in cash is a losing position.
- 5:29Interesting. Instead, he engineers a unique
- 5:32proxy for cash. His ultimate safety net is a
- 5:35massive position in Berkshire Hathaway.
- 5:38So he essentially outsources his defensive maneuvering.
- 5:41Exactly. In the Q 12026 filing, Berkshire
- 5:44makes up almost 22% of his portfolio.
- 5:47Wow. Yeah, that is roughly $4.38
- 5:50billion. And with Greg Abel now managing
- 5:53Berkshire defensively and sitting on their own gargantuan
- 5:55cash reserves of nearly $400 billion, which one gets the best
- 5:59of both worlds? He gets institutional grade
- 6:02liquidity and this absolute rock solid flooring of intrinsic
- 6:05value. Exactly, but he avoids the
- 6:07direct inflationary drag of sitting on pure cash himself.
- 6:10So Berkshire is the bedrock, but the undisputed king of his
- 6:14portfolio is Apple, sitting at an astounding weight of nearly
- 6:1737%. That's over $7 billion.
- 6:20A massive concentration. It is, but here is the thing
- 6:24that jumped out at me in the filing.
- 6:26His Apple share count actually dropped by over 10% this
- 6:29quarter. For a guy whose whole philosophy
- 6:32is buying monopolies and holding them forever, slashing a 10th of
- 6:36his biggest position looks like a sudden loss of faith in the
- 6:39iPhone. I know it looks like a sell off
- 6:41on paper, but his fundamental thesis on Apple hasn't changed
- 6:45at all. Apple possesses an unparalleled
- 6:49economic Moat built on high customer switching costs.
- 6:53Oh for sure, once you have the iPhone, the iPad, the Mac, and a
- 6:56decade of your family photos locked inside iCloud, the
- 7:00friction required to leave that ecosystem for an Android is just
- 7:03massive. Exactly.
- 7:04The share reduction wasn't a panic sell, it was simply the
- 7:07automatic, unemotional execution of his options strategy.
- 7:10OK, let me make sure I'm wrapping my head around this
- 7:12mechanically, because covered calls can confuse a lot of
- 7:16people. Yeah.
- 7:16They can get technical. So it's like keeping a non
- 7:18refundable deposit from a buyer who wants the option to buy your
- 7:21house in a year, but only if the neighborhood values absolutely
- 7:25skyrocket. If the values don't surge, you
- 7:29keep the house in the deposit. If they do surge, you are
- 7:32legally obligated to sell the house, but at a high price you
- 7:35were already thrilled to accept anyway.
- 7:37That is the exact mechanism he systematically writes covered
- 7:42calls on his existing core positions.
- 7:45So he sells another investor the right to buy his Apple shares at
- 7:49a higher strike price, and in exchange he collects an upfront
- 7:53premium, which is basically a massive passive yield.
- 7:57If Apple stock price surges past that strike price, those shares
- 8:00are automatically assigned, meaning they are sold away.
- 8:03So he's using options not to gamble, but to engineer a
- 8:06mathematically guaranteed discount.
- 8:08Precisely. He doesn't have to make an
- 8:10emotional decision to take profits.
- 8:12The maths simply executes itself.
- 8:14So he gets paid millions while he waits, and if it sells, it
- 8:17sells at his predetermined valuation.
- 8:19Exactly. But if his strategy with Apple
- 8:21and Berkshire is essentially just automated waiting, where is
- 8:24the action? Because looking at the rest of
- 8:26this 13 F filing, we're seeing a massive aggressive tech
- 8:30migration. Yeah.
- 8:31And here's where it gets really interesting.
- 8:33Yeah, because the most profound strategic shift in this
- 8:37quarter's filing is the migration of capital away from
- 8:40physical hardware manufacturing towards software lock in.
- 8:43The scale of this move is staggering.
- 8:45I mean he completely liquidated his entire position in ASML.
- 8:50Right, 9900 shares just gone. And he slashed his Taiwan
- 8:54Semiconductor, or TSMC holdings by nearly 90%.
- 8:58Yeah. Almost an 88% slash.
- 9:00These are the companies that possess virtual monopolies on
- 9:03the physical fabrication of the world's microchips.
- 9:05But simultaneously, he expanded his NVIDIA holding by over 90%,
- 9:10promoting it to a 12% portfolio weight.
- 9:13So why dump the physical monopolies for NVIDIA?
- 9:15It comes down to understanding the vulnerabilities of physical
- 9:18assets in the modern world. ASML and TSMC require colossal
- 9:23capital expenditures. They have to build actual
- 9:25factories. Exactly.
- 9:27Every incremental dollar of revenue requires them to build
- 9:30multibillion dollar physical factories.
- 9:32They are bound by the physics of manufacturing and the
- 9:35geopolitics of where those factories are located.
- 9:37Right, They are vulnerable to real world bottlenecks.
- 9:40A delayed shipment of concrete or geopolitical skirmish can
- 9:44halt their entire revenue stream.
- 9:47Exactly. Now contrast that with NVIDIA.
- 9:50The broader market often miscategorizes NVIDIA is just
- 9:53another hardware company making graphics processing units.
- 9:57But Duan sees it differently. Yes, Duan recognizes that
- 10:00Nvidia's true unassailable is actually its software ecosystem,
- 10:05ecifically CUDA C Johnny A is the foundational software layer
- 10:09that develoers have been using to write AI alications for
- 10:12Nvidia's processors for almost 2 decades.
- 10:14So even if a competitor somehow designs a chip that is slightly
- 10:17faster or cheaper than Nvidia's, it simply doesn't matter.
- 10:20Not at all. Because migrating away from CDU
- 10:22to require a company's entire software development team to
- 10:25spend years rewriting their entire enterprise code stack
- 10:28from scratch. Exactly.
- 10:30The switching cost is so astronomically high that
- 10:32customers are structurally locked in.
- 10:34But moving capital from TSMC to NVIDIA.
- 10:37He is ruthlessly prioritizing software enabled pricing power
- 10:40over physically bottlenecked industrial manufacturing, and
- 10:44the execution of this massive NVIDIA position is a master
- 10:48class in applying deep value mechanics to high growth tech
- 10:52stocks. Oh yeah, the sources outline the
- 10:54exact math of this trade and it is brilliant.
- 10:57He didn't just buy NVIDIA stock at the market price of around
- 11:00$116. No, he did not.
- 11:02He bought the stock and simultaneously wrote covered
- 11:06call options expiring in March 2026 with a strike price of
- 11:10$120.00. Right.
- 11:12And because Nvidia's volatility is so high, other investors paid
- 11:15him a massive premium of over $24.00 per share just for those
- 11:19call options. Which means his effective net
- 11:21acquisition cost was dragged all the way down into the low 90s,
- 11:24like $92.52. He's basically getting paid a
- 11:27massive rent check just for agreeing to sell his house at a
- 11:30higher price later. Exactly.
- 11:32He engineered his own margin of safety into an extremely
- 11:35volatile technology stock. It completely insulates his
- 11:38portfolio against a downside shock.
- 11:40But OK, if his entire thesis relies on extreme
- 11:44predictability, structural software lock in, and avoiding
- 11:48unpredictable variables, how do we explain the $1.2 billion
- 11:52elephant in the room? Let's look at Tesla.
- 11:55Yeah, this raises an important question.
- 11:57How does a strict value investor end up with a massive new
- 12:01initial core position in Tesla? We're talking 6.34% of his
- 12:05portfolio. Right, especially when he
- 12:07demands stable corporate cultures.
- 12:09And historically, Duan has been openly critical of Tesla.
- 12:12He publicly called it a dream valuation.
- 12:15Yes he did. And his rigid framework demands
- 12:17management teams with shared values and low volatility, which
- 12:21led him to actively avoid Elon Musk as a corporate leader for
- 12:24years. Yet out of nowhere, he is
- 12:26suddenly holding over 3.4 million shares.
- 12:29Did he just abandoned his principles?
- 12:31Well, the analysis points to two distinct theories that reconcile
- 12:34this paradox. OK, let's hear them.
- 12:35The first theory goes back to his option strategy,
- 12:38specifically cash secured puts. During periods of high market
- 12:42volatility, Duan frequently sells put options mechanically.
- 12:47This means he collects an upfront cash premium in exchange
- 12:50for promising to buy a stock if it crashes down to a specific
- 12:54lower price. Oh, I see.
- 12:56So if Tesla was trading at $200, Duan might have said I'll gladly
- 13:00buy this company at 140. Pay me $10 a share right now for
- 13:04that promise. Exactly.
- 13:05If the stock stays high, he keeps the millions in premiums
- 13:08and just walks away. But if Tesla took a sharp,
- 13:12sudden correction and crashed below 140, he was legally
- 13:15obligated to buy those shares. So he may have been
- 13:18automatically assigned those shares.
- 13:19He was forced to buy them, but at a massive predetermined
- 13:23discount that he felt offered a genuine mathematical margin of
- 13:26safety. Right, it removes the emotional
- 13:28hurdle of trying to buy the dip while a stock is in freefall.
- 13:31That explains the mechanics, but what about the fundamental
- 13:33business? I mean, Tesla is still a capital
- 13:36intensive hardware manufacturer. That leads to the second theory
- 13:39which suggests a bound fundamental reassessment of
- 13:42Tesla's Moat. Remember we discussed the
- 13:45software lock in of Nvidia's CDA with Tesla.
- 13:48Duan may have realized the mood is no longer the physical cars,
- 13:52it is the Full Self driving data network.
- 13:55Tesla has accumulated billions of miles of real world driving
- 13:59data. That is an expanding proprietary
- 14:01data advantage that creates A structural barrier to entry for
- 14:05any competitor trying to solve autonomous transportation.
- 14:08Yeah, he might be viewing Tesla not as a car company, but as an
- 14:12irreplaceable network asset. Which is such a vital lesson for
- 14:15you listening right now. Think about how often we allow
- 14:18our personal biases, like, you know, finding a CEO's public
- 14:21persona distasteful or chaotic all the time, to completely
- 14:24blind us to the changing mathematical reality of a
- 14:27company's data mode. Duan is willing to let changing
- 14:29facts, or at least his options, math overtime his personal
- 14:33distaste to the tune of $1.2 billion.
- 14:36It requires immense intellectual honesty to change your mind at
- 14:39that scale, and that flexibility is also reflected in the broader
- 14:43architecture of his portfolio, which is constructed as an
- 14:46extreme barbell strategy. Yeah, let's examine the two
- 14:50opposite ends of this barbell. On the heavy end you have the
- 14:52massive, highly concentrated mega caps we've just broken
- 14:55down, but sitting right next to them is Pin Duo Duo at a 10%
- 15:01weight. Yeah, Pin Duo Duo is the perfect
- 15:03illustration of his third rule. Sticking strictly to your circle
- 15:06of confidence, Duan has a deeply rooted personal connection to
- 15:10his founder, Colin Wong. Really.
- 15:12Hong actually accompanied Duan to his famous charity lunch with
- 15:15Warren Buffett way back in 2006. So Duan isn't guessing about
- 15:19Pinduo Duo's corporate culture. He understands its DNA
- 15:21intimately. He comprehends their new Pin MU
- 15:24strategy, which focuses on building an own brand ecosystem
- 15:27similar to Costco's Kirkland line, better than any Wall
- 15:30Street analyst possibly could. So that handles the multibillion
- 15:33dollar heavyweights, but the opposite end of the barbell is
- 15:36what I find most intriguing. He holds these microscopic
- 15:40allocations ranging from absolutely 0% up to barely 1/10
- 15:44of a percent of his total portfolio, right.
- 15:47We are talking about tiny slivers of enterprise software
- 15:51and AI infrastructure companies like Palantir, Snowflake,
- 15:55Crowdstrike and specifically Credo Technology.
- 15:58Yeah, and Credo Technology, which is at a .35% weight, is a
- 16:01prime example of why he does this.
- 16:04They currently hold an estimated 88% market share in active
- 16:07electrical cables. What are those?
- 16:09These cables are highly specialized physical components
- 16:13required for high speed data transmission inside massive AI
- 16:16server clusters. Physical latency, the actual
- 16:19time it takes data to travel across a wire, is a major
- 16:22bottleneck in AI computing right now.
- 16:24Wow. OK.
- 16:25And Credo's near monopoly makes them practically irreplaceable
- 16:28for the giant tech hyperscalers building these data centers.
- 16:31Taiwan scaled his position in Credo by over 400% this quarter,
- 16:34yet it still only makes up a fraction of a percent of his
- 16:36portfolio. But why bother?
- 16:38I mean when you are managing $25 billion, why go through the
- 16:43administrative hassle of holding in .04% position in Palantir?
- 16:47It's not going to move the needle on his overall returns
- 16:50even if the stock triples. Well, your strategic analysis
- 16:53refers to these as Sentinel positions.
- 16:56They are not speculative gambles hoping for a quick return.
- 17:00They function as real capital watch lists.
- 17:03Oh, that makes. Sense By allocating a tiny
- 17:05fraction of percent into these emerging software verticals, he
- 17:08structurally forces his fund to continuously track and analyze
- 17:12them. So what does this all mean?
- 17:13It means skin in the game, even if it is just a tiny scratch,
- 17:17alters your psychology exactly. When you have actual capital on
- 17:20the line, you don't just skim a headline, you read the 10K
- 17:23annual reports differently. You track the real world
- 17:26monetization of Palantir's AI platforms with the scrutiny of
- 17:30an owner, not just an observer. Right.
- 17:32He is deploying his radar today, ensuring that when these
- 17:35emerging business models finally mature, improve their modes, he
- 17:38is already prepared to deploy billions of dollars with
- 17:41absolute conviction. It is the purest application of
- 17:44Huawei. He is putting the analytical
- 17:46pieces in place effortlessly right now so that the heavy
- 17:49lifting of making a multi billion dollar capital
- 17:52allocation decision tomorrow is already done.
- 17:55Beautifully said. Before we get to our final
- 17:57takeaways, I want to briefly mention something for you
- 17:59listening to access more advanced researches,
- 18:02applications and our latest published personal intelligent
- 18:05multi agent Deep value intelligence.
- 18:08Welcome to explore and join Deep Values dot AI community.
- 18:13And to track total market temperature.
- 18:15Welcome to Search and download iOS App Buffett Indicator from
- 18:19Apple App Store. Disclaimer All analysis is based
- 18:22on publicly available sources and may be subject to revisions
- 18:25or differing interpretations. Right.
- 18:27All contents from this channel is for information and research
- 18:30purposes only and does not constitute investment advice.
- 18:33Conduct your own due diligence before making investment
- 18:36decisions. If you like to access wider
- 18:38topics including life and investing philosophy, please
- 18:40subscribe to our sister channel, Deep Values.
- 18:42Don't forget to like this video. Much appreciate it.
- 18:45Absolutely. All right.
- 18:46As we pull all these threads together, the brilliance of Duan
- 18:50Yongping's Q 12026 strategy becomes incredibly clear.
- 18:55He has successfully merged Charlie Munger's rigid,
- 18:58uncompromising discipline with the effortless execution of Wu
- 19:02Wei. Yeah, it's a master class.
- 19:03He engineers his own margins of safety through highly
- 19:06sophisticated options writing, completely removing emotion from
- 19:09his entry and exit points. And most importantly, he is
- 19:13ruthlessly migrating his capital away from the physical
- 19:16geopolitical bottlenecks of the old economy toward the
- 19:20unassailable software and data modes of the future.
- 19:22Exactly, and I'd say the most actionable take away for you
- 19:25listening right now isn't to log into your brokerage and blindly
- 19:29copy these exact stock. Picks right?
- 19:30Definitely don't do that it. Is to architect your own stop
- 19:33doing list, whether you apply it to your career, your personal
- 19:36finances, or even just your daily consumption of
- 19:38information. What are the three absolute non
- 19:41negotiable negatives that if followed religiously, will
- 19:44prevent you from making unrecoverable errors in your own
- 19:46life? Eliminate the fatal mistakes 1st
- 19:49and the compounding growth will take care of itself.
- 19:52Set up your own bumpers and the gutter balls disappear.
- 19:54Exactly. But I want to leave you with one
- 19:56final, slightly provocative thought to Mull over regarding
- 19:59these unassailable digital modes.
- 20:02We spent a lot of time exloring why Duan shifted billions into
- 20:06software ecosystems like Nvidia's CU Day, specifically
- 20:11because rewriting that code is too painful for developers.
- 20:14High switching costs create the Moat.
- 20:17But if we project out just a little further into the future,
- 20:20if generative AI eventually evolves to the point where an AI
- 20:23agent can, Florida State translate and rewrite massive,
- 20:26complex enterprise code bases from one language to another
- 20:30instantly, do software switching costs disappear entirely?
- 20:33Oh wow. And if that pain of migration
- 20:35vanishes overnight, what happens to the predictable software Moat
- 20:39that this entire $25 billion portfolio is currently betting
- 20:42its future on? That is a great question.
- 20:44Challenges The very definition of a permanent competitive
- 20:47advantage in the AI era. It's definitely something to
- 20:50keep an eye on as this technology accelerates.
- 20:52Until our next deep dive, keep asking the hard questions and
- 20:55stay curious. See you next time.