Dear readers,
As we enter the period for reviewing fund managers' latest portfolio updates, it’s valuable to analyze their recent additions and reductions to gain insight into their evolving investment strategies.
It’s important to note that these portfolio disclosures may not reflect their most up-to-date positions, as managers can adjust their holdings after the reporting date. Therefore, this information should be viewed as one component of a broader research process.
Below is a selection of respected fund managers, along with other notable investors, whose investment moves are worth tracking.
Source: Dataroma, compilation by TQI Capital
Q4 2024 Observations: A Significant Slowdown in Buying Activity
One clear takeaway from their Q4 2024 buying activity is a noticeable slowdown. My best guess? Uncertainty—both political (Trump’s re-election) and market-driven (valuation concerns).
His first month in office has already unsettled the markets, introducing trade tariffs (or at least the attempt to), speculating about buying Greenland, escalating conflicts, and even making outlandish claims about Canada being the US 51st states. Trump will always do what he does best, and as investors, our focus should be on the facts—ignoring the noise.
Capital Intensity & The MAG 7 Shift
Sources: X by Modest Proposal
Another striking observation: no major additions to the MAG 7 stocks—except for Alphabet. Perhaps the cheapest among them, yet also the most vulnerable to AI disruption. Alphabet has the most to lose!
However, it may not just be about AI challenges—it’s also about capital intensity. Historically, the MAG 7 thrived on their asset-light business models, allowing them to outperform most companies and grow into industry giants. But now, they are being forced to invest heavily in infrastructure.
This shift is costly, and their ROIC is under question. It’s unlikely to be zero, but an educated guess suggests it will be meaningfully lower than their previous returns. High capex with no clear monetization model is a concern. (at least no visible yet for now)
To put this into perspective, capital expenditures among these big tech companies have surged sixfold—from $51 billion in 2017 to $325 billion in 2024. This is akin to laying down an extensive railway network. The tracks, once built, can enable tremendous efficiency and scalability, but there’s always the lingering question: Will the train be allowed to run at full speed, or will regulatory barriers slow it down?
Much like how railroads were once seen as the backbone of industrial progress but later faced government intervention, taxation, and anti-monopoly scrutiny, today’s tech infrastructure expansion—particularly in AI and data centers—could be met with similar constraints. The investment is valuable, but will it be able to yield the expected returns? That remains uncertain.
Most importantly, the introduction of new technology is set to dramatically reduce the cost of training AI—and I firmly stand on this side. AI should be accessible to everyone, not just controlled by big tech. The democratization of AI will foster greater innovation, competition, and progress, rather than concentrating power in the hands of a few dominant players.
Lemming Effects
I’ve previously predicted that new technology will dramatically lower costs—DeepSeek might be just the beginning. Jeff Bezos often says he prefers to invest in things that are unlikely to change.
No one wants to be locked into a single ecosystem. Everyone prefers lower costs. The idea of only one winner is an unlikely scenario, no matter how strong the moat appears. Let me quote my previous post:
When will the music stop?
I have no idea. And frankly, I am one of the least qualified people to predict this.
However, the potential for a technological breakthrough that reduces reliance on GPUs or data centers in AI is worth considering. While timing is uncertain, I am confident it will happen—because technological progress follows an exponential curve. Betting against innovation has historically been a losing bet.
As history has shown, those who fail to learn from the past are condemned to repeat it. George Santayana’s words ring true not just in human history but in technological evolution as well. By TQI capital
DeepSeek & The Cost Equation
Sources: X by Beff -e/acc
For those unfamiliar, DeepSeek is an open-source AI efficiency model designed to reduce costs. Instead of relying on brute-force training, it prioritizes optimization and efficiency—a key differentiator.
What’s most impressive is that this breakthrough isn’t coming from a major tech giant but rather from an unknown hedge fund side project in China. (The above meme is funny!) This underscores the shifting dynamics in AI development, where innovation is no longer confined to the traditional powerhouses.
Sources: The Culture of Survival and Innovation by World Government summit (scroll to 15.27mins for his view on AI and capex)
Joe Tsai, Alibaba’s co-founder, recently shared an insightful perspective on the AI race, emphasizing that applications—not just raw model capability—will be the real battleground. He likened the AI arms race to wealthy parents spending hundreds of billions to train the ultimate child prodigy, questioning whether such an approach is truly necessary.
This analogy resonates with an emerging consensus: open-source collaboration will likely outpace even the most sophisticated closed AI systems. Instead of centralized control by a few tech giants, AI development may follow a more decentralized path—where efficiency, accessibility, and adaptability triumph over sheer computing power.
The key question remains: Will the future of AI be dictated by a handful of dominant players, or will a collective effort redefine the landscape? My bet is on the latter.
You don’t need a PhD in physics to sell clothes!
Changes in China sentiment
We can clearly see a shift in the China narrative—from "China is uninvestable" to "China is catching up in AI." Many fund managers are now making sizable bets on Chinese equities, recognizing the attractive valuations. Notable investors like David Tepper and Michael Burry have positioned themselves accordingly.
Of course, as always, it’s crucial to conduct our own due diligence—invest in what we understand and invest wisely! (Disclosure: I am vested in the China market!)
Conclusion
As we navigate the evolving market landscape, one thing is clear—fund managers are treading cautiously. The post-Trump rally and the shifting dynamics in AI and capital intensity will shape investment strategies in the coming months.
However, timing the market remains a difficult, if not impossible, task. For long-term investors, staying invested and focusing on fundamentals is the best approach.
As always, a sincere thanks to Dataroma for their excellent work in compiling 13F filings. Their platform remains a valuable resource for tracking fund manager activities.
Disclaimer: I might have positions in the above posts and receive no fees for writing the post. I am not affiliated or have any role with the company. Opinion is my own and this post is just for educational purpose. It is not an advice to buy or sell the stocks. Invest at your own discretion.
Thanks for the overview!
Guy Spier and Li Lu look lazy in this company :)
Glad you like it. Ya but for Li lu, i believe that it is just part of his fund and his major exposure should be in china.