[Red Envelope] Institutional rebalancing! Quantitative trading is rampant! Adjust your mindset, institutional golden opportunities will come too!

These past couple of days, market sentiment may not be very good [Taoguba]

There are two main reasons: rampant quant trading and institutional rebalancing!

These two reasons can actually be viewed together.

Because of institutional rebalancing, the uncertainty in individual stocks increases. When institutions sell, they do so across sectors, not targeting specific stocks. So even good stocks face the risk of being unfairly sold off during this phase.

On this basis, due to institutional rebalancing, the growth momentum of retail funds can’t get going. Many retail traders may be stuck in a few stocks, unable to buy or sell freely.

As a result, they might be passively locked in.

Institutional selling and retail lock-in, with no additional volume, lead quant traders to focus on small-cap stocks, playing their own game.

In other words, institutional rebalancing is the cause, and rampant quant trading is the result!

Based on this, when trading, pay attention to two points:

First, subjectively avoid high-position stocks in institutional sectors!

For example, the liquid cooling sector

Over the past year and a half of the bull market, the liquid cooling sector has quadrupled in value!

Remember, Zhang Kun, once a legendary figure in the liquor industry, only achieved a doubling of performance on paper.

Just think about how exaggerated this data is.

Second, the interests of public fund managers and investors are often misaligned!

The sector has increased fourfold. Core stocks like Yingweike, from low levels, have risen up to six times, and some more elastic stocks may have even more.

Moreover, from what I’ve learned, most institutions expect to finish their rebalancing by the end of Q1.

Now, it’s already mid-March, so the timing is about right.

Second point: Appropriately abandon overthinking and subjectively favor quant-preferred stocks!

Some teachers have pointed out that recent rallies mostly target sectors that experienced sharp declines the day before.

This is precisely a characteristic of quant trading.

For quant funds, a sharp drop means opportunity, because they are the largest market participants. Nothing can lift stocks without their involvement!

Generally speaking, March is the earnings announcement season, and historically, this period is mainly for performance-driven trading.

But this year, aside from poor earnings leading to losses, even stocks with good performance haven’t seen much upward movement.

For example, Baiwei Storage, with strong earnings, only saw a rise of less than three sectors.

The main reason is still related to institutional rebalancing.

Institutions are rebalancing, and semiconductors are a heavily rebalanced sector, heavily held by institutions.

If institutions want to sell high-performing sectors and buy sectors with less growth, semiconductors could easily trigger imagination.

Personally, I think semiconductors are still relatively low.

When might the market adjust? Probably around late March or early April, after earnings season ends!

Once earnings disappointments are out of the way and institutions finish rebalancing, the market will become easier to operate.

In the near future, focus will mainly be on quant stocks and sentiment-driven stocks.

Previously, I shared some logic: individual stocks can’t be discussed in detail, but I can give some conditions and reasoning—ask Doubao!

To avoid this post being swallowed, please give more likes! Share and support! Teachers, each tip 200 points!

First, about power sector

This sector is actually at a relatively low position. Historically, it has been valued as dividend stocks, so valuations are generally low.

Earlier this year, during the hype around computing power going overseas, the power sector was also briefly promoted.

This is unreasonable—power exports are just repackaged old wine, fooling the naive.

But why has this hype lasted so long?

It’s related to the upcoming two sessions discussing integrated computing power!

What does integrated computing power mean?

Simply put, computing power is in high demand because AI is developing rapidly. But data centers, supercomputing clusters, and large model companies are mostly concentrated in the eastern regions.

These power-consuming giants are competing with eastern prosperous areas for electricity, while leaving cheap western electricity unused, leading to resource waste.

Therefore, new policies now require coordinated development of computing and power!

New data centers and computing centers must use green electricity, improve energy efficiency, and use nearby cheap power.

In this context, what could have been solved with a bit more spending—raising prices to downstream—now requires green power, making it impossible without green energy.

This shifts the focus to infrastructure companies.

Some power companies, traditionally valued as infrastructure firms, need coordination and dispatching for power output.

Now, the market has shifted from a buyer’s to a seller’s market. I can decide which computing power company to sell electricity to.

In the future, companies with green electricity, especially state-owned enterprises, could be valued like tech firms.

From this perspective, their stock prices could multiply several times.

A very interesting stock here is GanNeng Shares.

This is one of the benchmarks in the photovoltaic sector, and also somewhat in eastern China, so today it was directly pushed to the daily limit!

No recommendation here—just sharing the logic and reasons for the limit-up!

But conversely,

Most domestic green electricity still comes from Xinjiang (desert PV, wind power), Inner Mongolia, Tibet (including Yajiang hydropower).

These areas generally have less favorable climate and conditions compared to the east.

Establishing subsidiaries there just for this purpose isn’t very practical.

If you insist on integrated computing and power, you’d have to sign agreements and transmit power over long distances!

What’s needed for long-distance transmission?

Of course, ultra-high voltage lines. We are globally leading in this area, with almost negligible losses!

From this perspective, future demand for computing power will grow exponentially, and electricity demand will follow suit. Will the components for ultra-high voltage lines also see a surge?

Could this lead to a major company emerging?

Transmission has peaks and troughs—for example, PV in Xinjiang, even with long sunlight hours, still has night-time downtime.

But large models don’t rest at night, right?

So energy storage is needed—storing excess electricity when abundant, discharging when scarce.

Currently, this is still a relatively future concept, with few similar views on the market.

In the current hype stage, it’s mainly about green electricity, and ultra-high voltage isn’t yet a focus.

Continuing on integrated computing power

Since we’ve discussed integrated computing power, let’s talk about cost issues.

We mentioned that western green electricity is cheap.

But once integrated computing power becomes widespread, behaviors that boost western economies will no longer keep electricity prices so low—market competition will drive prices up.

Whoever owns green energy resources and land can monopolize the cost of computing!

I want to mention a company.

This company isn’t doing integrated computing power, but is a core player in the computing sector.

Its current strategy is to operate with a light asset model.

Similar to Wanda Commercial Management—originally building real estate with large capital and long turnover.

Now, they cooperate with real estate developers, building Wanda malls on project sites, handling leasing and attracting tenants.

Wanda malls are developed by property developers, and rent from big brands isn’t small.

Because Wanda malls have good operational efficiency and attract consumers, property sales are easier. Developers pay advertising and other fees when selling units.

This company follows a similar approach: I have a computing center, you have one too, but your operation isn’t as good as mine. If you earn 10%, I earn 20%, so you might as well hand over your computing center to me.

You earn 10%, I earn 10%.

It seems I earn less, but I improve capital efficiency, acquire more computing centers, and can invest in upstream and downstream integration. Plus, the fixed assets aren’t mine—depreciation isn’t my concern.

So far, I’ve been talking about sellers.

Next, let’s discuss buyers.

How to create greater value with the same computing power?

In the past, a large model with minor modifications had to send samples to Nvidia.

Nvidia would fine-tune chips and graphics cards for that model.

But after DeepSeek released a new model, they sent samples to Huawei instead of Nvidia.

This means DeepSeek is trying to compensate for domestic hardware limitations through architectural changes.

In fact, the performance gap between domestically produced Ascend chips and Nvidia’s products isn’t huge.

Previously, these companies competed fiercely, willing to pay a 30-50% premium for just a 10% performance edge.

Now, DeepSeek’s move shows that domestic hardware has made breakthroughs, and with architectural adjustments, it might outperform Nvidia in the future!

Performance has improved, and costs have decreased.

From this, domestic hardware, especially Huawei’s Ascend series, has huge potential.

In recent years, terms like domestic substitution, independent controllability, and Huawei have often led to major stocks.

Including, but not limited to: Huali Chuangtong, Jierong Technology, Changshan Beiming, Shenzhen Huawei…

The world’s AI is mainly dominated by China and the US.

Our AI has, to some extent, surpassed the US. But because hardware has long relied on Nvidia, which profits and subsidizes its supply chain, the competitive advantage isn’t obvious.

However, if future hardware switches to domestic products, the gap between East and West could widen instantly.

US investors (mainly institutions) might shift funds to the East, enabling us to do more and develop faster!

If all those orders switch to domestic, the impact is significant.

If Huawei’s supercomputing clusters are deployed, will they need more liquid cooling? More design architecture?

Although liquid cooling is a sector where institutions are rebalancing, short-term opportunities might still exist.

Talking about Huawei, let’s briefly mention Lobster!

Lobster has been very popular recently.

But many people don’t really understand what Lobster is.

Briefly, it’s a software with high permissions that can access most apps on your computer and connect their functions.

(It’s basically high-level access, which can be done domestically, but domestic companies find it hard to pass compliance. Lobster is a personal program, initially non-profit, so it’s not constrained by these issues.)

What’s impressive about Lobster is that it can cut off error paths and ensure they are not repeated!

From this perspective, it has unlimited growth potential.

As long as it can eliminate errors, it can keep growing!

But from another angle, Lobster’s height is limited by the user’s own level.

If you don’t even know which path is correct, how can you help others?

So, from this perspective, stock trading with Lobster isn’t very scientific.

Most of the hype is from salespeople wanting to make money.

Main reasons:

First, Lobster’s trading ability depends on the user’s skill.

If you can consistently profit, Lobster might help, but not guarantee stable profits.

Second, if Lobster’s strategies become homogenized, it will still be vulnerable.

As we discussed last time, fixed formulas for trading will be beaten by quant funds, not by you!

Even if your program is identical to a quant’s, what fiber optic cables are used? Same speed? Same hardware?

Can your trading speed match?

They support quant trading when buying, and backstop when selling.

But once Huawei and similar companies get involved, privacy issues can be somewhat improved!

Could this lead to a new mainline?

I’m optimistic about this round of 20cm stocks, but currently, stocks like 688 are doing well, though retail support is weak.

Not sure if this cycle will shift away from retail investors toward institutional speculation!

Finally, to prevent posts from being swallowed, please like, share, and tip more!

See you at our live broadcast at 9 PM!