R博士

$ZHIPU originally the response came out at 6 o’clock; it was brushed off with light wording and a reset of the quota for a week. No wonder some big-name customers don’t buy it—after all, it’s the shareholders’ money that’s being burned. And in today’s market, what people dislike most is its exaggerated burn rate.

But it’s true that the incident has been downgraded. For users, a weekly quota really is appealing.

Now all big-model companies use this kind of playbook: reset the quota at the drop of a hat, and use money to silence people. Whether you can keep users later or not—who cares? This isn’t unique to Zhipu; it won’t be the last time either.

Of course, this also proves that AI is still in its early stages—many things aren’t fully developed yet, whether it’s humans or even “AI” like RSI. There are still huge swaths of fields that need to be filled in and improved.

People in the crypto circle should especially understand that!

$ZHIPU continue rolling the position; remember to set a stop-loss. I think this drop isn’t enough yet. The most conservative would be 93, 5 percentage points; the initial estimate is 88, and if it gets worse it may break the new low.

Recently it’s right in the spotlight! With this kind of windfall for the shorts, why aren’t they smashing it properly?

Set a stop-loss—for one, it’s discipline; and two, I’m afraid the company’s PR will be handled properly!

As for it—I think this is the real deal!

$ZHIPU slightly avoids some risks; it really is serious this time. Let’s see how this public relations situation plays out this round.

But $ANTHROPIC has had similar things before too—tracking codes, mandatory data retention policies, and so on. However, I feel like Zhipu has gone a bit too far this time.

1 0-Year Treasury yields break 5, and the toughest head-to-head logic has arrived—so is AI stronger, or are U.S. Treasuries stronger?

No matter how you try to push it down, you just can’t. And indeed, nowadays, besides AI, what else can you buy?

Risk assets turn into cash-flow-generating assets—the most certain kind!

The original conclusion was that a 10-year U.S. Treasury yield breaking 5 was only a matter of time, and once it breaks 5, tech stocks would inevitably fall.

But right now, it feels like AI is still strong tonight.

Keep observing for now—no rush to reach a conclusion.

The longer the 10-year Treasury yield stays around 5, the higher it will rise; tech stocks will still face pressure. This is the bears’ biggest weapon.

If it pulls back—if there aren’t as many rate hikes—if tensions in the Middle East ease—if these “ifs” happen, then which ones do you think would have to occur for tech stocks to rebound?

$SOXL This pullback can’t be avoided anymore—we’re just going to tough it out. Let’s see when things finally stabilize. I just realized I was down that much.

I didn’t add to my position; I only added a tiny bit and set a Dell short-term hedge. I’ll wait and see once this week’s situation becomes clear.

Recently, the hard-core events have been one after another—share allotment $ZHIPU . The Middle East situation is also stuck, and over the weekend another AI slowdown came out.

I don’t really have any better办法.

There’s no real reason to stop-loss, so I can only hold on. Honestly, at least for now, I think AI is still relatively resilient—I just don’t know if there will be any more events that continue to pressure it downward.

The new narrative has begun—are we ready to enter the sovereign AI moment?
This weekend has been packed with information.
~~
A long piece by Amodei traces OpenAI’s pause of its IPO, and it all points to one core issue: RSI and alignment problems. AI development is starting to feel uncontrollable.
~~
In specific terms: “By the end of this decade, there’s a 10% chance of killing all of humanity,” and “Within 6 to 12 months, this group of AI agents may be able to take control of the entire internet through a continuously running zombie network, causing losses of hundreds of billions of dollars—and then the scale of losses would keep growing.”
~~
Now, this narrative seems to be gaining some new traction—sovereign AI and safety AI.
~~

The market may first focus on the most directly reflected issues, such as how the slowdown affects AI hardware—e.g., reduced training has a greater impact on Nvidia. Sovereign AI also implies that pricing won’t be governed entirely by market behavior.
Second, OpenAI’s IPO delay: the market doesn’t fully buy its official, well-rehearsed explanations. The fact that large models aren’t worth that much money is also part of it. This is also a reflection of the market’s reaction to large-model valuations.
~~

I’ve been watching how smoothly AI stocks fall and adjust—there’s still a bit of resistance to the downside. Typically, markets move in the direction with the least resistance. If AI overall faces resistance as it moves down, will it break into new momentum driven by the new narrative? That’s something worth observing and hoping for!

From an early, almost barbaric style of development to a “big alliance” model of more standardized, regulated growth, I think this is the only path forward. In that process, there will definitely be battles over who gets to set the standards. But moving toward compliance and safety is never a decisive blow meant to destroy the other side—it’s instead something that enables broader, more confident adoption.

~~
Next, the two sides—AO—should keep pushing industry consensus. I’ll keep tracking their progress and how corresponding AI stocks perform.
~~
Also, another flag worth watching is PLTR. I haven’t done detailed research yet—I’ll just mention it for now. You can pay a little attention.
~

It’s still not clear to what extent this can be achieved. Once the focus shifts from competing on technological advancement to a broader dimension, things may become clearer.
~~
$ZHIPU Zhipu’s decline is more than expected—just follow along and do a bit along the way; nothing else much.

$ZHIPU It looks like the news about the rights offering should have come out earlier—the stock was run ahead of it, so the sell-off started early. It’s not because of DeepSeek, and it’s not because of the technical issues related to $ANTHROPIC .

Technical patterns will tell us about unknown information. For retail investors, this is a very useful reference indicator.

Glioseep’s big drop isn’t because it wants to raise funds through a rights offering and bond issuance—everyone knows that at this stage, large models rely on financing. The key point is that it needs to raise a large amount of funds again within just two months.

So is this about listing and printing money through stock offerings, or is it about steadily doing real business?!

Given the ongoing uncertainty from dilution, expecting the stock price to rebound is a rather troublesome scenario. I also honestly didn’t factor this situation in—but it’s something that’s bound to happen sooner or later.

Of course, it’s also possible that this is a business competition strategy. Whether it’s useful or not, prepare provisions for winter: at the same time, skim off the money in the market. For other competitors—especially those not yet listed—to raise funds or get listed, they’ll find that the water in the pool has already been drained out by a large amount, which is the advantage of listing early.

However, I think it’s more likely they should speak with one voice. Work together to achieve technical breakthroughs, make the cake bigger, and get the operating model running smoothly!

~~
Back to the stock price: I think if it rebounds, we should watch how much resistance it faces when it gets back above the rights offering price of 714—because shares from the rights offering can be sold anytime, unlike A-shares where placements are locked up.

If it falls, at what point will it stop declining? This could be the psychological price level for long-term investors in large models (and this level may also partially include the equity dilution from another potential rights offering!!)

These are price-side disturbances. On the value side, I currently don’t see much impact!

~~
DeepSeek’s new technology reduces reliance on HBM and NAND, and it has knocked down the share price of SanDisk a bit too. This doesn’t prove much. The market’s willingness to take on supply is still there, but the valuation is indeed high. Market participants generally don’t have especially strong conviction in holding the stock; once there’s a slight gust of wind, they think “the wolves are coming,” and money runs for the exits.

This creates a rather stuck, stalemated situation. My suggestion is to keep a low allocation for storage.

~~
Overall, there’s no need to make any special response yet. The logic analysis has already been covered in the earlier articles. But if you’re not personally in the situation, I also suggest waiting until the situation becomes clearer, then moving slowly.
~~
That’s it.

$DELL slaps your face, I really underestimated you. Tell me one thing—bulls rely on faith