
戈多 Godot
戈多 Godot
前 Manta Network 首席研究员,专注美股期权和基本面。
2Following
256followers
Feed
Livestream
Feed
There is also a ghost story: if the funding bill is not passed before December 11, the US government may shut down again.
Trump wants to increase US dollar defense spending and allocate funds for the Iran war. The Republicans hope to push the budget before losing control of Congress in January next year.
The Democrats want to delay the funding negotiations until after the new Congress takes office.
Then in February-March next year, there is also the issue of the US federal debt ceiling.
The 2011 script is repeating itself, really 🥲🥲

When I previously worked at a major internet company, I had multiple interactions with colleagues from Baidu.
Within the limited sample I had access to, here are the three most memorable impressions about Baidu.
1) BlueFocus was previously Baidu's exclusive PR supplier.
They quoted Baidu's Weibo content creators and KOLs at nearly double the price.
One of my internet mentors, formerly from Baidu's PR department, could get the same quotes at 60-70% of that price.
But he was pushed out.
The reason wasn’t that he took someone else’s kickbacks, but that he was too capable, making others look incompetent and unable to tolerate him.
2) One year, Baidu’s local HR wanted to organize an event with a great idea and confidently proposed it to headquarters.
Everyone thought it was good, but the headquarters HR rejected it.
The headquarters HR felt that since they were at headquarters, they didn’t want to cooperate with local teams on execution. Even if executed, the credit would go to others.
So they spoke ill, obstructed in various ways, and ultimately the event was not held.
From then on, headquarters HR and local HR started various conflicts, small fights when nothing was wrong, big fights when there was.
3) In 2015, I was at a self-media company, dining with Baidu people as a Baidu supplier.
At that time, apps like Jinri Toutiao, Yidian Zixun, Sohu News, etc., were booming on mobile.
I asked Baidu if they were worried and how they planned to respond.
He asked me back, what is Jinri Toutiao, why worry? They do news, we do search, not the same track.
I said, but everyone only has 24 hours a day, and there are over 30 frequently used apps on phones, attention is limited.
He said it didn’t matter, that was something the boss should consider.
Also, the Baidu mid-to-senior management or former mid-to-senior management I directly interacted with, when facing issues, their first reaction was: whose responsibility is this, and how to prove we are not responsible?
The second reaction was: whose turf is this? Whose turf is that? Should we handle this? If we do, will we get credit? Will it be visible to the boss?
As for whether it benefits the company, I never saw such considerations.
That’s about it.
How has Baidu managed to be ten steps ahead in every major industry technology narrative over the past decade or so, yet ultimately end up a hundred steps behind? It has strategic direction but lacks strategic resolve, poor tactics, and internal governance issues. In the end, its inherent advantages are being gradually exhausted!

Tonight Nvidia's earnings report is no longer that important; the conference call is more important.
Even if the answers are good, a rebound is not guaranteed, but if the answers are bad, it will definitely be a disaster.
Especially since @SemiAnalysis_ has downgraded AI asset spending for the coming years.
Guidance for next quarter, upcoming capital expenditures, and how Jensen Huang's departure will impact and be handled, and whether previous contracts will change.
Tonight is serious, it could even decide the fate of some Neocloud.
Just thinking about it is nerve-wracking 😢
$NVDA $NBIS $CRWV
$PANW $IOVA $HPE $DELL are worth watching.
All have increased more than 30% in 6 months or 3 months, ranking among the top gainers.
Afterwards, bulls take profits, and the stock price halves or nearly halves.
But recently, they have all found support at a certain level, and volatility has narrowed.
The fundamentals in the earnings reports have not deteriorated. The stock price will recover.
The logic is similar to buying MEME for the second wave of rise, but the certainty is much higher than MEME.
Stop loss should be set at the recent sideways range's previous low, preferably controlled within 3%.



Except for a few like AMD and ANET that have crossed above all moving averages, most semiconductors still struggle to break through EMA21 and EMA9.
Moreover, the rebounds are repeatedly pushed back. Strong at the opening but weakening by the close.
The situation still looks rather bleak.
The recent main theme in semiconductors is a rebound short.
Since 1958, in 17 midterm elections, the S&P has achieved positive returns six months after the election every time.
Within six months after the election, the maximum drawdown exceeded 5% only 3 times.
In 2018, there was continuous rate hikes combined with the US-China trade war.
About one and a half months after the election, the S&P dropped 14.7%, bottoming on December 24.
By February 15, 2019, it recovered the election day level, with a cumulative increase of 4.6% six months after the election.
————
In 2002, the tech bubble coincided with the risk of the Iraq war.
After the election, there was a four-month adjustment, with a maximum drawdown of 12.5% on March 11, 2003.
By April 23, it recovered the election day level, with a cumulative increase of 0.5% six months after the election.
————
Currently, there are no recession signals in the economy.
Because the current economy still has resilience, combined with inflation, rate hikes directly put pressure on the market.
————
Interest rate futures pricing shows about an 80% probability of a 25 basis point hike in December, and about an 84% probability of a cumulative 50 basis point hike by July 2027.
This differs from the Fed's dot plot and is an important reason why the market will fluctuate in the future.
Currently, the market's pricing of rate hikes is insufficient.
————
The current main theme in semiconductor trading is credit, funding costs, and leverage clearing.
The US 10-year Treasury yield remains high, and the high-yield bond credit spread has widened to 348 basis points.
High interest rates increase financing costs for data centers, and the widening credit spread further raises borrowing costs for companies like Neocloud.
Additionally, with Anthropic's IPO valuation repeatedly dropping, the market doubts the return on AI capital expenditures, putting semiconductors in a state of market cap destruction.
——————
As a result,
After the midterm elections on November 3,
Out of 41 local votes related to data centers, 40 leaned towards restricting data center development.
The newly elected Democratic governor of Texas also proposed suspending approval of new data center projects until the state legislature enacts relevant regulatory measures.
————
Semiconductors had already priced in very high growth expectations.
Crowded positions and leverage further amplified the decline.
From June 30 to November 3, future earnings expectations for semiconductors were raised by 30%, but the forward P/E ratio dropped by 35%.
High growth expectations, concentrated holdings, and leveraged trading make negative news more likely to trigger chain sell-offs.
————
If data center regulations lead to downward revisions in capital expenditures and semiconductor earnings expectations, the market may see a second round of declines.
In any case, it is still best to wait for the Nasdaq or Philadelphia Semiconductor Index to break above EMA21 and then retest without breaking before going long.
The recent main theme in semiconductors is a rebound short.
Since 1958, in 17 midterm elections, the S&P has achieved positive returns six months after the election every time.
Within six months after the election, the maximum drawdown exceeded 5% only 3 times.
In 2018, there was continuous rate hikes combined with the US-China trade war.
About one and a half months after the election, the S&P dropped 14.7%, bottoming on December 24.
By February 15, 2019, it recovered the election day level, with a cumulative increase of 4.6% six months after the election.
————
In 2002, the tech bubble coincided with the risk of the Iraq war.
After the election, there was a four-month adjustment, with a maximum drawdown of 12.5% on March 11, 2003.
By April 23, it recovered the election day level, with a cumulative increase of 0.5% six months after the election.
————
Currently, there are no recession signals in the economy.
Because the current economy still has resilience, combined with inflation, rate hikes directly put pressure on the market.
————
Interest rate futures pricing shows about an 80% probability of a 25 basis point hike in December, and about an 84% probability of a cumulative 50 basis point hike by July 2027.
This differs from the Fed's dot plot and is an important reason why the market will fluctuate in the future.
Currently, the market's pricing of rate hikes is insufficient.
————
The current main theme in semiconductor trading is credit, funding costs, and leverage clearing.
The US 10-year Treasury yield remains high, and the high-yield bond credit spread has widened to 348 basis points.
High interest rates increase financing costs for data centers, and the widening credit spread further raises borrowing costs for companies like Neocloud.
Additionally, with Anthropic's IPO valuation repeatedly dropping, the market doubts the return on AI capital expenditures, putting semiconductors in a state of market cap destruction.
——————
As a result,
After the midterm elections on November 3,
Out of 41 local votes related to data centers, 40 leaned towards restricting data center development.
The newly elected Democratic governor of Texas also proposed suspending approval of new data center projects until the state legislature enacts relevant regulatory measures.
————
Semiconductors had already priced in very high growth expectations.
Crowded positions and leverage further amplified the decline.
From June 30 to November 3, future earnings expectations for semiconductors were raised by 30%, but the forward P/E ratio dropped by 35%.
High growth expectations, concentrated holdings, and leveraged trading make negative news more likely to trigger chain sell-offs.
————
If data center regulations lead to downward revisions in capital expenditures and semiconductor earnings expectations, the market may see a second round of declines.
In any case, it is still best to wait for the Nasdaq or Philadelphia Semiconductor Index to break above EMA21 and then retest without breaking before going long.
By the way, during World War II, this is how the Italian Navy was doomed. Cynthia, an American spy, seduced a high-ranking Italian naval officer and obtained the complete naval codes.
This allowed Britain to easily defeat Italy in the Mediterranean without losing a single ship.
The key point is, she seduced again later—this time the press officer of the pro-Nazi Vichy French government—and successfully turned him. She even got him to willingly steal the French Navy codebook from a safe.
Later, she married this turned press officer and lived a secluded life together in France for the rest of their days.



