To The Moon

Two of the most highly anticipated company earnings reports after the bell on Tuesday, $Alphabet(GOOG)$ $Alphabet(GOOGL)$ and $Advanced Micro Devices(AMD)$ both ended up plunging on poor expectations, but while one was able to plow through, the other couldn't.

Google

Main issue:

• Cloud growth is poor, +30% YoY, quite a bit short of Consensus' 32% (buyers expect more);

• Capex is increasing instead of decreasing, with 2025 expected to be 75B, exceeding Consensus' 60B; raising Capex is not uncommon, mainly due to the lack of prior communication with the market and expectation management (the exact opposite of $Meta Platforms, Inc.(META)$ );

• In its guidance for Q1, management mentioned revenue headwinds, on the one hand, currency exchange rates and on the other hand, potential changes in cloud revenue growth rates

But I think there's a good chance that these issues will be downplayed by the market at this point for a number of reasons:

1. A major reason for the less-than-expected cloud growth rate is the capacity constraint issue (at least that's what the company says), so the investment in AI infrastructure (Capex increase) is to address this issue, so it's understandable;

2. As for the efficiency gains brought by DeepSeek, to a certain extent it can also benefit Google's own big models; and there's not a very direct hit on the demand for cloud services, more instead with $Microsoft(MSFT)$ Azure and $Amazon.com(AMZN)$ AWS competition, but the other two also have this problem;

3. The market this year, the main wind direction from the "AI hardware (chip)" want to "AI application layer" change, so if not AI monetization ability to be "continued to prove false", the market will still have money willing to rush in.There will still be funds willing to rush in.

AMD

Key Considerations

• Client business market share increased, confirming the overall rebound of the PC industry; meanwhile, gross margin was flat with market expectation, rebounding to 54%, and operating margin improved;

• AI arithmetic represented by the data center business, on the contrary, there are worries, although +8.7% QoQ, an increase of 300 million, but compared to the Q2 and Q2 of the chain of +500 million and 700 million U.S. dollars, in fact, is considered a decline

• Meanwhile guidance for Q1 of $6.8-$7.4B (market expected $7.05B) was down 3.4%-11.2% YoY, also representing the company's expectation that it is not very strong,;

The issue could fester even further as the

1. DeepSeek brings efficiency gains, which will have some impact on the demand for computing power, thus affecting upstream capital expenditures;

2. Custom chip ASIC will instead erode the market share of the computing chip, thus bringing a bigger blow to AMD, the second echelon of the GPU industry, on the contrary, the first echelon (unique) of $NVIDIA(NVDA)$ is not the main victim.

The essence of this logic is the leading effect: AMD's GPU chip as a replacement for NVDA, if the NVDA production capacity to go up, the market demand began to fluctuate, but instead the first to be abandoned by the customer; when the big manufacturers to reduce the NVDA's GPU demand (ceded to the ASIC chip), the other use of AMD chips, customers can instead go to the use of NVDA chips.