AMD quietly doubles, investors await Su Ma's earnings meeting tomorrow morning

Recently, the surge of AI in the US stock market is undeniable, with NVDA more than doubling since the bottom in April. The four CSPs confirmed continued high growth in Capex during their earnings calls, further strengthening the narrative around AI hardware.

Interestingly, AMD, which has been the worst performer during the AI wave over the past two years, has also doubled since its low point in April this year.

From the candlestick chart, AMD first dropped 50% from the peak of AI enthusiasm in October last year, and then doubled from the lowest point of AI this year, returning to the level of October last year.

1. You might be curious why AMD has risen so much. Did they receive some big AI orders?

This question will need to be addressed by CEO Lisa Su during tonight's earnings call. According to various foreign investment banks' tracking, there are many speculations.

Some say AMD may have already secured AI orders, and there have been recent reports of AMD products increasing in price; others suggest that some orders may not come until next year; and there are even more aggressive views that AMD could capture 20-30% of the AI inference card market in the future, potentially taking market share from NVDA.

Why is there such a big difference in expectations? It is precisely because the demand for computing power from cloud vendors is extremely strong, and NVDA's production capacity cannot meet the needs of various vendors in the short term. However, everyone is waiting for cards to use, so some overflow orders are being directed to ASIC or AMD's inference cards.

Looking at it this way, doesn't this logic seem familiar? If you can't buy NVDA's cards, you might turn to buy AMD's cards. As NVDA rises more, AMD also benefits from the same narrative.

2. But are there really orders? Each foreign investment bank is guessing...

For example, UBS mentioned in its report that thanks to the boost from PCs and data centers, there is an upward trend in Q2 earnings, with the most optimistic estimate being that DC revenue could reach $13 billion in 2026, which would roughly double compared to this year.

UBS stated that the price increase of products is real, and the remaining MI308X inventory can be sold to China, which could potentially bring in $100 million in net profit based on pure profit margins.

However, investors should not expect to see more guidance on GPUs for next year, as it is still too early for AMD to discuss next year. However, we expect sales to grow by 50%, and prices will rise accordingly, raising AMD's target price from $150 to $210.

From UBS's perspective, whether there are AI orders or not, AMD's position is already cheap enough. Assuming they can indeed secure AI orders, AMD, with a market value of over $200 billion, has significant upside potential compared to NVIDIA's $4 trillion market value.

Of course, AMD cannot be fully compared to NVIDIA for valuation; it is still a time for imagination. Everyone is guessing whether there are AI orders. If there are, even capturing just 10% of the market share would be enough to reverse AMD's predicament.

3. The Most Optimistic Narrative Circulating in the Market about AMD

Some experts have stated in interviews that OpenAI's own ASIC may not meet demand in the short term and are negotiating with AMD. OpenAI hopes to achieve a 50% market share for both NVIDIA and AMD in its future demand.

AMD has recently performed well in optimizing and adapting its cards, improving rapidly, and from a cost-performance perspective, it also has certain advantages. It is expected that out of 1 million cards, NVIDIA may occupy 60% and AMD 40%.

If AMD's cards perform as advertised in the future, a 50-50 split is possible.

This aggressive little memo also mentioned that AMD and Oracle have committed to delivering 250,000 units, and Microsoft is expected to purchase about 400,000 MI350 and MI355 units from AMD, which would reach NVIDIA's procurement volume. This totals 650,000 units, including the MI350 and MI355X.

The authenticity of this memo is currently unknown, but based on feedback from foreign investment banks, the numbers should not be that large. If it truly reaches the same procurement volume as NVIDIA, it would be impossible for the market to still be in a guessing phase...

Assuming this memo is true, or if half of the numbers are true, then AMD is still very cheap, and even a further 50% increase would not be excessive So, tonight's AMD earnings call is the time of the biggest expectation gap, let's speculate on a few scenarios.

1. There are indeed orders, and the performance is good, but the orders are not that many. The market might digest this and continue to rise, maybe there will be more orders next year?

2. There are indeed orders, and the guidance exceeds expectations, then it could immediately rise by 5-10%.

3. The worst outcome, the guidance shows not many orders, and the revenue meets buyer expectations, then it might drop back a bit.

However, it has doubled from April to now, which definitely makes sense.

At least AMD's market value is not expensive, reasoning that demand continues to explode, and AMD indeed has reasoning cards. Theoretically, the data gap with NVDA is not particularly large, and there are catalysts for selling to Chinese internet stocks in the future. Compared to NVDA's 40 trillion, Broadcom's 10 trillion, AMD's current market value of 280 billion is too low.

This kind of turnaround logic means that as long as it is verified that a portion of the orders are real, the valuation model will immediately switch from a turnaround to a growth stock. The authenticity of this story may be 50%, but the odds of the story are far higher than 50%, even exceeding 100% is not an exaggeration.

From the perspective of implied volatility in options, the market expects the volatility of tonight's earnings report to be around ±8.4%. Let's wait and see if the predictions from various foreign banks are correct.

Risk Warning and Disclaimer

The market has risks, and investment requires caution. This article does not constitute personal investment advice and does not take into account the specific investment goals, financial situation, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article align with their specific circumstances. Investing based on this is at your own risk