February 3–7: NAS100 Market Outlook & Has the Market Peaked? February 3–7, 2025 (The following is solely personal opinion, not investment advice. Please conduct your own assessment before making any decisions.)
This is a market context update using Elliott Wave Theory. As shown in the chart, considering the Nasdaq bull market from the 2008 bottom to the present, there is a high probability that we are currently in a B Wave rebound within the ABC corrective wave. However, it remains uncertain how high the B Wave can rise.
According to Elliott Wave Theory, the C Wave correction should fall below the bottom of the A Wave, which means for the Nasdaq cash index, the price could drop to at least 10,600, representing a 50%+ correction. Of course, this process may take several years.
Scenario 1: The Market Has Not Peaked Yet
I am watching price action within the 22,465–23,001 range. If this scenario holds, the weekly price needs to stay above 19,200, and the daily price must remain above 20,300. From a macro trend perspective, the market can still be cautiously bullish.
Scenario 2: The Market Has Already Peaked
If the price breaks below 20,300 quickly and remains below it, this could be the first technical confirmation that the market has topped. If this happens, we should see further consolidation around 21,000 next week, followed by a breakdown. However, if the price finds support above 21,200 and breaks upward, there is still a possibility of new highs and further gains.
Last week, Nasdaq 100 moved in line with my previous analysis. On Friday, the price rebounded to the 21,600–21,700 target range before declining. If this decline breaks 20,300 swiftly, the probability of the market already peaking increases significantly. Typically, downward movements are faster than upward ones. The market took four days to rebound to the target range, so if the current trend holds, we should see a breakdown early next week (Monday–Tuesday) to confirm the trend. If the price fails to break lower between Monday and Wednesday and finds strong support or rebounds, the market could push higher and break new highs—this needs further observation.
Deepseek & AI Industry Observations
Over the past few days, I’ve read more about Deepseek and industry insights. One of the key cost-control strategies for training models is distillation, where a large model is used to train a smaller model. This suggests that Deepseek’s smaller models depend on larger ones and have certain limitations.
Many industry experts have also stated that Deepseek’s emergence does not reduce AI investment and that the market may have overestimated its impact.
As I am not an AI expert, I cannot verify the accuracy of these claims or whether Deepseek’s impact is overestimated. However, Anthropic CEO Dario mentioned in a post last week that AI model training costs are decreasing by 75% per year.
A model that costs $100M to train in 2024 will only cost GETTEX:25M in 2025, and by 2026, the cost may drop below $5M.
Dario also noted that this cost reduction trend is accelerating, meaning that early investments in AI without strong technical barriers and market dominance could be wiped out.
The next few years will see explosive AI development, but many AI investments may fail.
Competitive Perspective: Can the U.S. Maintain Its AI Lead?
From a competitive standpoint, I do not believe the U.S. can maintain its AI dominance solely through investment.
I strongly agree with Nvidia CEO Jensen Huang’s AI development roadmap:
The Four Stages of AI Development
1. Perception AI
2. Generative AI
3. Agentic AI
4. Physical AI
According to Jensen, we are still in the Generative AI stage, where training models mainly rely on internet data. At this stage, the U.S. holds a huge advantage due to its leading semiconductor technology.
However, when AI enters the physical realm, internet data will no longer be sufficient, and stable, large-scale data sources will be key to AI success. In this phase, China and India, with their vast populations, will have a major advantage.
Since the Huawei incident, China has accelerated its chip R&D and manufacturing. While still lagging behind TSMC, tighter trade restrictions will further drive development in this sector.
If a competitor to Nvidia emerges in China, what impact would that have on the high-valued U.S. tech sector?
AI Bubble & Investment Strategy
I believe Deepseek is just the beginning of exposing the AI bubble. While its direct impact may be overestimated, the fact that AI training costs are falling rapidly is undeniable.
Investment Perspective
From an investment standpoint, I believe that market price action is the ultimate confirmation.
- For long-term views, take a macro approach.
- For short-term decisions, use technical analysis.
