**■ Money Show ‘Kang Jung-soo’s AI Investment Strategy’ – Kang Jung-soo, Director of Bluedot AI Research Center**
**[Review and Analysis of the Previous Day’s Market]**
**◇ Let’s start by discussing the closing status of the New York Stock Exchange overnight. What did you make of it?**
– All three major indices of the New York Stock Exchange rose… Nasdaq up 2.69%
– Philadelphia Semiconductor Index at 6,703.20… up 4.63%
– Alphabet up 6.3%… “Continuous praise for Gemini 3.0”
– Sam Altman: “OpenAI now chasing Google”
– Musk acknowledges Gemini’s performance with “Congratulations”
– Gemini 3.0 suggests new possibilities in the AI industry
– Google constructs Gemini system centering on its chip TPU
– Broadcom up 11%… “key partner in Google’s TPU manufacturing”
– CME FedWatch: “85.1% probability of a 25bp rate cut in December”
– Amazon invests $50 billion in AI infrastructure for US government
– Xi Jinping: “Maintaining positive momentum in US-China relations”
– US Commerce: “Trump will make the final decision on H200 exports”
– Waller, Federal Reserve Governor: “Supports rate cut in December”
– US Commerce: “Saudi semiconductor agreement could be a model”
– US Senate urges investigation into ‘Meta scam ads’ to SEC, etc.
– Michael Burry criticizes Nvidia after starting a paid blog
– MS struggles to expand AI chatbots despite ‘dominance’ in the corporate market
– US Treasury: “Economy remains resilient despite 43-day shutdown impact”
**◇ As concerns over the high valuation due to AI bubble theories remain persistent, market uncertainties continue. How do you view the current fear surrounding AI?**
– Disputes over ‘AI bubble theory’… What’s the fact?
– AI bubble theory emerges due to shareholder return structure and circular trading risks
– US big tech issues corporate bonds worth 132 trillion won in two months
– Bond prices of Meta, Oracle, etc., fall due to bond flooding
– CDS trading volumes of big tech like Oracle reach billions of dollars
– Increase in real demand indicators and order backlogs…”Not an AI bubble”
– Nvidia reaffirms sustainability of AI demand with its performance
– Berkshire invests 6 trillion won in Google Alphabet stocks
– Governments establish national-level ‘Sovereign AI’
– Jensen Huang: “Nvidia GPUs sold out”… rebuts AI bubble theory
**◇ Nvidia also directly refuted the AI bubble theories. It is said to have unusually distributed an official rebuttal document to major investors. It emphasized that most of the debates over bubbles and performance stem from data errors or misunderstandings?**
– Jensen Huang rebuts AI bubble theories point by point?
– Jensen Huang: “Only a small portion from circular sales… Inventory at normal levels”
– Sent a 7-page ‘AI Bubble’ rebuttal document to global shareholders
– Jensen Huang: Rebuttal on AI bubble theories “mostly due to data errors”
– Rebuttal against ‘AI Bubble theories’… Jensen Huang: “Unfair to say it fuels the bubble”
– Nvidia directly counters controversies over sales bonds, inventory, and cash flow
– Nvidia: “Days sales outstanding is 53 days”
– Nvidia: “Delinquent sales bonds are negligible”
– Nvidia: “Quarterly operating cash flow figure is $23.8 billion”
– Nvidia: “Structurally impossible to inflate sales”
– Nvidia states SoftBank and Peter Thiel as mere individual investors
**◇ OpenAI is still struggling with deficits. According to financial reports, it is projected that annual operating losses could reach $74 billion by 2028. Among AI entrepreneurs and investors, it has also been identified as the second most likely AI company to collapse first.**
– AI bubble theories repeatedly resurface… Due to OpenAI?
– OpenAI identified as disreputable in surveys among AI entrepreneurs and investors
– OpenAI, second most likely to collapse due to deficits among AI companies
– Companies likely to collapse during deficits… Perplexity and OpenAI ranked first and second
– OpenAI’s deficit could also determine the fate of many big tech companies in the future
– OpenAI is expected to record a deficit of about $74 billion by ’28
– OpenAI’s circular trading structure increases financial burdens
– OpenAI earns 6 trillion won in H1 but incurs 10 trillion won in operating losses
– OpenAI CEO Altman: “Will continue investment despite deficits”
**◇ As soon as Google released ‘Gemini 3 Pro’, reviews emerged in the tech industry that it surpassed OpenAI’s latest model. This time, did Gemini really secure a definitive victory, and how do you view the critical performance differences between the two models?**
– Google’s ‘Gemini 3’ speeds up… Major features?
– ‘Gemini’s counterattack shakes ‘ChatGPT’s dominance’?
– Google narrows the gap in generative AI technology and monthly user numbers
– ‘Gemini 3 Pro’ excels in image interpretation and coding
– Image generation tool ‘Nano Banana Pro’ gaining popularity
– ‘Nano Banana Pro’: “More like a photograph than a photo”
– Gemini evolves to understand user intentions better
– ‘AI Overview’ expands to a level of 2 billion users per month
– Gemini’s 37.5% correct answer rate on ‘humanity’s last test’
– Altman: “Google’s AI advancement could cause economic headwinds”
– Google secures a lead by solving ‘pre-training’ issues
– Pre-training involves learning vast data for initial AI architecture
**◇ IBM was the first to create a smartphone, but it was Apple that ultimately dominated the market. There are predictions that Google could reshape the AI market by leveraging its strength as a ‘full-stack’ AI company. How do you foresee the competitive landscape of the AI market unfolding?**
– Gemini VS ChatGPT… Who will be the final winner in the AI competition?
– ‘Full-stack AI’ Google… Securing the entire ecosystem of AI technology
– ‘Gemini’ increases monthly users from 450 million to 650 million
– Gemini’s technological prowess VS ChatGPT’s popularity
– Sam Altman calls Gemini 3’s edge “temporary”
– Gemini’s integration with search engines and Chrome could expand synergies
– ChatGPT can run Spotify and Booking.com internally
– ‘ChatGPT Atlas’ launched… Challenges the search market directly
– OpenAI fights back by incorporating health and shopping: “ChatGPT as a super app”
**◇ There is also analysis that Google’s strategy of expanding TPU usage on its cloud could pose a greater burden on AMD, which is in a chaser position, rather than Nvidia, the dominant market player. What are your thoughts?**
– Google’s own TPU… More burden on AMD than Nvidia?
– ‘Gemini 3 Pro’ trained on Google’s own TPU
– Nvidia possesses the CUDA ecosystem and overwhelming market share
– Google upgrading AI infrastructure with its own TPU and CPUs
– Google’s strategy to increase computing speed and reduce costs using TPUs
– “Expensive GPU luxury”… Accelerating ‘escape from Nvidia’ with efficiency
– AMD once held 20% of the GPU market… Now below 10%
– Google could absorb potential AMD customers and demand internally
– Competition in the AI accelerator market intensifies, increasing pressure on AMD
**◇ Among big tech companies, Meta is notably aggressive in investing in AI and the metaverse, but challenges and skepticism about profitability persist in the market. When do you foresee Meta’s bold bets translating into corporate value?**
– AI and stock market barometer—Not Nvidia, but Meta?
– Meta developing ‘morning AI briefing’ in ChatGPT style
– AI big tech’s 132 trillion won storm of bonds… Meta’s credit concerns rise
– Meta to invest $880 trillion in the US AI industry over three years
– Despite AI turbulence… Retail investors focus on Nvidia and Meta
– BNP Paribas: “Meta’s AI infrastructure expansion erodes profitability”
– Meta once wholeheartedly dedicated to building the metaverse… Now lagging in AI investment
– Advertising-centric business model leads to slower profit recovery
– Meta largely relies on Nvidia products for AI training GPUs
**◇ Let’s move on to Tesla. Tesla’s stock price has only risen modestly by about 5% since the beginning of the year. When and where do you see the momentum for a trend reversal in stock prices emerging?**
– When and where will Tesla’s trend reversal momentum come from?
– Tesla begins developing the next-generation ‘AI6’… AI5 mass production imminent
– Musk: “Tape-out of Tesla AI5 semiconductor imminent”
– Tesla introduces supervised FSD in Korea… Applies HW4 specifications
– 41% of EV owners across 30 countries say “Tesla boycott”
– Tesla’s ‘door handle’ safety controversy continues and faces lawsuits in the US
– Peter Thiel sells Nvidia shares… Reduces Tesla’s proportion
– Tesla pursues excluding Chinese parts in US-produced EVs
**◇ Tesla’s ‘Cyber Cap’ has yet to receive sales approval, while Chinese company Baidu’s robotaxi has reportedly reached its break-even point?**
– Which Chinese robotaxi company has surpassed the break-even point?
– CNBC: “Robotaxi, China is far ahead of the US”
– Baidu operates over 1,000 units of ‘Apollo Go’
– Baidu’s robotaxi surpasses the break-even point, in contrast to the US
– Waymo operates over 2,500 robotaxis in test runs in the US
– Waymo plans to start paid operations from next year
– Tesla begins limited autonomous robotaxi service
– Tesla’s robotaxi operates with safety officer onboard
– Robotaxi market expected to grow over 36 trillion won by 2030
(For detailed information, please watch the video.)
**Your tip can become a news story.**
**SBS Biz is waiting for your tips.**
**Website = [https://url.kr/9pghjn](https://url.kr/9pghjn)**