올해 읽은 최고의 보고서. 10분만에 요약합니다

0:00 Well, today I have one or two of the greatest reports ever . This is an AI-related report from a place called Kotu. Now, if you think of this company as a combination of venture capital and private equity, you can think of it as a company. In particular, we invest heavily in technology-focused unicorn companies . But I recently published a report on AI, and it's over 100 pages long. But I really like this content, so I wanted to share it with you . I'll compress it a lot and just give it to you, so I hope you watch it. So, first, this report covers the AI super cycle. It logically develops which companies are doing well now and which companies and industries to keep an eye on in the future. The beginning of the story is simple. What are the biggest drivers in the AI market today? It's demand. And the demand is growing at an unimaginable level.

0:46 Amazon, Microsoft, Google, Meta, Oracle, and Tesla are the six companies that invest the most in AI . These are AI infrastructure GPUs, right? The amount they announced they would invest here in January of this year was $31 billion. It was about 400 trillion won in our currency. But what happened during the May earnings season when the company was announcing its earnings? This increased by a whopping 20% to $365 billion . They said they would burn an additional 80 trillion won in just four months. In just a few months, the amount of money equivalent to a national budget was increased. Okay, so here's where the first question should come in. No, why are big tech companies pouring so much money into this ? Isn't this crazy? The answer was right there in this phrase. This is what Microsoft CEO Satya Adelaide said during the earnings call. This time, we processed over 100 trillion tokens, a

1:31 fivefold increase over the previous year . And what's even more amazing is that 50 trillion tokens were processed in the past month alone. Demand literally exceeded expectations. If we keep predicting, the current AI demand is outpacing it, and the upward slope is like a graph of a sine function . From their perspective, Nvidia chips are not an expensive luxury. It costs about 40,000 dollars in chips, but if you buy it for 40,000 dollars and keep it for about two years, you will get back all your investment. So, you ca n't not buy the GPU rate. And now, the second issue is survival . In this technological war, not having a state-of-the-art AI home is like going into battle without a gun. I'm hesitant to invest now because research, development, and service are all now impossible. It's the same thing as saying we're going to get out of Big Tech. However, the recent

2:17 stock performance has been a bit disappointing. Over the past five years, American Big Tech has really taken off. In the case of Nvidia, it had a historically impressive 100% increase, but this year it is a bit disappointing . Oh, and the NASDAQ didn't go up much overall. And then the market starts to reverse again . While M7 was faltering, new companies began to emerge . No, if you ask me which companies came up, you would already know. AI Foundation Model. In particular, companies that create AI-based models, such as OpenAI, XAI, and Entropy, have really risen to prominence, and below them, data infrastructure companies like Coreweave and Databricks, and now companies that create B2C user applications and robotics are rising to prominence . In fact, I think you could find dozens of related companies, but I'll just

3:03 pick a few and explain them. First of all, first of all, this is Emrobin. The country went up almost tenfold and now it's slightly adjusted, so you're probably wondering what kind of company this is? When we use our cell phones, we can see advertisements continuously appearing in apps . The people who put up those ads are none other than Emrobin. There are ads in the mobile application. There is a very high chance that it is Emrobin. These guys are really good at matching ads that are tailored to the user based on AI, so advertisers are really happy with their RAS and are receiving a lot of attention . I really liked it. And the company itself, Emrobin, says it has gained a lot of efficiency thanks to AI. If you look now, in 2021, sales per capita were $3.6 million, but by 2025, sales per capita have more than doubled to almost $7.6 million . yes. When you look at this, is n't it really amazing? One employee

3:50 generates over 10 billion won in sales . I think this is why corporate value increases. And the second company is Notion. Everyone who has used Notion knows that the AI function has been available for about two years now. However, it is said that the number of users using the function has increased from 10% in the past to 51% now . And the next company I'd like to introduce to you is Open Evidence, which is now being used in the medical field. It is said that currently, more than 30% of internists in the United States are using it. What kind of service is this? It's a medical professional who has drastically reduced hallucinations for doctors . If you input the patient's condition, treatment diagnosis, etc., it will find references and provide an accurate answer . This is more specialized for medical purposes than the currently commercially available Jaeminna GPT, so it has a level of sophistication that few other chapoes can surpass . Recently, all the doctors who used GPT have

4:36 switched to this because it is so good. If the medical field has come out, the legal field should come out too. This is the most representative example of legal tech . Currently used by 34 of the top 100 law firms in the U.S., it is showing an absurd growth rate of 240% per year. This is also very useful in practice. Unlike existing chatbots, it is said to have the best performance in the legal field because it was trained using a huge number of legal documents . It is said that they can view various international laws together and are already far superior to the average human lawyer in analyzing cases and writing documents . I thought it would be difficult to become a professional, but now that I hear from people who are actually adopting this, I think it won't be easy to become a professional. It really performs great. More than 30% of people are already using it, and word of mouth has spread so much that it's

5:21 spreading so quickly in the field that it's practically impossible not to use it. And then I think it might be the hottest theme recently. There's something called Neo Cloud, an AI-specific cloud that's emerging to replace existing hyperscalers like Google, Amazon, and Microsoft . Among them, in the case of Together AI, token usage has increased by a whopping six times in just nine months. It's safe to say that this is almost a six-fold increase in sales . And besides Together, I know that there are about 10 unlisted Neoclouds that are emerging these days . By the way, the leader of this Neocloud is Core Weave. The stock price has more than tripled in the three months since the IPO. The results showed that sales increased sevenfold in just one year . The reason Core Weave is doing well is because it is the first to receive Nvidia chips . And it 's also the stock in which Nvidia has invested the most. Anyway, from this huge cloud to this tiny cloud, everything is

6:07 growing at an insane rate. And where did this AI go? This is the world of developers. What's really dramatic here is that the term "vibe coding" is trending . What this means is, just do some coding and do whatever. Please make a site like this . If you look, this is not a site I created myself, but a salty carrot site I created using AI . How is it? Doesn't that sound a bit plausible? From the developers' perspective, a monthly fee of $200 or $300 is n't that expensive. No, if the AI assistant does more than two people's work, then the truth is that $200 or $300 isn't that bad . I'm also using the Ultra version of Claude and Jamina now. Now I'm so used to it that I can't do my job without using AI. I think many developers would agree with this . Thanks to this, in the case of coding AI, the annual recurring revenue almost quadrupled in one year . In the case of Cursor, the company's value exceeded 30 trillion won in just one year after I started talking about it .

6:53 This is going up even faster than Open AI . But as I use it, I feel like it can't be helped. Okay, let's summarize this. The biggest thing that AI recommends is market knowledge labor. Then, when the market sizes of medicine, law, and development are combined, the human rights cost reaches 8 trillion won per year. Many people still naively expect that AI won't take their jobs . But the reality is quite cold. Well, it might not have a big impact for the next year or two. But will you still have a job in five years ? I also think I can be replaced . Almost 95% chance. Well, even as of now, anyone familiar with AI would think that two people can do the work of one person. Of course, now the number one spot will double. However, as AI develops further, there will clearly be limitations compared to the market size that grows . Ultimately, any

7:39 job that can be replaced by AI will certainly be replaced . This is the realm of necessity. Okay, now let's talk about investing . Two months after Trump's tariff shock in May, which plunged 20%, it has now reached a new high. Ultimately, where does liquidity go when interest rates fall and fiscal expansion is in place ? It's heading towards AI. I am very confident in my report and I am also very confident in my report. And while we'll have to wait and see, the inflation figure, the Consumer Price Index, is quite stable despite the tariff rhetoric. Let me tell you another interesting story. Professor Paul Krukuman, who would later receive the Nobel Prize in Economics in 1998. What did this person say? Even in 2005, he was still being criticized and kicked the blanket for saying that the Internet's impact would be smaller than that of the invention of the fax machine

8:25 . Historically, humanity has underestimated the power of technological advancement . In 1993, I estimated that the Internet would increase productivity by 0.8% to 2.8% in 2000. But in fact, in 2000, it went up to 2.0%. No, you might say that 1.2% isn't that much of a difference, but how much of a difference does 1.2% make? Debt-to- GDP ratio was cut by almost half, from 61% to 34% . So, we can say that the country's financial health has improved by about two times. So now, to some extent, the logic that Trump is putting forth makes sense . Well, I think AI is now a cognitive revolution, a new form of industrial revolution . Well, if AI increases productivity, labor costs will be reduced first. As labor costs decrease, inflation rates will naturally fall

9:11 . Then the base rate will fall and interest rates will decrease. Then companies will increase investment and hire more people. So, in conclusion, we can create a virtuous structure that increases GDP and expands tax revenue . If we quantify this, it would result in a productivity increase of approximately 2.8% . As we saw earlier, the internet grew by 2%, but the current consensus among companies is that it will have a bigger impact than the internet . So, this asset management company tells you to invest in the following and gives you a list of about 40 . If you look, there are now about 40 companies, from big tech to Bitcoin, and even start-ups like OpenAI and SpaceX. If you have a lot of money, you can just invest in this fund . But now that it is the third fund, it needs to have a lot of money . And now, let me introduce you to the recent trends in venture capital and third-party fund deals. First of all, IPOs are huge right now. It is expected to

9:58 increase by almost double compared to last year, and it is not just IPOs that are increasing. That's not the problem, M&AD is now looking like it's on track to hit record highs . As you can see, some very large deals are being sold right now. If you look at it, now huge companies like Wizna and Scale are being sold to big tech companies for around 50 trillion won. In conclusion, I had thought earlier that Big Tech might be hesitating, but if you look closely, Big Tech is now acquiring AI companies, so I personally think that M7 will still do well . It practically has everything you need for the AI era . We have money, capital, technology, and data . Of course, I'm not discouraging investing in small businesses, but wouldn't investing primarily in stable big tech companies yield sufficient results even in 2025 ? So, I'll end the video by recommending Big Tech again and again.

10:44 And it takes more effort and time to make this video than you might think. If you could subscribe and like, I would be very grateful. And if you have any disagreements, please leave a comment below. Yeah.