Speaker 1: Welcome to La Economía. The eruption of the Chinese artificial intelligence DeepSeek has unleashed an earthquake in the markets. It is a much cheaper application, free of charge, and according to experts, more efficient than the models of the great American technologies. DeepSeek, this open source code and training it cost, according to its creators, just 6 million dollars, compared to the thousands of millions it has cost to train Chad GPT or Jiminy. In Wall Street, the manufacturer of Chips NVIDIA, the world leader in artificial intelligence components and programs, fell up to 17%, recording the largest loss in stock market value in the history of the stock market. As a whole, the Nasdaq index fell 3.4%, also dragging the world of cryptocurrencies. It seems that the Chinese artificial intelligence has left the United States in evidence. We are looking for the details. Ana Nieto from New York. NVIDIA is the great loser of the great Chinese eruption that affects eruptors in the United States. What other values have been negatively affected?
Speaker 2: Yes, the stock market today has clear losers. The most striking, without a doubt, is NVIDIA, with many consequences for its weight within the indexes. Its valuation has been the star of Wall Street in recent years in anticipation of the strong demand for highly sophisticated and very expensive microprocessors. But if the takeoff of NVIDIA has been spectacular, today its fall has been for Wall Street's history books, because it has been evidenced that spending may not be necessary, and therefore all income and profit forecasts of this company are questioned. Today, the company's capitalization has been reduced by more than half a billion dollars, which is equivalent to the capitalization of a company like Boeing, but four times. Other companies in the sector, such as AMD, Broadcom, Marvel Technologies, as well as the ADRs of companies such as Taiwan Semiconductors or ASML from Holland, have also suffered strong losses that have negatively impacted the SP500 and NASDAQ indexes. Other two companies, such as Apple and Meta, have, however, had a very good evolution today. Those who have also seen how their future scenario has been moved by DeepSeek are the energy companies that were exposed to the great impulse that artificial intelligence was going to give. Many of them have also suffered strong losses of two percent.
Speaker 1: Incredible everything that DeepSeek has moved. Ana, there are big technical questions, but also strategic. What are the most immediate consequences?
Speaker 2: The first ones are in the markets, they are being evaluated by investors. In another context, the president of Microsoft has said, he has reduced the Jevons paradox, in which it is said that the higher the efficiency of a resource, the higher its consumption. That is, it is expected that there will be a much faster adoption of artificial intelligence, since it is much easier to get there. With regard to Washington, two questions. The first, the limitation of exports from the United States to China of this material, of these microchips. And the second, the meaning of the great investments, especially that of Stargate, presented last week by Donald Trump.
Speaker 1: Very interesting everything you tell us. Thank you very much, as always, for the analysis, Ana.
Speaker 2: Thank you, Rosa.
Speaker 1: What is particular about DeepSeek, the Chinese artificial intelligence that has shaken the markets? We have the pleasure of talking about this issue with the IT engineer Matías Carlson, who accompanies us from Buenos Aires. Matías, thank you for your availability. First of all, Matías, DeepSeek, the Chinese GPT chat, shakes the markets and Silicon Valley. It is an open source application, we say, in contrast to OpenAI. How is this relevant? What advantages does it have in practice?
Speaker 3: Good morning, very happy to be here. Well, it is very relevant that it is open source. Basically, that it is open source means that anyone in the world can access the code, can modify it, can improve it. This even generates that a large scientific community around the world can collaborate with the DeepSeek code, compared to OpenAI, which is a closed code, and obviously only depends on the people they have working inside to be able to improve that code. Others who made a very big bet were Meta, for example, with its Lama model, who also made a bet for open source.
Speaker 1: And did it work? Also, how is DeepSeek working?
Speaker 3: It also worked, what happens is that DeepSeek really had a very large technological breakthrough in terms of what training is. Always the flag that is very much in this type of open source things is also the transparency that it can generate, because one can see the code and see that there is nothing hidden, nor is it sending any message anywhere, nor is there bias.
Speaker 1: Matías, it seems that DeepSeek is at the level of the best American models. Why has it needed a much smaller amount of chips than the other technological giants?
Speaker 3: Well, that's a very good question. Basically, the need led them to reinvent the way of training the models. They had a very limited access to NVIDIA boards, they did not have access to the same boards that OpenAI suddenly has, to train from China. So basically what they did was to make a much better performance in terms of training. This was based on compression of models, on selective training. They made a very modular model, unlike OpenAI, which has a model of a complete neural architecture. So every time they train, they train their entire model. This, to make a small analogy, let's imagine that every time I go to look for something, I turn on a light from a bedroom in a building. In the case of DeepSeek, in the case of OpenAI, I have to turn on the lights of the entire building, the energy consumption that this entails. In the case of DeepSeek, they only activate the modules they need at the moment they are training it.
Speaker 1: What does this revolution for OpenAI mean, which has invested billions of dollars, and for the artificial intelligence macro project announced by Trump, Stargate?
Speaker 3: Well, initially it seemed like bad news. I think in the long term it will be good news. Let's think that the Stargate project has a lot of compute that is beastly. That will allow, let's imagine that if OpenAI or different technological companies begin to adopt this modular and compression model that DeepSeek used, which as it is open source, apart they can copy it, what they will be able to do is something much bigger and much more interesting in the medium and long term. In the short term, it seemed like bad news, and in fact they fell into the actions of envy, but I am sure that in the medium and long term, they will be able to do something much more interesting. What I do want to highlight is that it seems like with open source a much larger community is generated and a very fast acceleration in terms of the technology that can be improved in the code.
Speaker 1: Matías Carlson, computer engineer, thank you very much for your time.
Speaker 3: A pleasure.
Speaker 1: Increases pressure on Brussels for tariffs of up to 35% against Chinese electric cars. Now the American manufacturer Tesla, of the magnate Elon Musk, denounced the measure before the General Tribunal of the European Union. Tesla, like other automobiles, produces vehicles in China that later exports to the European bloc. And since Brussels made the decision to impose tariffs against electric cars manufactured in the Asian giant, those of Tesla pay a 7.8% tariff. The European Commission alleges that China unfairly subsidized manufacturers, causing significant damage to the European market. And we go with a summary of other outstanding economic notes. In the United States, some entrepreneurs already record losses in production because migrant workers have not shown up for fear of deportations ordered by President Trump. This was reported by the American Business Immigration Coalition, an organization that groups more than 1,200 companies and whose members have already reported a 50% deceleration in their operations. The Mexican economy increased its trade deficit by 50% in 2024 according to a report by the National Institute of Statistics and Geography. The country's imports exceeded exports, reaching $8.2 billion, above the also negative data of 5.470 million in 2023. This is already the fourth consecutive year of deficit for Mexico. And for the first time in history, in 2024, chicken consumption exceeded meat in Argentina, a country traditionally known for its roast. This change in eating habits is related to inflation, economic adjustment policies and the impact on the purchasing power of the population.
Speaker 4: Surrounded by vast meadows with huge herds of cattle, Argentina is still known for its roasts and juicy cuts of meat. However, 45 million Argentines have consumed more chicken than beef during the last year, for the first time since statistical records were made.
Speaker 5: The reality is that you consume more chicken because the meat is much more expensive, it yields much less, you are buying a chicken, you can take out the breast, with the female breast, with the rest you can make a sauce, then if you think about it, you have two meals.
Speaker 4: This change is reflected both in restaurants, supermarkets and butchers. The constant inflation in the country, added to the severe cuts implemented by the government, have pushed more than half of the population into poverty. Today, a kilo of minced meat costs 6,000 pesos, 5,000 pesos,
Speaker 6: which is the lowest price for meat, versus a kilo of thigh meat, 2,000 pesos, or a kilo of chicken, 2,500 pesos. So, there is the difference, and today people are taking great care of their wallets. The salaries continue to rise, the inflation continues, and people ask for offers, and we end up offering the chicken, which is the lowest price. This is contributed by Argentine ranchers prioritizing meat export,
Speaker 4: without an interest in reducing prices for the local market. Before, when export was partially closed, the producer of cricket said,
Speaker 7: why am I going to export meat, I'm going to sell it to the market, the producer of cricket said, why am I going to produce more, if I produce more, there will be more meat, and the price will go down in the local market, because I won't be able to export what is left. At this moment, the reasoning is the opposite. Well, now I have to bet on growing, because as I produce more, more will be exported, and if more is exported, I will have more possibilities of increasing the price of meat.
Speaker 4: Thus, beef, the emblem of Argentine gastronomy, will continue to be an unattainable luxury item for many average workers in the country.
Speaker 1: And this is all in La Economía. Thank you for joining us again.
Generate a brief summary highlighting the main points of the transcript.
GenerateGenerate a concise and relevant title for the transcript based on the main themes and content discussed.
GenerateIdentify and highlight the key words or phrases most relevant to the content of the transcript.
GenerateAnalyze the emotional tone of the transcript to determine whether the sentiment is positive, negative, or neutral.
GenerateCreate interactive quizzes based on the content of the transcript to test comprehension or engage users.
GenerateWe’re Ready to Help
Call or Book a Meeting Now