Main to remember
- Microsoft, Meta and IBM have all highlighted healthy growth in their AI companies in the last three months of 2024.
- Microsoft and Meta, who cumulatively expect to invest nearly $ 150 billion in the infrastructure during their respective financial years of 2025, held their aggressive spending plans.
- The leaders expressed the confidence that the incredibly profitable AI model of the Chinese start-up Deepseek, who frightened investors earlier this week, was finally a good thing for their business.
Artificial intelligence was the accent put on Wednesday when the Microsoft (MSFT) and Meta (META) technology launched the first Big Tech income round in 2025.
Analysts have pointed out questions of questions about their investments in AI, how many AI contributes to income and their reflections on the Chinese start-up Deepseek, which rocked Wall Street earlier this week with its super efficient AI model .
AI activity is booming
The leaders were impatient to show investors that artificial intelligence contributes to their results.
Meta CFO Susan Li said Meta AI had reached 700 million monthly active users in the fourth quarter while the ADAVer + Marketing Service Ai-alive by the Company increased by 70%. Meta’s quarterly revenues and benefits have both exceeded the expectations of analysts.
Cloud income in the second quarter of Microsoft in the second quarter has slightly fell in expectations, but CEO Satya Nadella noted that the IA income rate of the company had reached $ 13 billion in the quarter, -Dess the $ 10 billion mark he provided in October.
The CEO of IBM (IBM), Arvind Krishna, said that the company had reserved nearly $ 2 billion in AI activities in the last quarter of 2024; This represents around 40% of the $ 5 billion reserved since the creation.
Tech Titans is engaged in their IA expenses
The leaders of Microsoft and Meta were standing from their plans to spend tens of billions of dollars this year in IA infrastructure.
“We are planning to continue investing against solid request signals,” said Microsoft’s financial director Amy Hood about his capital expenses on the company’s profits. “However, the growth rate will be lower than the 2010 financial year.”
Microsoft plans to spend $ 80 billion on infrastructure during its year 2025. So far, two quarters, the company has spent $ 30.7 billion, an increase of 56% compared to a year ago .
The meta-PDG Mark Zuckerberg has also defended the massive investment in the AI of its company, which could total $ 65 billion this year. “I continue to think that the investment very strongly in Capex and Infra will be a strategic advantage over time,” he told analysts on Wednesday evening.
The expenses of the Big Tech AI were already a source of anxiety at Wall Street, where for half a year, the questions swirled to find out if the demand will correspond to the magnitude of the investments in the infrastructure. These doubts were amplified this week when the markets took note of Deepseek, whose Open Source model would occur as the American models and the features at a fraction of the cost.
Before Wednesday’s profits reports, analysts doubted that hyperscalers like Microsoft and Amazon (AMZN) would modify their investment prospects. Deepseek efficiency could feed the development of AI applications and more stoke demand, giving cloud suppliers even more reasons to continue to add computer capacity aggressively.
Silicon Valley impressed by Deepseek, not frightened
The markets may have been shaken by Deepseek, but technology leaders were not.
“In a certain sense, what is happening with AI is not di ff erent that what was happening with the regular calculation cycle,” said Nadella de Microsoft of Deepseek’s efficiency. “We ourselves observe significant efficiency gains in training and inference for years now.”
Nadella admitted that Deepseek, whose model competes with the abilities of OpenAi supported by Microsoft, “had real innovations” which, according to him, should benefit her business. “When the prices of the tokens drop, the prices for calculating inferences drop. This means that people can consume more, and there will be more written applications ”which are supported by the infrastructure and services of Microsoft.
Zuckerberg echoes the praise of Nadella for Deepseek, saying that he and his team still digest “a certain number of new things they have done” that “we hope to implement in our systems”.
Zuckerberg added that Deepseek’s success has confirmed Meta’s decision to make his large language model, Llama, Open-Source. Deepseek was proof, he said that “there will be an open source standard worldwide. And I think that for our kind of national advantage, it is important that it is an American standard. »»
In another sign of an almost imperturbable confidence in the future of American AI leaders, the Japanese technology investor Softbank would be in talks to invest up to $ 25 billion in Openai, which would do the most Large Investor of Chatgpt-Maker. It would be at the top of the Soft bank of $ 15 billion already committed to spending a joint venture with OpenAi and Oracle (ORCL), announced last week.
Update – January 30, 2025: This article has been updated with information concerning the potential investment of SoftBank in Openai.
