The wave of large technology companies reported quarterly earnings in recent weeks, including six The so-called magnificent 7 firms after a short time Chinese Ai Startup Deepseek shook the stock exchange, as investors saw his new, cheap artificial intelligence models as a terrible sign for American technology shares.
Deepseek’s innovations are so shocked by the US technology industry that AI DARLING Nvidia BootyNvdaIn the case of 600 billion dollars shaved his market hat in a day – There is the largest loss in the history of the stock market as investors who are worried about large technologies, they will slow down its costs on AI.
About the calls with investors, the heads of these firms quickly praised the deep deep artificial intelligence models, rejected them, or tried to avoid the topic at all.
The reactions of US technology managers for sale, which affected most of their shares fluctuated from protection. While a very agreed Deepseek News is a sign that AI’s expenses reaffirmed their commitments that have larger investments in capital expenditures and other AI infrastructure, despite the costs of those costs.
Here’s what they say.
Microsoft (:Hook) Home Satya Nadella quickly received Deepseek, on January 29, defined by its opening speech on its opening call.
Nadella pointed to Microsoft’s move on the last AI model of Deepseek’s AI, AI foundry and Github, adding that it was “automated red team and security scanning”. He said customers would soon be able to deploy Deepseek models on Microsoft’s AI PCs.
“I think Deepseek had real innovations,” Nadela said, adding that he sees that AI is “commodent”.
“For such a hyperscalik like us, like a computer platform like us, all this is good news, as far as I am concerned.”
Microsoft President and Chief Executive Officer Satya Nadella During the May 20, Microsoft Mystery, 2024 May 20. (Photo by Jason Redmond / AFP) ·Jason resets through Getty Images
Meta (:Meta) – which has been recently described as “The most well-installed company that benefits from generating AI“Given its advertising business. He has shipped his stock in his new AI model Deepseek’s new AI model, which is called R1, on the occasion of the news day, the shares are growing by almost 2%.
Chief Executive Officer Mark Uck Werberg was relatively non-sinking about Deepseek Frenzy.
To the question whether the potential of lower costs AI models will affect the capital expenditures of the meta, said the chief executive The company’s last quarterly earnings.
Meta Ceo Mark Zuckerberg is trying to the Orion Ar Glasses at the annual Meta Connect event, California Menlo Park, 2024. On September 25. Reuters / Manuel Orbegozo ·Reuters / Reuters
To the question of what is the low cost of Deepseek for Google (GardenLawyer Sundar Pichay said that the Deepseek team “has done a very good job”, said Google’s Gememini Flash models are better.
“I would say and our 2.0 flash models, our 2.0 flash-thinking models, they are the most effective models there, including comparing Deepseek v3 and R1.”
“And I think that our strength of our full development is a lot [Google makes its own custom AI chips as well as AI models and the software that runs on them]Final optimization, our obsession according to the value of the survey, “Pichas added.
“I think we are so excited about the opportunity AI, we know that we can run emergency cases because it will actually become more use.”
Amazon (:Amzn) Executive Director Andy Manat said that it believes that Deepseek models will not decline in AI investments.
“Sometimes people make assumptions that if you are able to reduce the value of any type of technological component, in this case we really talk about the conclusion. [running AI models]In some way it will bring less costs in technology. And we have never seen it is. “
Anatian Atingians, the aggression pointed out the company’s aggressive costs to develop its cloud infrastructure in the early 2000s, even as expenses.
Amazon’s Executive Director Andy Athy Speaks at AWS RE Based Street: 2024 – a conference held by Amazon Web Services, in the Venetian Las Vegas, located in Las Vegas in Nevada. (Photo: Noah Berger / Getty Images for Amazon Web Services) ·NOAH BERGER GETTY IMAGES
He is right, at least for now.
Meta, Alphabet, Amazon and Microsoft call on their earnings to predict that artificial intelligence education and conclusion costs will Spend $ 325 billion in accumulative cumulative in 202546% have increased since last year. Amazon is the biggest mole of the band.
The investors were not satisfied. Amazon’s shares decreased by 4% Friday Following the comments of the leaders, they expect to stimulate capital expenditures by 35% and more than $ 100 billion.
AMD (:DashGeneral Director Lisa Sust expressed its conviction that new innovations in Deepseek models are not investing in AI, which indicate the $ 500 billion AI infrastructure project provided by SoftBank (:Sittb), Oracle (Sole), and Openai.
“All these initiatives require new calculations and create unprecedented growth opportunities in our business in AMD.”
In the organization, Lisa Sus, CEO, in 3444, Chelsea on December 11, 2024 in New York. (Photo during NOAM GALAI / GETTY IMAGES) ·NOAM GALAI VITTY PICTURES
Perhaps the most enthusiastic about Deepseek innovations, was the qualification (Qcom) CEO Corentian Ammon, which did the job that such developments could accelerate the schedule for AI to hold a new channel of the demand for smartphones and PCs.
“Deepseek-R1 and other similar models are developing faster, they become smaller, more capable and efficiently and can now launch directly on the device,” Amon said.
“When we enter the AI ​​SERVICE, we expect to continue in the cloud before the course will run more and more on the device, which will stimulate more targeted, target, targeted, targeted, targeted, targeted and effective models and programs that. We expect to continue to raise the adoption, and in turn require a number of devices for qualitative assemblies. “
In CEO and President Cristiano Ammon, the first day of the North Summit is participating in the first day of the website of Lisbon in Portugal, 2024. November 12. (Photo by RITA FRANCA / NURPHOTO) ·NURPHOTO VIA GETTY Images
ARM (:Arm) Executive Director Rene Hadj also claims that the UK Chip’s chip designer will use a lower price for consumer devices, even to take a crushing in NVIDIA.
“As a great product as [Nvidia’s] Grace Blackwell is that you will never be able to put it on a cell phone, never can put it into the ears. You can’t even put it in a car. But the hand is in all places. “
Laura Bratton is a journalist of Yahoo Finance. Follow him on Blureky @ Laurabratton.bsky.social. Send him to laura.bratton@yahooinc.com.
English