What a major tech executive said about Deepseek as we were contemplating the ban.
The wave of major tech companies has reported quarterly revenue in recent weeks. The 7 magnificent companies – right after that Chinese AI startup Deepseek rattles the stock market as investors saw the new, inexpensive artificial intelligence model as an ominous sign of US tech stocks.
Deepseek’s innovation has shaken up the US tech industry. nvidia (NVDA)) $600 billion cuts market capitalization in one day – The biggest loss in stock market history – Investors worry about spending on AI hardware will be slower as they worry about Big Tech.
On revenue calls with investors, executives from these companies quickly praised Deepshek’s artificial intelligence models, either dismissing them or trying to get around the topic entirely.
The response to the sale of US high-tech executives – which affected most of the stocks – ranged from defensive to excitement. Deepseek’s news has largely agreed that AI costs are indications that they will eventually collapse, but reaffirms its commitment to spending large amounts of money on capital expenditures and other investments in AI infrastructure in 2025 I did. come.
This is what they said.
Microsoft (msft) CEO Satya Nadella quickly accepted Deepseek and mentioned the company in her opening remarks on the January 29th post-revenue call.
Nadella pointed to the move to deploy Deepseek’s latest AI models on developer platforms Azure Ai Foundry and Github, adding that it has passed “automatic red team, content safety integration, security scans.” He said that customers can run Deepseek’s model locally on Microsoft’s AI PCS.
“I think there were some real innovations in Deepseek,” Nadella said, adding that he sees AI as “commoditized.”
“For hyperschools like us, which are PC platform providers like us, this is all good news as far as I’m concerned.”
Microsoft Chairman and CEO Satya Nadella will speak at a Microsoft briefing event held at Microsoft in Redmond, Washington on May 20, 2024. ・Jason Redmond via Getty Images
Meta (Meta) – Recently, one high-tech analyst has been saying,The most appropriate companies that utilize generative AI“Given the advertising business, stocks rose with the debut of Deepseek, a new AI model called R1, and stocks rose nearly 2% on the day of the news.
CEO Mark Zuckerberg has not changed relatively about Deepseek Frenzy.
When asked whether the potential for a low-cost AI model would affect Meta’s capital expenditures, the CEO said over the phone: The company’s latest quarterly revenue: “I don’t know – it’s probably too early to have a really strong opinion on the trajectory of infrastructure and equipment and what that means.”
Meta CEO Mark Zuckerberg was testing out Orion AR glasses at Meta Connect’s annual event held on September 25, 2024 at its headquarters in Menlo Park, California, USA. Reuters/Manuel Orbegozo ・Reuters/Reuters
“I’m also thinking in light of recent news, some of the new competitors, Deepseek from China… that’s one of the things we’re talking about.
“And I think it’s important that it is the American standard because of our kind of national advantage.”
Apple (whose “secular perspective” on the “Deepseek situation”)aapl) CEO Tim Cook said in a call after the revenues on January 30th that “innovations that drive efficiency are good,” and the iPhone maker said, “a very wise and intentional approach to our spending.” He said he was taking the following.
When asked about what DeepSeek’s low-cost model means for Google (Goog), CEO Sundar Pichai said Deepseek’s team is “doing a very good job,” but thinks Google’s Gemini Flash model is better.
“It’s some of the most efficient models, such as comparing both the 2.0 flash model and the 2.0 flash thinking model, as well as the DeepSeek V3 and R1.”
“And I think a lot of that is the strength of full stack development (Google creates its own AI chips and AI models and software that runs them), end-to-end optimization, cost obsessive queries. Each one added Pichai.
“Part of the reason I’m so excited about AI opportunities is that I can drive extraordinary use cases to make it feasible because it actually reduces the cost of using it. I know.”
Amazon (amzn) CEO Andy Jassy said he believes Deepseek’s model will not cause a recession in AI investments.
“If people can reduce the costs of any type of technology component, then in this case we’re actually talking about reasoning (executing AI models) in technology.
Jassy pointed to the company’s aggressive spending on cloud infrastructure development in the early 2000s, even if costs were reduced.
Amazon CEO Andy Jassy will speak at the keynote address at AWS Re: Invent 2024, a conference hosted by Amazon Web Services, held at Venetian Las Vegas in Las Vegas, Nevada, on December 3, 2024. (Photo of Noah Berger/Getty Images on Amazon Web Services) ・Noah Burger via Getty Images
He’s right, at least for now.
Despite Meta, Alphabet, Amazon and Microsoft expect to reduce AI training and inference costs with revenue calls, they Accumulated $325 billion in 2025an increase of 46% from the previous year. Amazon is the group’s largest spender.
Investors were not pleased: Amazon’s stock fell 4% on Friday Following the executive commentary, it expects capital expenditures to increase by 35% to exceed $100 billion.
AMD (AMDCEO Lisa Su expressed her belief that new innovations such as the Deepseek model would not cut down on AI investments, supporting SoftBank for the recently announced $500 billion Stargate AI infrastructure project. I pointed it out (sftby), Oracle (orcl), and openai.
“All of these initiatives require a huge amount of new computing and create unprecedented growth opportunities for AMD across our business.”
AMD CEO Lisa Su will speak on stage as he closes his dinner at the Chelsea Piers on December 11, 2024 at the current dinner at the Chelsea Piers. (Photo: Noam Galai/Getty Images for Time) ・Noam Galai Vitty Images
Perhaps he was the most enthusiastic about Deepseek’s innovation (Qualcomm)QCOM) CEO Cristiano Amon claimed that such developments could speed up the timeline of AI and drive a new wave of demand for smartphones and PCs.
“DeepSeek-R1 and other similar models have recently demonstrated that AI models have developed faster, become smaller, more capable, more efficient and can be run directly on devices,” Amon said. Ta.
“As we enter the age of AI inference, training continues in the cloud, while inference is increasingly running on devices, making AI more accessible, customizable and efficient. Models and applications are adopted It will drive the rise of the Qualcomm platform across a variety of devices.
Qualcomm CEO and President Cristiano Amon will be attending the first day of the web summit held in Lisbon, Portugal on November 12, 2024 (photo by Rita Franca/Nurphoto via Getty Images) ・Nurphoto via Getty Images
arm(arm) CEO Rene Haas also claimed that UK-based chip architecture designers will benefit from low-cost AI for consumer devices when they take JAB on Nvidia.
“A great product like (Nvidia’s) Grace Blackwell will not be able to put it in a phone, you can’t put it in your earphones. You can’t even put it in your car. But the arms will be in all those places there is.”
Laura Bratton is a reporter for Yahoo Finance. Follow her at Bluesky @laurabratton.bsky.social. Email her to laura.bratton@yahooinc.com.
