Connect with us

Published

on

Dado Ruvic | Reuters

Chinese artificial intelligence firm DeepSeek rocked markets this week with claims its new AI model outperforms OpenAI’s and cost a fraction of the price to build.

The assertions — specifically that DeepSeek’s large language model cost just $5.6 million to train — have sparked concerns over the eyewatering sums that tech giants are currently spending on computing infrastructure required to train and run advanced AI workloads.

But not everyone is convinced by DeepSeek’s claims.

CNBC asked industry experts for their views on DeepSeek, and how it actually compares to OpenAI, creator of viral chatbot ChatGPT which sparked the AI revolution.

What is DeepSeek?

Last week, DeepSeek released R1, its new reasoning model that rivals OpenAI’s o1. A reasoning model is a large language model that breaks prompts down into smaller pieces and considers multiple approaches before generating a response. It is designed to process complex problems in a similar way to humans.

DeepSeek was founded in 2023 by Liang Wenfeng, co-founder of AI-focused quantitative hedge fund High-Flyer, to focus on large language models and reaching artificial general intelligence, or AGI.

AGI as a concept loosely refers to the idea of an AI that equals or surpasses human intellect on a wide range of tasks.

Much of the technology behind R1 isn’t new. What is notable, however, is that DeepSeek is the first to deploy it in a high-performing AI model with — according to the company — considerable reductions in power requirements.

“The takeaway is that there are many possibilities to develop this industry. The high-end chip/capital intensive way is one technological approach,” said Xiaomeng Lu, director of Eurasia Group’s geo-technology practice.

“But DeepSeek proves we are still in the nascent stage of AI development and the path established by OpenAI may not be the only route to highly capable AI.” 

How is it different from OpenAI?

Read more DeepSeek coverage

In a technical report, the company said its V3 model had a training cost of only $5.6 million — a fraction of the billions of dollars that notable Western AI labs such as OpenAI and Anthropic have spent to train and run their foundational AI models. It isn’t yet clear how much DeepSeek costs to run, however.

If the training costs are accurate, though, it means the model was developed at a fraction of the cost of rival models by OpenAI, Anthropic, Google and others.

Daniel Newman, CEO of tech insight firm The Futurum Group, said these developments suggest “a massive breakthrough,” although he shed some doubt on the exact figures.

“I believe the breakthroughs of DeepSeek indicate a meaningful inflection for scaling laws and are a real necessity,” he said. “Having said that, there are still a lot of questions and uncertainties around the full picture of costs as it pertains to the development of DeepSeek.”

Meanwhile, Paul Triolio, senior VP for China and technology policy lead at advisory firm DGA Group, noted it was difficult to draw a direct comparison between DeepSeek’s model cost and that of major U.S. developers.

“The 5.6 million figure for DeepSeek V3 was just for one training run, and the company stressed that this did not represent the overall cost of R&D to develop the model,” he said. “The overall cost then was likely significantly higher, but still lower than the amount spent by major US AI companies.” 

DeepSeek wasn’t immediately available for comment when contacted by CNBC.

Comparing DeepSeek, OpenAI on price

DeepSeek and OpenAI both disclose pricing for their models’ computations on their websites.

DeepSeek says R1 costs 55 cents per 1 million tokens of inputs — “tokens” referring to each individual unit of text processed by the model — and $2.19 per 1 million tokens of output.

In comparison, OpenAI’s pricing page for o1 shows the firm charges $15 per 1 million input tokens and $60 per 1 million output tokens. For GPT-4o mini, OpenAI’s smaller, low-cost language model, the firm charges 15 cents per 1 million input tokens.

Skepticism over chips

LinkedIn co-founder Reid Hoffman: DeepSeek AI proves this is now a 'game-on competition' with China

Nvidia has since come out and said that the GPUs that DeepSeek used were fully export-compliant.

The real deal or not?

Industry experts seem to broadly agree that what DeepSeek has achieved is impressive, although some have urged skepticism over some of the Chinese company’s claims.

“DeepSeek is legitimately impressive, but the level of hysteria is an indictment of so many,” U.S. entrepreneur Palmer Luckey, who founded Oculus and Anduril wrote on X.

“The $5M number is bogus. It is pushed by a Chinese hedge fund to slow investment in American AI startups, service their own shorts against American titans like Nvidia, and hide sanction evasion.”

Seena Rejal, chief commercial officer of NetMind, a London-headquartered startup that offers access to DeepSeek’s AI models via a distributed GPU network, said he saw no reason not to believe DeepSeek.

“Even if it’s off by a certain factor, it still is coming in as greatly efficient,” Rejal told CNBC in a phone interview earlier this week. “The logic of what they’ve explained is very sensible.”

However, some have claimed DeepSeek’s technology might not have been built from scratch.

“DeepSeek makes the same mistakes O1 makes, a strong indication the technology was ripped off,” billionaire investor Vinod Khosla said on X, without giving more details.

It’s a claim that OpenAI itself has alluded to, telling CNBC in a statement Wednesday that it is reviewing reports DeepSeek may have “inappropriately” used output data from its models to develop their AI model, a method referred to as “distillation.”

“We take aggressive, proactive countermeasures to protect our technology and will continue working closely with the U.S. government to protect the most capable models being built here,” an OpenAI spokesperson told CNBC.

Commoditization of AI

However the scrutiny surrounding DeepSeek shakes out, AI scientists broadly agree it marks a positive step for the industry.

Yann LeCun, chief AI scientist at Meta, said that DeepSeek’s success represented a victory for open-source AI models, not necessarily a win for China over the U.S. Meta is behind a popular open-source AI model called Llama.

“To people who see the performance of DeepSeek and think: ‘China is surpassing the US in AI.’ You are reading this wrong. The correct reading is: ‘Open source models are surpassing proprietary ones’,” he said in a post on LinkedIn.

“DeepSeek has profited from open research and open source (e.g. PyTorch and Llama from Meta). They came up with new ideas and built them on top of other people’s work. Because their work is published and open source, everyone can profit from it. That is the power of open research and open source.”

WATCH: Why DeepSeek is putting America’s AI lead in jeopardy

Why China's DeepSeek is putting America's AI lead in jeopardy

– CNBC’s Katrina Bishop and Hayden Field contributed to this report

Continue Reading

Technology

Here’s what Elon Musk said about tariffs and their potential effect on Tesla

Published

on

By

Here's what Elon Musk said about tariffs and their potential effect on Tesla

U.S. President Donald Trump talks to the media, next to Tesla CEO Elon Musk with his son X Æ A-12, at the White House in Washington, D.C., U.S., March 11, 2025. 

Kevin Lamarque | Reuters

Elon Musk said on Tuesday that he doesn’t like high or unpredictable tariffs, but any decision on what happens with them “is entirely up to the president of the United States.”

Speaking on his company’s first-quarter earnings call, with tariff-related uncertainty swirling across the economy, Musk said Tesla is in a relatively good position, compared to other U.S. automakers, because it has “localized supply chains” in North America, Europe and China.

Musk said Tesla is the “least-affected car company with respect to tariffs at least in most respects.”

Tesla reported troubling quarterly earnings and sales on Tuesday, including a 20% year-over-year drop in automotive revenue and a 71% plunge in net income. The company also said that it wasn’t providing any guidance for 2025 at least until its second-quarter update.

While Musk is one of President Donald Trump’s closest advisers, tariffs are the one issue where he’s partially broken with the administration. He recently called Peter Navarro, Trump’s top trade adviser, a “moron” and “dumber than a sack of bricks.”

On Tuesday’s call, however, Musk said, “If some country is doing something predatory with tariffs,” or “if a government is providing extreme financial support for a particular industry, then you have to do something to counteract that.”

Tesla’s stock price has been hammered since the president floated his plan for widespread tariffs earlier this month, and that was after the shares plunged 36% in the first quarter, their worst performance for any period since 2022.

Because Tesla manufactures cars that it sells in the U.S. domestically, the company isn’t subject to Trump’s 25% tariff on imported cars. But Tesla counts on materials and supplies from China, Mexico, Canada and elsewhere for manufacturing equipment, automotive glass, printed circuit boards, battery cells and other products.

Musk said he offers his advice to the president on tariffs.

“He will listen to my advice. But then it’s up to him, of course, to make his decision,” Musk said. “I’ve been on the record many times saying that I believe lower tariffs are generally a good idea.”

He added that he’s an advocate for “predictable tariff structures,” as well as “free trade and lower tariffs.”

Musk said Tesla’s energy business faces an “outsized” impact from tariffs because it sources lithium iron phosphate battery cells, used in his company’s cars, from China.

“We’re in the process of commissioning equipment for the local manufacturing of LFP battery cells in the U.S.,” he said. But he said the company can “only serve a fraction of our total installed capacity” with its local equipment.

“We’ve also been working on securing additional supply chain from non-china based suppliers, but it will take time,” he said.

Musk called Tesla the most “vertically integrated car company” but said that there are still plenty of parts and materials that come from other countries. Even though it’s built a lithium refinery in Texas, “we’re not growing rubber trees and mining iron yet,” he said.

WATCH: Tariffs on batteries out of China can end up being really costly

Tariffs on batteries out of China can end up being really costly for tariffs, says Fmr. Tesla President

Continue Reading

Technology

Tesla CEO Musk says time he spends on DOGE will drop ‘significantly’ next month

Published

on

By

Tesla CEO Musk says time he spends on DOGE will drop 'significantly' next month

Elon Musk, CEO of Tesla Inc., in the Oval Office of the White House in Washington, D.C., on Feb. 11, 2025.

Aaron Schwartz | Bloomberg | Getty Images

Tesla CEO Elon Musk began his company’s earnings call on Tuesday by saying that his time spent running President Donald Trump’s Department of Government Efficiency will drop “significantly” starting in May.

Musk, who has watched Tesla’s stock tumble by more than 40% this year, said he’ll continue to support the president with DOGE “to make sure that the waste and fraud that we stop does not come roaring back.”

After spending almost $300 million in the 2024 campaign to help return Trump to the White House, Musk created DOGE and joined the administration with a mission to drastically reduce the size and capability of the federal government.

He said he’ll continue to spend a “day or two per week” on government issues “for as long as the president would like me to do so.”

Musk’s commentary came after his company reported disappointing first-quarter results, including a 20% year-over-year slump in automotive revenue and 71% plunge in net income.

In addition to challenges the company already faced, such as competition out of China and an aging fleet of electric vehicles, Tesla has recently been hit with protests in the U.S. and Europe and brand damage due to Musk’s ties to Trump and his support of Germany’s far-right AfD party.

“The protests that you’ll see out there, they’re very organized,” Musk said on Tuesday’s call. He claimed, without evidence, that some people are likely protesting “because they’re receiving fraudulent money” or are “recipients of wasteful largesse.”

On its website, which was last updated on Sunday, DOGE says its cuts have led to an estimated $160 billion in savings. However, Musk’s estimates of savings have been challenged, and DOGE has deleted some of the largest purported savings.

Over that same stretch, Tesla has lost roughly $600 billion in market cap.

DOGE has also made cuts at agencies charged with oversight of his companies. They include the SEC, Federal Aviation Administration and National Highway Traffic Safety Administration.

The White House said in early February that Musk was serving as a “special government employee,” a designation with fewer requirements when it comes to conflict-of-interest disclosures and ethics policies.

The Department of Justice says the title is for anyone expected to work for the government for 130 days or less in a year. The Trump administration will hit its 130th day at the end of May.

Job cuts from DOGE’s work have come from across the government, at agencies including the Internal Revenue Service, National Park Service, Consumer Financial Protection Bureau, and the departments of Agriculture, Education, Energy, Health and Human Services, Homeland Security, and Veterans Affairs, according to the Associated Press.

As of February, staffers from DOGE had pushed top-ranking officials at the Department of Education out of their offices, rearranged the furniture and set up white noise machines to muffle their voices, according to employees at the agency. U.S. senators expressed concern that DOGE had possibly gained access to federal student loan data on tens of millions of borrowers.

Also in February, the Trump administration said that USAID would shut down as an independent agency and be moved under the State Department.

WATCH: Musk needs to recommit to Tesla

Elon Musk needs to recommit to Tesla and say he's leaving Trump administration: Wedbush's Dan Ives

Continue Reading

Technology

Meta could take a $7 billion hit this year because of Trump’s tough China tariffs

Published

on

By

Meta could take a  billion hit this year because of Trump's tough China tariffs

This photo illustration created on Jan. 7, 2025, in Washington, D.C., shows an image of Mark Zuckerberg, CEO of Meta, and an image of the Meta logo.

Drew Angerer | AFP | Getty Images


Meta’s core online advertising business could take a $7 billion hit this year due to President Donald Trump’s tough China tariffs impacting retailers in the country.

That’s according to a MoffettNathanson research note published Tuesday that analyzes the potential impact of China-linked retailers like Temu and Shien slashing their Facebook and Instagram advertising budgets amid the U.S. and China trade dispute.

The MoffettNathanson analysts pointed to Meta’s latest annual report in which the company revealed that its China revenue was $18.35 billion in 2024, equating to a little over 11% of total its total sales. Like other analysts, MoffettNathanson believe Temu and Shien comprise the bulk of Meta’s China business, and if those online retailers cut back on their ad campaigns this year, the social networking giant’s 2025 ad sales could be impacted by $7 billion.

Meta did not immediately respond for a request for comment.

There are already signs of a pullback, the analysts wrote, citing a CNBC report about Temu reducing its U.S. advertising spending and seeing a big drop in its Apple App Store rankings following Trump’s China tariffs.

“China’s importance to Meta’s business cannot be overstated,” the analysts wrote in the note. “While Meta does not provide a country-level breakdown of revenue within Europe, we logically can presume that China is Meta’s second-largest revenue source after the United States — a remarkable position for a country where Meta has no users or active platforms.”

Meta could be in even more trouble if the broader markets heads into a recession this year, as some analysts and corporate financial chiefs have predicted. A “truly prolonged economic downturn” combined with the U.S. and China trade dispute “could wipe $23 billion in 2025 advertising revenues off Meta’s books and crush our 2025 earnings by -25%,” the analysts said.

“As noted earlier, we believe Meta is particularly exposed to a pullback in ad spend from Chinese advertisers,” the analysts said. “In a scenario where a recession is triggered or exacerbated by escalating trade tensions, Meta would face a dual headwind: cyclical advertising weakness and a targeted decline in Chinese ad spend.”

The MoffettNathanson analysts still maintain a Buy rating on Meta, said they have but decreased their target price by $185 to $525.

Meta shares have dropped about 19% to $499.36 since Trump was officially sworn in as U.S. president for the second time.

The company reports its first-quarter earnings next Wednesday.

WATCH: Fmr. DOJ antitrust chief: Antitrust enforcement is most important.

Fmr. DOJ antitrust chief: Antitrust enforcement is most important in times of tech inflection points

Continue Reading

Trending