Investors have closely watched Nvidia’s week-long GPU Technology Conference (GTC) for news and updates from the dominant maker of chips that power artificial intelligence applications.

The event comes at a pivotal time for Nvidia shares. After two years of monster gains, the stock is down 15% over the past month and 22% below the January all-time high.

As part of the event, CEO Jensen Huang took questions from analysts on topics ranging from demand for its advanced Blackwell chips to the impact of Trump administration tariffs. Here’s a breakdown of how Huang responded — and what analysts homed in on — during some of the most important questions:

Huang said he “underrepresented” demand in a slide that showed 3.6 million in estimated Blackwell shipments to the top four cloud service providers this year. While Huang acknowledged speculation regarding shrinking demand, he said the amount of computation needed for AI has “exploded” and that the four biggest cloud service clients remain “fully invested.”

Morgan Stanley analyst Joseph Moore noted that Huang’s commentary on Blackwell demand in data centers was the first-ever such disclosure.

“It was clear that the reason the company made the decision to give that data was to refocus the narrative on the strength of the demand profile, as they continue to field questions related to Open AI related spending shifting from 1 of the 4 to another of the 4, or the pressure of ASICs, which come from these 4 customers,” Moore wrote to clients, referring to application-specific integrated circuits.

Piper Sandler analyst Harsh Kumar said the slide was “only scratching the surface” on demand. Beyond the four largest customers, he said others are also likely “all in line looking to get their hands on as much compute as their budgets allow.”

Another takeaway for Moore was the growth in physical AI, which refers to the use of the technology to power machines’ actions in the real world as opposed to within software.

At previous GTCs, Moore said physical AI “felt a little bit like speculative fiction.” But this year, “we are now hearing developers wrestling with tangible problems in the physical realm.”

Truist analyst William Stein, meanwhile, described physical AI as something that’s “starting to materialize.” The next wave for physical AI centers around robotics, he said, and presents a potential $50 trillion market for Nvidia.

Stein highliughted Jensen’s demonstration of Isaac GR00T N1, a customizable foundation model for humanoid robots.

Several analysts highlighted Huang’s explanation of what tariffs mean for Nvidia’s business.

“Management noted they have been preparing for such scenarios and are beginning to manufacture more onshore,” D.A. Davidson analyst Gil Luria said. “It was mentioned that Nvidia is already utilizing [Taiwan Semiconductor’s’] Arizona fab where it is manufacturing production silicon.”

Bernstein analyst Stacy Rasgon said Huang’s answer made it seem like Nvidia’s push to relocate some manufacturing to the U.S. would limit the effect of higher tariffs.

Rasgon also noted that Huang brushed off concerns of a recession hurting customer spending. Huang argued that companies would first cut spending in the areas of their business that aren’t growing, Rasgon said.

Nvidia CEO Jensen Huang on Thursday walked back comments he made in January, when he cast doubt on whether useful quantum computers would hit the market in the next 15 years.

At Nvidia’s “Quantum Day” event, part of the company’s annual GTC Conference, Huang admitted that his comments came out wrong.

“This is the first event in history where a company CEO invites all of the guests to explain why he was wrong,” Huang said.

In January, Huang sent quantum computing stocks reeling when he said 15 years was “on the early side” in considering how long it would be before the technology would be useful. He said at the time that 20 years was a timeframe that “a whole bunch of us would believe.”

In his opening comments on Thursday, Huang drew comparisons between pre-revenue quantum companies and Nvidia’s early days. He said it took over 20 years for Nvidia to build out its software and hardware business.

He also expressed surprise that his comments were able to move markets, and joked he didn’t know that certain quantum computing companies were publicly traded.

“How could a quantum computer company be public?” Huang said.

The event included panels with representatives from 12 quantum companies and startups. It represents a truce of sorts between Nvidia, which makes more traditional computers, and the quantum computing industry. Several quantum execs fired back at Nvidia after Huang’s earlier comments.

A third panel included representatives from Microsoft and Amazon Web Services, which are also investing in quantum technology and are among Nvidia’s most important customers.

Nvidia has another reason to embrace quantum. As quantum computers are being built, much of the research on them is done through simulators on powerful computers, like those that Nvidia sells.

It’s also possible that a quantum computer would require a traditional computer to operate it. Nvidia is working to provide the technology and software to integrate graphics processing units (GPUs) and quantum chips.

“Of course, quantum computing has the potential and all of our hopes that it will deliver extraordinary impact,” Huang said on Thursday. “But the technology is insanely complicated.”

Nvidia said this week that it will build a research center in Boston to allow quantum companies to collaborate with researchers at Harvard and the Massachusetts Institute of Technology. The center will include several racks of the company’s Blackwell AI servers.

Quantum computing has been a dream of physicists and mathematicians since the 1980s, when California Institute of Technology professor Richard Feynman first proposed the idea behind a quantum computer.

While classical computers use bits that are either 0 or 1, the bits inside a quantum computer — qubits — end up being on or off based on probability. Experts predict that the technology will be able to solve problems with massive amounts of possible solutions, such as deciphering codes, routing deliveries or simulating chemistry or weather.

No quantum computer has yet beat a computer at solving a real, useful problem. But Google claimed late last year that it discovered a way to do error correction.

One question at the panel centered around whether quantum computing might one day threaten companies like Nvidia that make computers based on transistors.

“A long time ago, somebody asked me, ‘So what’s accelerated computing good for?’” Huang said at the panel. Accelerated computing is a phrase he uses to refer to the kind of GPU computers that Nvidia makes.

“I said, a long time ago, because I was wrong, this is going to replace computers,” he said. “This is going to be the way computing is done, and and everything, everything is going to be better. And it turned out I was wrong.”

Klarna, the buy now, pay later lender that’s headed for an initial public offering, said on Thursday that it’s signed on DoorDash as a partner, another sign of momentum for public market investors.

It’s DoorDash’s first BNPL alliance and gives users of the restaurant delivery service a new way to pay for meals. Klarna said in a press release that DoorDash customers will be able to pay in full at checkout, split payments into four equal interest-free installments, or defer to dates that align conveniently with payday schedules.

Klarna, which is headquartered in Sweden, filed its prospectus last week to list on the New York Stock Exchange. Revenue last year increased 24% to $2.8 billion, and adjusted operating profit was $181 million, swinging from a loss of $49 million a year earlier. CNBC reported on Monday that Klarna will be the exclusive provider of buy now, pay later loans for Walmart, taking a coveted partnership away from rival Affirm.

“Our partnership with DoorDash marks an important milestone in Klarna’s expansion into everyday spending categories,” said David Sykes, Klarna’s chief commercial officer, in Thursday’s release.

Klarna, founded in 2005, said in its prospectus that it has 675,000 merchant partners in 26 countries. It’s among the most hotly anticipated IPOs of the year following an extended stretch of historically little activity for new offerings.

Investors have closely watched Nvidia’s week-long GPU Technology Conference (GTC) for news and updates from the dominant maker of chips that power artificial intelligence applications.

The event comes at a pivotal time for Nvidia shares. After two years of monster gains, the stock is down 15% over the past month and 22% below the January all-time high.

As part of the event, CEO Jensen Huang took questions from analysts on topics ranging from demand for its advanced Blackwell chips to the impact of Trump administration tariffs. Here’s a breakdown of how Huang responded — and what analysts homed in on — during some of the most important questions:

Huang said he “underrepresented” demand in a slide that showed 3.6 million in estimated Blackwell shipments to the top four cloud service providers this year. While Huang acknowledged speculation regarding shrinking demand, he said the amount of computation needed for AI has “exploded” and that the four biggest cloud service clients remain “fully invested.”

Morgan Stanley analyst Joseph Moore noted that Huang’s commentary on Blackwell demand in data centers was the first-ever such disclosure.

“It was clear that the reason the company made the decision to give that data was to refocus the narrative on the strength of the demand profile, as they continue to field questions related to Open AI related spending shifting from 1 of the 4 to another of the 4, or the pressure of ASICs, which come from these 4 customers,” Moore wrote to clients, referring to application-specific integrated circuits.

Piper Sandler analyst Harsh Kumar said the slide was “only scratching the surface” on demand. Beyond the four largest customers, he said others are also likely “all in line looking to get their hands on as much compute as their budgets allow.”

Another takeaway for Moore was the growth in physical AI, which refers to the use of the technology to power machines’ actions in the real world as opposed to within software.

At previous GTCs, Moore said physical AI “felt a little bit like speculative fiction.” But this year, “we are now hearing developers wrestling with tangible problems in the physical realm.”

Truist analyst William Stein, meanwhile, described physical AI as something that’s “starting to materialize.” The next wave for physical AI centers around robotics, he said, and presents a potential $50 trillion market for Nvidia.

Stein highliughted Jensen’s demonstration of Isaac GR00T N1, a customizable foundation model for humanoid robots.

Several analysts highlighted Huang’s explanation of what tariffs mean for Nvidia’s business.

“Management noted they have been preparing for such scenarios and are beginning to manufacture more onshore,” D.A. Davidson analyst Gil Luria said. “It was mentioned that Nvidia is already utilizing [Taiwan Semiconductor’s’] Arizona fab where it is manufacturing production silicon.”

Bernstein analyst Stacy Rasgon said Huang’s answer made it seem like Nvidia’s push to relocate some manufacturing to the U.S. would limit the effect of higher tariffs.

Rasgon also noted that Huang brushed off concerns of a recession hurting customer spending. Huang argued that companies would first cut spending in the areas of their business that aren’t growing, Rasgon said.

The Metropolitan Transit Authority will stop selling and refilling those formerly-ubiquitous MetroCards by the end of the year in favor of the OMNY system, MTA Chair and CEO Janno Lieber told Crain’s New York Business Wednesday.

MetroCards have been around since 1994, but now seem destined to go the way of the subway token, which stopped being used in 2003.

MTA officials previously said they planned to say goodbye to MetroCards in 2027, but now have provided an estimated date when they will stop selling and filling the cards, and that’s at the end of 2025.

OMNY’s popular tap-and-go system has been around since 2019 and the service includes the ability to tap your phone to pay to purchase an OMNY tap card that passengers can buy and reload.

Commuters will still be able to use their existing MetroCards with whatever funds they have on them until sometime in 2027.

Nvidia CEO Jensen Huang downplayed the negative impact from President Donald Trump’s tariffs, saying there won’t be any significant damage in the short run.

“We’ve got a lot of AI to build … AI is the foundation, the operating system of every industry going forward. … We are enthusiastic about building in America,” Huang said Wednesday in a CNBC “Squawk on the Street” interview. “Partners are working with us to bring manufacturing here. In the near term, the impact of tariffs won’t be meaningful.”

Trump has launched a new trade war by imposing tariffs against Washington’s three biggest trading partners, drawing immediate responses from Mexico, Canada and China. Recently, Trump said he would not change his mind about enacting sweeping “reciprocal tariffs” on other countries that put up trade barriers to U.S. goods. The White House said those tariffs are set to take effect April 2.

“We’re as enthusiastic about building in America as anybody,” Huang said. “We’ve been working with TSMC to get them ready for manufacturing chips here in the United States. We also have great partners like Foxconn and Wistron, who are working with us to bring manufacturing onshore, so long-term manufacturing onshore is going to be something very, very possible to do, and we’ll do it.”

Shares of Nvidia have fallen more than 20% from their record high reached in January. The stock suffered a massive sell-off earlier this year due to concerns sparked by Chinese artificial intelligence lab DeepSeek that companies could potentially get greater performance in AI on far-lower infrastructure costs. Huang has pushed back on that theory, saying DeepSeek popularized reasoning models that will need more chips.

Nvidia, which designs and manufactures graphics processing units that are essential to the AI boom, has been restricted from doing business in China due to export controls that were increased at the end of the Biden administration.

Huang previously said the company’s percentage of revenue in China has fallen by about half due to the export restrictions, adding that there are other competitive pressures in the country, including from Huawei.