Cerebras Systems’ monster debut on Thursday didn’t just place it among tech’s biggest-ever IPOs — it was a crystal clear signal of unstoppable demand for chips to power AI, as tech giants scramble to find alternatives to the costly, sold-out graphics processing units made by Nvidia.

Cerebras closed its first day trading on Wall Street with a market cap just below $100 billion, putting it near the few companies to close above that mark, such as Facebook-parent Meta and Alibaba. The stock closed 10% lower on Friday, its first full day of trading.

Here’s what you need to know about this hot Nvidia competitor.

Cerebras makes a different type of chip than the classic Nvidia GPU, and it’s the size of a dinner plate.

“We build the biggest chips in the semiconductor industry,” Cerebras CEO and Co-Founder Andrew Feldman told CNBC on Squawk Box Thursday. “Big chips process more information in less time and deliver results more quickly.”

Until now, Nvidia has been winning the AI chip race because its GPUs serve as general-purpose workhorses, excelling at the parallel math necessary for training large models. But we’ve now arrived at the era of agentic AI, where inference is key. While training teaches the AI model to learn from patterns in large amounts of data, inference uses the AI to make decisions based on new information.

Inference can happen on less powerful chips programmed for more specific tasks, such as Cerebras’ WSE-3. It falls into a category of chips known as custom ASICs — application-specific integrated circuits. It’s an increasingly crowded space, with in-house ASICS now made by the likes of Google, Amazon, Meta and Microsoft.

Cerebras said the WSE-3 is 57 times larger than the largest GPU, and has 50 times the number of transistors.

The most advanced AI chips are made using Taiwan Semiconductor Manufacturing’s 2-nanometer process node, currently only possible in Taiwan. Cerebras’ chip is also made at TSMC, but on its less advanced 5-nanometer node.

Founded in Silicon Valley in 2016, Cerebras first filed to go public in 2024, but withdrew that submission when it faced scrutiny for its heavy reliance on a single customer, Microsoft-backed AI firm G42 in the United Arab Emirates.

With the firm’s successful IPO on Thursday, Feldman and hardware technology chief Sean Lie, two of its co-founders, became billionaires based on their holdings.

For years, Cerebras sought to sell chips to companies, but now largely operates the chips inside its own data centers as a cloud service, pitting it against cloud providers Google, Microsoft, Oracle and CoreWeave.

Cerebras and OpenAI announced a $20 billion cloud deal in January that expires in 2028, while Amazon Web Services announced in March that it’s using Cerebras chips in its data centers.

“For our fast inference product, there’s so much demand that our biggest challenge is actually trying to supply it. We are adding as much manufacturing and data center capacity as we possibly can, and we’re still sold out into 2027,” Cerebras CFO Bob Komin told CNBC Thursday.

While hyperscalers make their own in-house ASICs, Cerebras more closely competes with firms that specialize in making them for others. Chief among those is Groq. In its largest purchase to date, Nvidia paid $20 billion for Groq’s tech in December, then announced custom Groq Language Processing Units at GTC in March.

SambaNova and D-Matrix are two other notable Cerebras competitors looking to capitalize on unprecedented AI chip demand.

SambaNova counts Hugging Face and Meta among the customers of its SN50 chips, while Intel participated in a $350 million funding round for SambaNova in February. Intel CEO Lip-Bu Tan has served as SambaNova’s chairman since 2017.

Cerebras’ IPO also paves the way for other custom ASIC start-ups looking to go public, such as Rebellions.

The South Korean chipmaker raised $400 million from the likes of Samsung, at a valuation of $2.34 billion, in March as it prepares for an IPO.

Watch: Breaking down AI chips, from Nvidia GPUs to ASICs by Google and Amazon

At a time when Hollywood is torn between fear of artificial intelligence stealing jobs and the pressure to cut costs, a new kind of hybrid production studio is launching with the latest AI tools.

Innovative Dreams is a new production services company, backed by Amazon Web Services and Luma, a generative AI startup, that combines cameras and a giant LED wall on a soundstage with tools to apply AI from pre-production, to shooting, into post-production. By combining virtual production, motion capture, and a variety of AI tools including Luma, Google’s Nano Banana, and Bytedance’s SeeDream, Innovative Dreams says it can significantly cut down both on costs and time.

“We visually design and explore the world, then we take the footage that we filmed and start mapping that performance capture to these digital assets,” CEO Jon Erwin explained. “You’re fusing a performance with a piece of [digital] wardrobe that you like. The cool thing is the actor’s performance, the camera, the lens choice– that’s all getting through.” Erwin says this approach merges a traditional filmmaking process into a digital world, rather than replacing cameras and actors with prompts.

Innovative Dreams was born of Erwin’s production studio, Wonder Project, after he used AI to produce historical scenes in far-flung locales for its biggest show, “House of David.” (The show is available via Amazon Prime Video.) With Innovative Dreams, Erwin is going all-in on the potential for AI and virtual production to create films and shows with massive scale, without ever leaving a sound stage — and he aims to keep production in Southern California.

“It was a game-changer in House of David, so we came back from that experience thinking that other people must have been doing the same thing,” Erwin said. “We quickly realized that people weren’t.” 

The first project using this new workflow is an upcoming three-part series called “The Old Stories: Moses,” starring Ben Kingsley, set to debut this spring. The three-episode series, shot in a single week on the virtual soundstage, shows the actors in 40 locations, by putting footage from around the world into the screens in the production facility. Erwin says it would have taken a traditional production five or six weeks to shoot –and there wouldn’t have been the budget to go to so many locations.

AI video generation requires such massive amounts of computing capacity, Innovative Dreams brought on AWS as an investor and partner; AWS provides cloud and AI infrastructure to power the real-time hybrid production tools used on set, part of its broader effort to work with the entertainment industry.

“We’re providing the … tools that are going to allow for filmmakers to be able to work in ways that they simply couldn’t have been able to before and produce content much faster, much cheaper, and collaborate in ways that would accelerate production cycles at scale,” says Samira Bakhtiar, general manager of media, entertainment, games and sports at AWS.

Innovative’s other big investor and partner is AI company Luma. Valued at over $4 billion, Luma has a new agent tool that brings together multiple AI generation services into a collaborative workspace. And Erwin says they’re giving feedback to the company.

“In allowing Luma to invest and getting directly in touch with a lot of these companies and having these collaborative conversations, we’re able to actually shape the tools that we use in a pretty profound way,” he said.

But the rise of new AI tools sparks even more concerns of job losses in an industry already struggling. The Covid pandemic brought production to a halt, and then the writers and actors guild strikes shut down production again for months in 2023. Los Angeles County has lost over 40,000 entertainment industry jobs since 2022, with production activity in the city sinking to the lowest level since 1995. The guilds’ standoff with the studios was in large part due to actor and writer concerns about AI compromising their intellectual property and stealing their jobs.

“This industry has been battered by one shock after another. Construction, consolidation, cost cutting, cuts in content spending,” said entertainment attorney Jonathan Handel. “Everything is down by 25% to 35% compared to pre-COVID.”

But now, the ability to digitally create sets, wardrobe, and makeup raises questions about the potential destruction of jobs for costumers, set designers and makeup artists. 

“The question of how much job displacement there’ll be versus how much job augmentation will exist, is one that just has not played out yet and is still making people very nervous,” said Handel.

But Erwin says he thinks that the hybrid production capabilities of Innovative Dreams won’t accelerate job losses.

“There’s just an alarming lack of green lights, especially in America,” Erwin said. “I think this is a method that allows us to film here again.” 

While Erwin suggests the best industry workers will adapt their skill set to this new AI-powered world, Handel notes that AI could impact entry-level jobs, shrinking on-ramps to an already tough industry. But Erwin is bullish about AI being a tool that enables the industry to survive.

“I think this is necessary to bring jobs back to LA,” Erwin said. “We’re inventing a new method to fix something that’s become unsustainable.”

Watch the video to learn more.

An underappreciated step in the chipmaking process is poised to become the next bottleneck for artificial intelligence.

Every microchip used to power artificial intelligence must be put into hardware that can interact with the outside world. But right now, almost all of this chipmaking step, known as advanced packaging, happens in Asia, and capacity is in short supply.

It’s now taking center stage as Taiwan Semiconductor Manufacturing Co. prepares to break ground on two new plants in Arizona and Elon Musk taps Intel for his ambitious custom chip plans.

“It can emerge as a bottleneck very quickly if people are not making the CapEx investments proactively to account for the surge in fab output that’s going to be coming in the next couple of years,” said John VerWey of Georgetown University’s Center for Security and Emerging Technology.

In a rare interview, TSMC North America packaging solutions head Paul Rousseau told CNBC that the numbers “are growing very substantially.”

Its most advanced method currently in use is called Chip on Wafer on Substrate, or CoWoS, and Rousseau said it is increasing at a stunning 80% compound annual growth rate.

AI giant Nvidia has reserved a majority of the most advanced capacity available at TSMC, which is the volume leader in packaging.

But Intel is technologically on par with the Taiwanese giant.

The U.S. chipmaker has struggled to solidify a major external customer for its chip fabrication business, but its packaging customers include Amazon and Cisco.

On Tuesday, Musk also tapped Intel to package custom chips for SpaceX, xAI and Tesla at his ambitious Terafab plant planned for Texas.

Intel does the majority of its final packaging in Vietnam, Malaysia and China. Portions of Intel’s most advanced packaging happen at U.S. facilities in New Mexico, Oregon and at a site in Chandler, Arizona, where CNBC got a tour in November.

The process has come into the spotlight as AI pushes the density, performance and efficiency needs of chipmakers racing to make the best hardware for inference workloads. As transistor density approaches physical limits, new methods of packaging silicon can help.

“It’s really the natural extension of Moore’s Law into the third dimension,” Rousseau said.

For decades, individual chips, known as dies, were removed from a single wafer and packaged into a system that connects to devices like computers, robots, cars and mobile phones. More advanced packaging methods took off as chip complexity exploded in the last few years with the advent of AI.

Now, multiple dies such as logic chips and high-bandwidth memory are packaged together into one larger chip, like a graphics processing unit, or GPU. Advanced packaging is used to connect all those dies together and allow them to communicate with each other and the broader system.

“Up until about 5 or 6 years ago, nobody was doing this,” said chip analyst Patrick Moorhead of Moor Insights & Strategy, adding that packing used to be “an afterthought” that companies would assign to junior engineers.

“Now, obviously, we know it’s as important as the die itself,” he said.

The bottleneck

Nvidia has reserved the majority of TSMC’s leading CoWoS technology, and capacity is so heavily booked that TSMC has reportedly outsourced some steps to third-party companies that specialize in simpler parts of the process, such as ASE and Amkor.

ASE, which is the world’s largest outsourced semiconductor assembly and test company, sees advanced packaging sales doubling in 2026. ASE is building a large new site in Taiwan, where subsidiary SPIL also held a grand opening for another new packaging site, attended by Nvidia CEO Jensen Huang, last year.

TSMC is also ramping up two new packaging facilities in Taiwan, in addition to building two packaging facilities in Arizona.

Right now, TSMC sends 100% of chips to Taiwan to be packaged, even those made at its advanced chip fabrication plant in Phoenix, Arizona. TSMC didn’t disclose a timeline for completion of the U.S. packaging sites.

“To have that capability right next to the fab in Arizona is going to make their customers very happy,” leading packaging researcher Jan Vardaman of TechSearch International told CNBC.

That’s because it will lower turnaround time by avoiding the need to ship it back and forth between Asia and the U.S., she added.

Intel already does some packaging near its new advanced 18A chip manufacturing plant in Arizona.

The U.S. chipmaker has yet to secure a major external customer for making chips at its 18A fab, but foundry services head Mark Gardner told CNBC that the company has had customers for packaging since 2022, including Amazon and Cisco.

Nvidia is also looking to package at Intel as part of its $5 billion investment in the chipmaker that came weeks after the U.S. government invested $8.9 billion in 2025.

“Chip companies want to show the U.S. administration that they will do business with Intel, and the lower risk path with Intel is to do packaging,” Moorhead said.

When asked if Intel could find a major chip manufacturing customer through the back door of advanced packaging, Gardner said there’s “an inroad to that” with some customers.

“There’s benefits of everything being in one place,” he said.

Musk could be an early adopter for both chipmaking and packaging at Intel.

An Intel LinkedIn post on Tuesday said the company’s “ability to design, fabricate, and package ultra-high-performance chips at scale” would help Musk’s Terafab reach ambitions of producing 1 terrawatt of annual compute to power AI.

Evolving from 2D to 3D

Many chips, like central processing units, are made with 2D packaging. More complex chips like GPUs need something extra, which is the realm of TSMC’s CoWoS, a form of 2.5D packaging.

For these chips, an additional layer of high-density wiring called an interposer adds tighter interconnections so high-bandwidth memory can mount directly around the chip, effectively eliminating what’s often referred to as the memory wall.

“You just can’t get enough memory inside your compute chip to fully utilize it. So when we introduce CoWoS, we are able to bring the HBM memory right beside the compute in a very efficient way,” said TSMC’s Rousseau.

TSMC pioneered its 2.5D technique in 2012 and it has since gone through several iterations. TSMC said Nvidia’s Blackwell GPUs are the first product to be made with its latest generation, CoWoS-L.

It’s this latest capacity that has everyone worried because Nvidia has reportedly reserved the majority of it.

Intel’s leading packaging technology is called embedded multi-die interconnect bridge, or EMIB. It works similar to the process used by Taiwan Semi, but with silicon bridges in place of an interposer. 

By “embedding these really small pieces of silicon just where they’re needed,” Intel’s Gardner said, “there’s a cost advantage.”

All the players are also working on what comes next: 3D packaging.

Intel calls its method Foveros Direct, while TSMC’s is called System on Integrated Chips, or SoIC.

“Instead of having the chip side by side, now we put them one on top of the other,” explained Rousseau, adding they “can really behave as if they’re one chip and that provides a whole other level of performance gain.”

Rousseau said it’ll be a couple of years before we see TSMC packaging products with SoIC.

Meanwhile, memory companies like Samsung, SK Hynix and Micron have advanced packaging factories of their own, where they use 3D packaging to stack dies into high-bandwidth memory. 

As they hustle to get chips out the door, memory and logic chipmakers are also looking to replace bumps with copper pads in a new method called hybrid bonding, boosting the number of chips that can fit in a stack.

“Instead of a bump, we could do a pad-to-pad connection, which is almost no distance at all, and so that gives us better power performance,” explained Vardaman. “It also gives us better electrical performance since the shortest path is the best path.”

Watch: How advanced packaging is advancing AI chips in the third dimension

Oracle reports third-quarter earnings on Tuesday, and it will be an unofficial test for the artificial intelligence trade.

Following the announcement of a $50 billion financing plan at the beginning of February that included debt and equity, investors have been eager to understand the pace of dilution for current stockholders.

“The cadence matters,” said Gil Luria, equity analyst at DA Davidson, told CNBC.

Of all the hyperscalers that are leaning into AI cloud computing, Oracle has had to rely the most on financing measures to fund its ambitious data center build-out plans. Its latest debt raise included a $5 billion convertible preferred offering and roughly $25 billion in senior notes at different maturities, according to a credit investor who spoke to CNBC. The deal was oversubscribed, indicating strong demand.

Oracle’s ability to deliver data center assets to OpenAI, its main customer, is of utmost importance to investors.

Late Friday, Bloomberg reported that talks to expand its deal with OpenAI in Abilene, Texas, fell through. A source familiar with the situation told CNBC that Oracle’s deal to deliver eight sites to OpenAI remains on track and on schedule. The source asked not to be named in order to discuss a confidential matter.

OpenAI executive Sachin Katti later posted on X that while it contemplated expanding its presence in Abilene, it would look to other markets across the U.S.

“We considered expanding it further, but ultimately chose to put that additional capacity in other locations,” Katti wrote. “Today we have more than half a dozen sites under development across multiple states, including the site we’re building with Oracle in Wisconsin, where the first steel beams went up just this week.”

Katti is in charge of spearheading OpenAI’s compute infrastructure and previously held the role of AI chief and chief technology officer at Intel.

The market has become hypersensitive to any developments tied to Oracle’s $300 billion deal with OpenAI.

News of the deal initially sent Oracle’s stock up by 35% last September, its biggest intraday gain since 1992. The deal reinforced Oracle’s position as a major contender in the AI cloud computing space, putting it alongside Amazon, Google and Microsoft.

However, in late fall, Oracle surprised the market by raising a significant amount of debt, fueling investor fears that its AI build-out would be costly and put pressure on its balance sheet.

Oracle’s 5-year credit default swaps widened as bond investors were skeptical of the enterprise software company’s ability to hold on to its investment-grade credit rating, currently two notches above junk.

Credit default swaps are like insurance for investors, with buyers paying for protection in case the borrower can’t repay its debt. Bond investors told CNBC that they’ve become a popular way to hedge the risk tied to the AI trade.

Wall Street is looking for a clear read on the return on investment of Oracle’s AI bet when the company reports earnings Tuesday, in addition to any future capital raises.

Analysts have also speculated that the company may undergo consolidation measures to streamline costs.

TD Cowen wrote in a note to clients on Jan. 26 that “our channel checks indicate that Oracle is evaluating multiple paths forward to address financing questions including a 1) a RIF (Reduction in workforce) of 20-30K employees which could drive ~$8-10B of incremental free cash flow.”

Analysts there added that divestitures and securing vendor financing deals could also be in the cards.

Watch the video to learn more.

The Seahawks may have won the Super Bowl, but Anthropic also walked away with bragging rights, according to data analyzed by BNP Paribas.

The maker of the Claude chatbot saw visits to its site jump 6.5% following its Super Bowl advertisement that took a swing at rival OpenAI’s decision to bring ads to ChatGPT.

The Claude gain, which took it into the top 10 free apps on the Apple App Store, beat out chatbot and artificial intelligence competitors OpenAI, Google Gemini and Meta. Daily active users also saw an 11% jump post-game, the most significant within the firm’s AI coverage.

OpenAI’s ChatGPT had a 2.7% bump in daily active users after the Super Bowl and Gemini added 1.4%.

Claude’s user base is still much smaller than ChatGPT and Gemini.

Advertisements from artificial intelligence brands took center stage at the big game, which brought in an audience of 125 million in the U.S., as platforms made their pitch to lure more customers in a competitive market that’s showing no signs of slowing.

The high-stakes rivalry between OpenAI and Anthropic added a new layer with the duelling ads as the AI labs head toward potential IPOs later in the year.

Both companies have also gotten more vocal in public in recent weeks, with executives openly criticizing each other’s businesses.

OpenAI CEO Sam Altman attacked Anthropic’s Super Bowl ad campaign. In a post to social media platform X, Altman called the commercials “deceptive” and “clearly dishonest.”

Both Anthropic and OpenAI are also fiercely competing to win over enterprises and the best coding talent, while raising record amounts of private funding.

This week, Anthropic closed a $30 billion funding round at a $380 billion post-money valuation, more than twice its valuation at the time of its September raise.

But OpenAI leads the pack with the largest ever private tech fundraising round, which it set last year. The company is also in talks to wrap what could be a $100 billion funding round, CNBC previously reported.

— CNBC’s Kate Rooney and Drew Troast contributed to this report

SAN FRANCISCO — Daniela Amodei has an energy that’s hard to place — warm, unhurried, immediately present. She swept into a sunlit room on the ground floor of Anthropic’s headquarters in December, sat down, and immediately apologized for her mug.

“Is my, like, gigantic novelty mug going to be distracting if I have it?” she said. “I’ll defer to the wise camera people in terms of what’s in the shot.”

It was a disarming kind of normal for someone operating at this level.

Five years ago, she and her brother, Anthropic CEO Dario Amodei, led an exodus from OpenAI.

They took a core group of senior researchers and started Anthropic with a contrarian bet: That safety and business success weren’t in tension, that the real money wasn’t in viral consumer products, and that the artificial intelligence race would be won by the company that knew when to slow down.

“We really just felt more like we were running towards something than running away from something,” Daniela said of the decision to leave.

The co-founders had known each other for years before OpenAI. Her brother and co-founders Tom Brown and Chris Olah worked together at Google Brain. Others overlapped at various labs.

Anthropic’s headquarters is on “AI Alley” in San Francisco, a 230,000-square-foot glass high-rise in the shadow of Salesforce Tower. The neighborhood is a corridor of startups and tech giants reshaping the American economy from a handful of downtown blocks.

Anthropic is now valued at $183 billion and on track to nearly double that, according to a newly signed term sheet, with Microsoft and Nvidia joining its cap table.

Revenue has grown ten times annually for three consecutive years, thanks in large part to the fact that the company’s AI assistant, Claude, has become the model of choice for enterprises that care about reliability as much as capability.

Daniela Amodei mentioned, almost in passing, that Claude once helped diagnose her with a bacterial infection after multiple doctors missed it.

With her human-scale, here-and-now energy, she is by most counts the opposite of the AI founder archetype personified by OpenAI’s Sam Altman and Tesla/xAI CEO Elon Musk.

Even her brother is a version of the technical visionary who speaks in timelines to superintelligence and treats interviews like TED talks.

But Daniela Amodei is different.

If her brother is the one looking for a path to the horizon, she is pouring the foundation that lets him get there.

The customer list reads like a who’s who of global enterprise, with Novo Nordisk, the Norwegian sovereign wealth fund — the world’s largest — plus Bridgewater, Stripe, and Slack, all running Claude at scale.

Sameer Dholakia, a partner at Bessemer Ventures who invested in Anthropic, said the bet made sense for one simple reason: enterprise customers don’t churn the way consumers do.

“We really love the focus that they had, and candidly, the focus of Anthropic on safety and trust, we knew was going to play really well with the enterprise buyer, and that’s proven to be true.”

While OpenAI still leads in scale and cultural cachet, with ChatGPT a household name boasting nearly 900 million weekly active users, Anthropic is closing the gap fast — and in some areas is already ahead.

Today, Anthropic’s revenue is roughly 85% business. OpenAI is more than 60% consumer.

“One of the values and the things that we talk about a lot internally is just how not to believe the hype,” Daniela Amodei told CNBC. “For us, it’s never been about seeking attention or sort of being in the headlines. We’re really here to do the work.”

“As Anthropic goes, so goes generative AI,” said Alex Kantrowitz, founder of the independent publication Big Technology. “They have the most pure bet on this technology working. If Anthropic is able to make it work, all of GenAI is going to work. And if not, we’re going to have some serious problems.”

Anthropic’s lead in enterprise is proving durable. And it’s a lead the Amodei siblings have built together.

“It’s genuinely a privilege to run Anthropic with my sibling,” Daniela Amodei said. “We’ve known each other our whole lives — or my whole life, at least. He had four years without me, poor guy.”

“Dario and I really help each other,” she added. “He’s great at pushing me to think about the big picture… I am helpful in thinking about, like, how do we build an organization that is enduring, that’s sustainable, that’s filled with great people who really want to do the type of work that we set out to do five years ago.”

There’s always more work to do, Amodei said, and the models are getting exponentially smarter.

She framed it as a lesson from the last generation of tech: If social media companies could go back in time, knowing what their platforms would unleash, would they have done anything differently?

Anthropic, she said, is trying to answer that question now: “To talk about the risks and try to mitigate them” while it still can.

Daniela Amodei is betting it can. She’s built her whole company around that conviction.

Watch the video to learn more.

Members of the Federal Reserve rate-setting committee say they are factoring increased labor productivity into their economic forecasts as artificial intelligence technology becomes more widely adopted.

Fed Chair Jerome Powell addressed this topic in his December news conference, saying that in past technology waves “there’s always been more work and higher productivity and incomes have risen. What will happen here? We’re going to have to see.”

Economists and investors say generative AI tools in particular have potential to increase worker productivity and shake up the labor market. Powered by machine learning, these tools may improve over time as more people use them to augment their work, according to researchers writing in the National Bureau of Economic Research. 

“This is because AI can learn. And human beings can also try to utilize AI more effectively, and train AI to suit each person. And the resulting productivity gain is huge,” said Ping Wang, a professor of economics at Washington University in St. Louis and co-author of “Artificial Intelligence and Technological Unemployment.” 

Wang and his co-author, Tsz-Nga Wong, a senior economist at the Federal Reserve Bank of Richmond, modeled various scenarios for AI’s development. In an “unbounded growth” scenario, in which the technology becomes fully developed over many decades, 23% of workers lose employment and labor productivity increases by as much as three to four times.

“Over the next decade which is more like an intermediate run, labor productivity will increase by about roughly 7% per year,” said Wang in an interview with CNBC. He noted that this is a hypothetical scenario that may not unfold.

The potential effects could affect the employment side of the Federal Reserve’s dual mandate. The Federal Open Market Committee in December forecasted a federal funds rate settling near 3% over the longer run. This may be a moderately accommodative posture relative to an estimated medium-run neutral interest rate at 3.7%, according to Cleveland Fed economists. 

Some investors see similarities between today’s rush to build data centers and a capital expenditures boom on network components in the 1990s.

“The fact that we see a run up in valuations makes us a little more cautious about future returns,” said Dan Tolomay, chief investment officer at Trust Company of the South in an interview with CNBC.

Watch the video to learn more about how AI affects the Fed’s economic outlook.

Lucid Motors gets rave reviews from critics. But it’s sorely lacking customers.

That’s a problem the company can’t afford.

The Arizona-based EV maker has top-shelf tech, deep-pocketed backers, and highly praised cars. However, it has struggled to meet production targets, and has been unable to steal the spotlight away from established luxury brands with century-old pedigrees.

Lucid is ramping up production of its high-end, three-row, Gravity SUV, though it has sold only a few hundred units so far in 2025. The Gravity’s production ramp has faced a slew of challenges, primarily supply chain shortages.

Yet the company already has plans for another vehicle aimed more at the middle of the market, where it would compete with the top-selling Tesla Model Y SUV. And it’s investing in self-driving cars for consumers while working on a robotaxi fleet with Uber and self-driving tech maker Nuro.

In the process, Lucid is burning through a lot of money. The company’s third-quarter results were worse than Wall Street expected, with a net loss of close to $1 billion.

“Their gross profit has been getting kind of worse,” said Tom Narayan, an analyst at RBC Capital Markets. “A lot of people are doing the math. How long can the company keep losing cash?”

Adding to its challenges is a tougher environment for all EV manufacturers. Demand has fallen short of expectations, and many automakers are pulling back. EVs have lost key support from the federal government, including a $7,500 tax credit, funding for charging, and restrictions on state level programs that incentivize automakers to produce zero-emission vehicles.

A ‘fantastic car’

Lucid’s first vehicle, a sedan call the Air, is the most popular vehicle in its segment, according to Cox Automotive. Through the third quarter, it was the third best-selling full-size luxury sedan, and the top selling electric one, according to the company. The Air is frequently a “critic’s pick.” No other EV can touch the 512-mile range of the Air Grand Touring, one of its top trim levels.

Last year, Lucid delivered 10,241 vehicles, the majority of which were Air sedans, a 71% increase from 2023. U.S. EV leader Tesla delivered 1.8 million.

Unfortunately, sedans have consistently compared to SUVs, crossovers, and pickups–which now all but dominate the roads. Of the top 10 best-selling models in the U.S., seven are from those three segments, according to Edmunds.

“It was a fantastic car,” said Sam Abuelsamid, vice president of market research for Telemetry. “It still is a fantastic car. But it came to market at kind of the wrong time.”

In 2023, Tesla’s Model Y was the best-selling vehicle in the world, according to JATO Dynamics. It sold more than 265,000 units in the U.S. through the third quarter of 2024, according to Cox Automotive, about 100,000 more than the Model 3 sedan. Among EVs, the Model 3 is unusually popular for a sedan. Behind the two Tesla Models, the three vehicles that round out the top five are crossovers — Chevrolet Equinox, Ford Mustang Mach-E and Hyundai Ioniq5.

The Model S, the Air’s closest Tesla competitor in terms of size, performance, and price, sold just over 4,500 units in the same period.

The highest volume EVs, such as the Model Y, are also less expensive than the Air, which starts just above $70,000 and runs up to about a quarter of a million dollars. The Model 3 is nearly half that. The average EV transaction price in November was just above $59,000, according to Cox Automotive.

“There’s just not enough of a market for those premium electric sedans right now,” Abuelsamid said.

Lucid sold slightly more than 300 Gravity SUVs in the US through the third quarter of 2025, according to Cox Automotive. Like the Air, the Gravity has a high price tag. Nevertheless, the company has said the Gravity stands to attract six times as many customers as the Air.

“We’ll see if that’s the case,” said Narayan. “The latest numbers I’ve seen show it kind of equalizing the sales of sedans.”

Lucid interim CEO Mark Winterhoff told CNBC in an interview that the company has seen “a very good uptick in demand when it comes to the Gravity as compared to the Air.” He added that most customers are configuring the car in ways that run the price up above $100,000.

Production trouble

Demand might be strong, but Lucid also has to get the vehicle into customers’ hands. Gravity’s launch earlier in 2025 was beset by shortages of key materials like magnets, aluminum, and chips, Winterhoff said on the company’s third-quarter earnings call.

“We haven’t been able to produce as many as we wanted up until this point,” Winterhoff told CNBC. “We’re very confident right now that we solved those problems.”

Deliveries have risen for seven straight quarters, culminating in a 47% percent jump over the third quarter of 2024. Lucid has added a second shift to the final assembly section of its factory to meet demand.

The company said demand has been resilient despite worries the EV market is stalling after the federal EV incentive ended on Sept. 30.

“In October, our delivery numbers went up,” Winterhoff said. “Whereas in many other pure EV players or even EVs for incumbent players that also have [internal combustion] vehicles, the deliveries dropped down drastically.”

Still, analysts say it’s a tougher time to make EVs than it was when Tesla was ramping up the Model 3 and Model Y.

“They were the only game in town,” Narayan said. “So there wasn’t competition there. They also benefited from battery prices falling significantly. And they got a lot of government support. Today we’re in a very different world.” 

If gross profit does keep getting worse, the company will eventually have to return to investors, Narayan said. Lucid is currently about 55% owned by the Saudi Public Investment fund, according to FactSet.

In the third quarter, Lucid and the PIF agreed to increase a delayed draw term loan credit facility from $750 million to roughly $2 billion. A DDTL is a loan the company can draw on over time, rather than all at once. That brings its total liquidity to $5.5 billion. The company has said it has enough to get through the first half of 2027.

“So far, the Saudis have put many billions of dollars into Lucid, and they’ve been very patient through Lucid’s struggles as they try to ramp up production and sales,” said Abuelsamid. “It’s unclear how long they will continue to be patient.”

The company also received a $300 million investment from Uber in September to develop a platform for robotaxis with a third partner, autonomous driving tech developer Nuro. On top of that, Uber plans to buy 20,000 Gravity vehicles for the self-driving fleet.

Separately, Lucid has a partnership with Nvidia to develop what it calls the “the first true eyes-off, hands-off, and mind-off (L4) consumer owned autonomous vehicle.” L4 means Level 4, nearly the highest level of autonomy in the current system devised by the Society for Automotive Engineers.

“You need to invest first and then you reap the benefits later on,” Winterhoff said. “And therefore we have to do a lot of things in parallel.”

Smaller SUV on the way

Despite Gravity’s production challenges, the company is already at work on its next vehicle — a mid-size crossover priced closer to the industry average of about $50,000.

This could boost volumes, but deepen losses.

“If gross profit is negative at a vehicle transacting around $100,000 or more on average, what Lucid’s Air and Gravity are currently going for, what will it look like when the company is selling a vehicle that is closer to half of that?” Narayan said.

“One answer could be, well, it’s a much bigger scale, so you have better operating leverage,” Narayan said. The argument, he said, is that Lucid’s technology enables it to squeeze a lot more range out of a battery than competitors.

“Our vehicles are way more efficient by 30% to 40% than the competition, which means smaller battery use smaller battery to achieve the same range,” said Emad Dlala, senior vice president of engineering and digital at Lucid. “That means a lower [bill of materials] cost. That means better margins.”

Dlala said the Lucid Air Pure has a battery size of a Tesla Model Y, but has a range of 420 miles, about a hundred miles more than the Tesla Model Y.

Lucid says the money it’s spending goes to developing an array of technologies and manufacturing processes that are hard to copy.

“Lucid has about over 10 years of powertrain history,” Dala said. “This didn’t come overnight. There’s lots of IP patented technologies across powertrain, vehicle and software,” which lead to “gains in efficiency across the board,” he said.

That even extends to the way the cars are made.

“We use manufacturing processes here that no one else in the industry uses in powertrain,” said Adrian Price, senior vice president of operations at Lucid.

“Lucid is very highly vertically integrated,” he said, meaning it brings much of the car-making process in house. “We have a lot more of our own subassemblies. That’s also how we get the performance that other people don’t. A lot of that engineering is outsourced by other major automakers. Not only do we do the engineering, but we also do the manufacturing, and we can control some of the Lucid secret sauce by doing it in-house.”

One area where Lucid appears to lag competitors is brand awareness.

“A lot of people don’t know what Lucid is,” Narayan said. “That’s the biggest, I think, hurdle that they have.”

The company has given more thought to marketing, and raising its profile in the luxury market, where it’s going up against names like Mercedes-Benz, Porsche, and BMW.

“The buyer of a luxury brand, typically, they like things like heritage,” Narayan said. “The brand matters. So, you know, how do you create that from scratch?”

“We are very well known and have tons of accolades in the motor press, Winterhoff said. “So everybody who’s interested in cars knows about it. But that’s not the majority of vehicle buyers.”

So, it has begun a new marketing and advertising strategy for the Gravity, with Hollywood actor Timothee Chalamet as its first “global brand ambassador.”

“We’re shifting from focusing on our vehicles, about the capabilities of our vehicles, to what does it mean to own a Lucid and what does it say about you?” Winterhoff said.

Winterhoff said he’s confident the company can successfully ramp the Gravity, while building and eventually delivering its mid-size vehicle.

“That is the point where I think this is a sustainable business going forward,” he said. “We have a clear plan also to profitability.”

Watch the video to get an exclusive inside look at Lucid’s Gravity production and to find out what’s next for the automaker.

After the search for survivors and recovery of victims in tragic aviation accidents — like that of a UPS cargo plane shortly after takeoff from Louisville Muhammad Ali International Airport in Kentucky last month — comes the search for flight data and a cockpit voice recorder often called the “black box.”

Every commercial plane has them. Aerospace giants GE Aerospace and Honeywell are among a few companies that design them to be nearly indestructible so they can help investigators understand the cause of a crash.

“They’re very crucial because it’s one of the few sources of information that tells us what happened leading up to the accident,” said Chris Babcock, branch chief of the vehicle recorder division at the National Transportation Safety Board. “We can get a lot of information from parts and from the airplane.”

Commercial aircraft have become very complex. A Boeing 787 Dreamliner records thousands of different pieces of information. In the case of the Air India crash in June, data revealed both engine fuel switches were put into a cutoff position within one second of each other. A voice recording from inside the cockpit captured the pilots discussing the cutoffs.

“All of those parameters today can have a very huge impact on the investigation,” said former NTSB member John Goglia. “It’s our goal to to provide information back to our investigators who are on scene as quick as we can to help move the investigation forward.”

This crucial data can also help prevent future accidents. A crash can cost airlines or plane manufacturers hundreds of millions of dollars and leave victims’ families with a lifetime of grief.

But in some circumstances black boxes were destroyed or never found. Experts say further developments such as cockpit video recorders and real-time data streaming are needed.

“The technology is there. Crash worthy cockpit video recorders are already being installed in a lot of helicopters and other types of airplanes, but they’re not required,” said Jeff Guzzetti, aviation analyst and former accident investigator for the Federal Aviation Administration and NTSB. “There’s privacy and cost issues involving cockpit video recorders but the NTSB has been recommending that the FAA require them for years now.”

Watch the video to learn more.

— CNBC’s Leslie Josephs contributed to this report.