Connect with us

Published

on

A demo setup of racks of AI servers connected with Credo cables, displayed at the Open Compute Summit in San Jose, California.

Credo

In July, Elon Musk posted photos from inside an xAI data center called Colossus 2, which the artificial intelligence startup aims to turn into a massive supercomputing facility in Memphis, Tennessee.

Musk’s pictures, posted to his X feed, didn’t show off the pricey Nvidia racks that are filled with powerful graphics processing units. Rather, he focused on the wires behind the servers, including one image with thousands of neatly organized purple cables connecting the computers together.

Those purple cables are the signature offering of Credo, a 17-year-old Silicon Valley-based semiconductor company whose name rarely gets mentioned alongside the leaders of the AI boom.

But Wall Street has taken notice.

Credo shares have more than doubled this year to $143.61 after soaring 245% in 2024. The company’s market cap, which was about $1.4 billion at the time of its IPO in 2022, now sits at close to $25 billion. Credo is angling to position itself as a key supplier in the trillion-dollar AI infrastructure expansion, and is benefiting as the money flows downstream.

The stock jumped 5% on Friday after analysts at JPMorgan Chase initiated coverage with the equivalent of a buy rating and a $165 stock price. They said the active electrical cable (AEC) market, which Credo pioneered, is on pace to hit $4 billion by 2028, as all the major hyperscalers invest in data center buildouts.

“The industry outlook is supported by increasing deployments from major companies such as Amazon, Microsoft, and xAI as well as broadening adoption, including Meta and more,” the analysts wrote. They predict annualized revenue growth for Credo of at least 50% through 2028.

Big Tech’s AI spending spree is fueling growth. But what happens if the money stops flowing?

Revenue in fiscal 2025, which ended in early May, more than doubled to $436.8 million. The company also turned profitable, recording net income of $52.2 million after losing $28.4 million the prior year. Analysts are expecting sales to more than double again in fiscal 2026 to almost $1 billion, according to LSEG.

Credo’s purple AECs cost between $300 and $500 each, depending on bulk discounts and other negotiations, according to an estimate from the 650 Group, an industry researcher. They are sturdy, moderately thick copper cables wrapped in a braided covering with big connectors containing chips on each side.

Much of the excitement around Credo is driven by the AI boom, which to this point has been driven by a handful of hyperscalers that are rapidly building data centers for future expected workloads. Analysts expect $1 trillion in spending on AI data centers by 2030, but any pullback from the major cloud providers or scaling back in OpenAI’s plans could hurt many suppliers, including Credo.

For now, projections are way up and to the right.

Expanding opportunity

Previous servers typically had one or two processors on a motherboard. Individual servers today can have up to eight, and the most powerful AI models require potentially millions of GPUs all working together as one.

Each GPU needs its own connection to the switch, the term for a computer that routes data around the cluster, often mounted on the top of a server rack.

Nvidia’s latest products slot several of these boards together to comprise a system with 72 GPUs. Next year’s fastest racks will have twice as many, and the following year, a Kyber rack will have 572 GPUs, Nvidia says.

“In the past, Credo’s opportunity was one cable per server, but now Credo’s opportunity is nine cables per server,” said Alan Weckel, an analyst at 650 Group. He estimates that Credo has 88% of the market for AECs, which are also made by Astera Labs and Marvell.

Many GPUs are connected by fiber optic cables powered by components made by companies like Broadcom and Coherent. AECs offer an alternative to fiber optic cables. They have chips called digital signal processors on both sides that use sophisticated algorithms to pull data out of the cable, enabling much longer lengths than traditional copper cables. Credo’s longest AEC is seven meters long.

Credo CEO Bill Brennan, who joined the company in 2013, told CNBC that hyperscalers are choosing his company’s cables because they’re more reliable than fiber optic cables. He said customers are trying to avoid what’s called a “link flap,” where one part of an AI cluster goes offline because the optical cable connecting them fails, costing hours of pricey GPU time.

“It can literally shut down an entire data center,” Brennan said.

He said Credo is increasingly working with hyperscalers in the early stages of planning large AI clusters, especially as some designs become denser, allowing more servers to be connected by shorter cables.

“When you connect with these hyperscalers, the numbers are very large,” Brennan said.

Credo’s AEC leadership team, Hal Hawthorne, Don Barnetson, Ameet Suri, and Ryan Cai.

Corey Bentley, Credo

The company doesn’t name its hyperscaler clients, but analysts have cited Amazon and Microsoft as customers. Amazon Web Services CEO Matt Garman posted an image on LinkedIn of the company’s Trainium AI chip racks on Friday that appeared to show Credo’s purple cables.

Credo says it expects three or four customers to make up more than 10% of revenue each in the coming quarters, including two new hyperscale customers this year.

Amazon and Microsoft declined to comment. Meta and xAI didn’t respond to requests for comment.

At a conference for data center professionals in San Jose this week, Credo presented alongside a representative from Oracle Cloud. An example rack of Nvidia GPUs designed by Meta displayed at the show prominently featured Credo’s purple cables.

“Every time you see a new announcement of a gigawatt data center, you can rest assured that we view that as an opportunity,” Brennan told investors on an earnings call in September.

It’s a market that everyone in AI networking is targeting.

Analysts at TD Cowen estimated earlier this month that the market for AI networking chips could be worth $75 billion per year by 2030. Major players include Nvidia and Advanced Micro Devices, which both have their own networking businesses and have the power to dictate which technologies are part of their broader systems.

‘Insatiable demand’

Christina Locopo | CNBC

The AEC business didn’t take off until the AI boom in the early 2020s, because data centers didn’t yet need its technology, Brennan said.

However, there was early excitement in the air when Musk’s car company came knocking in 2017. Tesla wanted help with its Dojo AI supercomputer and needed chips with more bandwidth than what was available at the time.

Now, Credo is hoping to use its foothold with its active copper cables to branch out into additional product lines, including intra-rack connections, or what’s called “scale-up” networking. The company announced new transceivers and software for optical cables this week.

“You’ve got this market pull like we’ve never had before,” Brennan said. “If you could deliver the next generation right now, it would be consumed. Generation after that, it would be consumed. You’ve got this insatiable demand from the AI cluster world.”

WATCH: OpenAI in Abilene, Texas

OpenAI CFO Sarah Friar: 'More compute, more revenue' in response to concern on Oracle, Nvidia deals

Continue Reading

Technology

Companies are blaming AI for job cuts. Critics say it’s a ‘good excuse’

Published

on

By

Companies are blaming AI for job cuts. Critics say it’s a 'good excuse'

More companies are announcing AI-driven layoffs from Salesforce to Accenture.

Twenty20

From tech to airlines, large global companies have been slashing staff as the real-world impact of artificial intelligence plays out, spooking employees. But critics say AI has become an easy excuse for firms looking to downsize.

Last month, tech consultancy firm Accenture announced a restructuring plan that includes quick exits for workers that aren’t first able to reskill on AI. Days later, Lufthansa said it was going to eliminate 4,000 jobs by 2030 as it leans on AI to increase efficiency.

Salesforce also laid off 4,000 customer support roles in September, saying that AI can do 50% of the work at the company. Meanwhile, fintech firm Klarna has reduced staff by 40% as it aggressively adopts AI tools.

Language-learning platform Duolingo has stated that it will gradually stop relying on contractors and use AI to fill the gaps.

The headlines are grim, but Fabian Stephany, assistant professor of AI and work at the Oxford Internet Institute, said there might be more to job cuts than meets the eye.

Previously there may have been some stigma attached to using AI, but now companies are “scapegoating” the technology to take the fall for challenging business moves such as layoffs.

“I’m really skeptical whether the layoffs that we see currently are really due to true efficiency gains. It’s rather really a projection into AI in the sense of ‘We can use AI to make good excuses,'” Stephany said in an interview with CNBC.

Companies can essentially position themselves at the frontier of AI technology to appear innovative and competitive, and simultaneously conceal the real reasons for layoffs, according to Stephany.

“There might be various other reasons why companies are having to get rid of part of their workforce … Duolingo or Klarna are really prime candidates for this because there has been overhiring during Corona [Covid-19 pandemic] as well,” the professor said.

Some companies that flourished during the pandemic “significantly overhired” and the recent layoffs might just be a “market clearance.”

“It’s to some extent firing people that for whom there had not been a sustainable long term perspective and instead of saying “we miscalculated this two, three years ago, they can now come to the scapegoating, and that is saying ‘it’s because of AI though,'” he added.

This pattern has sparked conversation online. One founder, Jean-Christophe Bouglé even said in a popular LinkedIn post that AI adoption is at a “much slower pace” than is being claimed and in large corporations “there’s not much happening” with AI projects even being rolled back due to cost or security concerns.

“At the same time there are announcements of big layoff plans ‘because of AI.’ It looks like a big excuse, in a context where the economy in many countries is slowing down, despite what the incredible performance of stock exchanges suggest,” said Bouglé, who co-founded Authentic.ly.

Feeding the fear of AI

Jasmine Escalera, a careers expert, said this concealment is “feeding the fear of AI” with employees globally concerned about their jobs being replaced as a result of AI.

“So we already know that employees are scared because companies are not being honest, open and communicative about how they’re implementing AI,” Escalera told CNBC Make It. “Now companies are openly stating ‘We’re doing this [layoffs] because of AI’ so it’s feeding the frenzy.”

Escalera said big companies need to be more responsible as they set the tone for what’s the norm in business decision making and avoid greenlighting “bad behavior.”

A Salesforce spokesperson clarified to CNBC that the company deployed its own AI agent, Agentforce, which reduced the number of customer support cases and eliminated the need to “backfill support engineer roles,” they said.

View taken inside a Lufthansa Airbus A350 airplane on March 19, 2025.

Lufthansa to cut 4,000 jobs as airline turns to AI to boost efficiency

“We’ve successfully redeployed hundreds of employees into other areas like professional services, sales, and customer success,” the Salesforce spokesperson added.

Klarna directed CNBC to its co-founder and CEO Sebastian Siemiatkowski’s comments on X where he explained that the company shrank its workforce from 5,500 to 3,000 people in two years but “AI is only part of that story.”

Siemiatkowski linked the workforce reduction to slimming down its analytics team to one “success team,” with many then leaving by natural attrition as well as the reduction of the company’s customer success team.

Lufthansa and Accenture declined to comment on the matter and did not share any further details on their AI restructuring strategy. Duolingo did not respond to CNBC’s request for comment.

Mass AI layoffs are not here

The Budget Lab, a non-partisan policy research center at Yale University, released a report on Wednesday which showed that U.S. labor has actually been little disrupted by AI automation since the release of ChatGPT in 2022.

The lab examined U.S. labor market data from November 2022 to July 2025 using a “dissimilarity index” which measured how much the occupational mix—the share of workers in different jobs—has shifted since AI’s debut and compared it to other technological shifts such as the introduction of computers and the internet. It found that AI hasn’t yet caused widespread job losses.

Additionally, New York Fed economists released research in early September which showed that AI use amongst firms “do not point to significant reductions in employment” across the services and manufacturing industry in the New York–Northern New Jersey region.

It found that 40% of service firms said they were using AI this year, up from 25% last year, while manufacturing firms saw a similar jump from 16% last year to 26% this year, but very few were using AI to layoff workers.

Only 1% of the services firm reported AI as the reason for laying off workers in the past six months, down from 10% that had laid off workers using AI in 2024. Meanwhile, 12% of services firms said AI made them hire less workers in 2025.

By contrast, 35% of services firms have used AI to retrain employees and 11% have hired more as a result.

Stephany said there isn’t much evidence from his research that shows large levels of technological unemployment due to AI.

“Economists call this structural unemployment, so the pie of work is not big enough for everybody anymore and so people will lose jobs definitely because of of AI, I don’t think that this is happening on a mass scale,” he said.

He added that concerns about technology putting an end to human work can be seen throughout history.

“It reoccurred this century alone a dozen times, you can go back to ancient times where Roman emperors put hold to certain machines because they were worried about this and always the contrary happened. The machine made companies, industries more productive.

“It allowed for the emergence of entirely new jobs. If you think about the internet 20 years ago, nobody would have known what a social media influencer is, what an app developer is because it didn’t exist.”

Read more about companies conducting AI layoffs below:

A logo sits illuminated at the Accenture booth in Mobile World Congress 2025 on March 03, 2025 in Barcelona, Spain.

Accenture plans on ‘exiting’ staff who can’t be reskilled on AI amid restructuring strategy

Continue Reading

Technology

Close to half of Kalshi user base experienced glitches, delays during Saturday college football games

Published

on

By

Close to half of Kalshi user base experienced glitches, delays during Saturday college football games

The Kalshi logo arranged on a laptop in New York, US, on Monday, Feb. 10, 2025.

Gabby Jones | Bloomberg | Getty Images

Close to half of Kalshi’s user base experienced glitches and delays on Saturday during college football games, a major source of trades, as some said they were temporarily unable to process orders.

In a message sent to a user obtained by CNBC, the predictions market service’s website apologized for any inconvenience and said it was “looking into” the issues traders were experiencing. 

“The Exchange is experiencing temporary delays,” the message read. “Balances and positions may not be accurately reflected at this time.” 

One user shared a screen recording and screenshots with CNBC that showed they were unable to see their balance or bets while the issues persisted.

A number of users on X reported the website was down when they were trying to place bets on college football games, with some saying they had open orders that wouldn’t process. When CNBC visited the website, it wouldn’t load, showing only a green K with a spinning circle around it for more than 20 minutes. The platform later loaded.

“Earlier today, Kalshi experienced minor glitches that temporarily affected some user experiences. No exchange outage occurred, no funds were affected, and the issues are now resolved,” the company said in a statement.

Earlier, a spokesperson denied there was an outage and said the exchange “never stopped functioning properly.” He added that there has been no impact on clearing, advanced trading, or institutional trading.

“There were some glitches and delays on our web and app product, which affected less than half of our user base,” the spokesperson said. 

A little over a week ago, Kalshi announced a $300 million Series D funding round that valued the company at $5 billion, more than double its $2 billion valuation in June after its Series C round. 

The round was co-led by Andreessen Horowitz (a16z) and Sequoia Capital, with participation from Paradigm. Additional backers included Coinbase Ventures, General Catalyst, Spark Capital and CapitalG. 

The company, founded in 2018, rose to prominence by offering bettors the ability to trade on a wide range of real-world events, from football games to who President Donald Trump could pardon this year.

WATCH: Kalshi CEO on $2B valuation: We’re one of the fastest growing companies in America

Kalshi CEO on hitting $2B valuation: We're one of the fastest growing companies in America

Continue Reading

Technology

AI headshots are changing the way job seekers are seen and get hired in tough labor market

Published

on

By

AI headshots are changing the way job seekers are seen and get hired in tough labor market

AI headshots are becoming popular on LinkedIn and in professional portfolios as job seekers look for affordable profile pictures to give them an edge.

Since first impressions happen almost entirely through a screen, a clean, appealing photo is as important as a strong resume. And in a competitive job market, a good headshot can make a big difference. But professional photography has long been a financial barrier for many job applicants with an average starting cost for a professional headshot in the U.S. that can easily run up to hundreds of dollars.

Now job seekers are using fast and inexpensive AI tools to replace costly studio sessions.

“When I was at Yale, it was $200 for a 15-minute session for students,” said Melanie Fan, head of growth at Plush, an AI-powered online search platform for personalized shopping. “It was really expensive. The process of getting the pictures back, rendering them, looking at which ones I looked the best in, and then sending it back to the photographer for edit.”

This frustration has fueled the explosion of AI headshot tools like InstaHeadshots, PhotoPacksAI, HeadshotPro and Aragon AI, services that promise a professional image in minutes starting at under $50. Users simply upload selfies, pick a background, and receive dozens and no photographer is needed.

“After I changed my LinkedIn photo, the amount of inbound I’ve been getting from companies has skyrocketed,” Fan said. “Three to four times more messages from companies.”

Design company Canva recently launched its own AI headshot feature, with the goal of offering users a quick way to create realistic headshots and still be able to retouch or restyle them.

According to a recent Canva job market research report, 88% of job seekers believe a polished digital presence influences hiring decisions, which is up 45% from the year before. This is in line with the general uptick in use of AI as part of the application and hiring process, with 90% of hiring managers saying they have used AI to help with the hiring process, and 96% of job seekers who used AI in the application process saying they received callbacks.

Danny Wu, Canva’s head of AI products, said the goal wasn’t to replace real photography, but to make high quality imagery attainable to everyone no matter the budget or location. Once a user uploads an image, Canva can use AI for adjusting or changing the background, placing something in a different place, and for styling. “This is just a more accessible way to get professional and unique headshots,” Wu said.

Risks and questions about authenticity among HR recruiters

Anyone with a phone can get a LinkedIn-ready headshot, but the technology’s rapid adoption has created new questions about ethics and trust. Many candidates fear looking fake or deceptive and recruiters are on the lookout for AI-generated portraits that look overly smooth or stylized, saying authenticity matters the most.

“It is perceived as risky to use an AI headshot,” said Sam DeMase, ZipRecruiter career expert. “While recruiters accept them, a bad AI-generated headshot will put off most recruiters,” DeMase said. “A poorly done AI-generated headshot is easily recognized, reads as inauthentic, and can hurt the candidate’s chances of being selected.”

However, recruiters are struggling to tell if a headshot is AI produced, and the technology will only get better. “It’s becoming more and more difficult to tell whether a headshot has been enhanced or generated by AI,” DeMase added.

Chris Bora, founder and principal AI architect of Bora Labs and a former Meta engineer, said he built his own headshot generator, Nova Headshot, after being disappointed by existing options. “Some made me look taller and skinnier,” Bora said. “The other ones, they made me look lighter, so it wasn’t really me,” he said. “You don’t need to spend thousands to look professional anymore. You just need a tool that makes you look like yourself on your best day. With Nova, it takes less than ten minutes,” Bora said.

Amber Collins, an AI headshot user, said she still feels uneasy about it, especially since not every app gets it right. “There are a lot of bad apps out there,” Collins said. “Seven fingers, half a necklace, and the rest of it is gone from your neck. I feel guilty using AI. There’s a stigma. I’d 100% prefer to get actual get headshots done,” Collins said.

But ultimately, she says, the benefits outweighed the risks. “In this economy, you have to be mindful of where you’re going to put your money. I don’t need to have my face out there excessively, but having a couple of really good, solid, professional looking headshots is worth it to me,” Collins said.

Wu said the goal for job applicants seeking a headshot should be to use Canva’s tool to balance realism and creativity without losing their identity.

The tension between tech innovation and accessibility on the one hand, and authenticity on the other, will remain.

A LinkedIn spokesperson told CNBC what while the platform does allow the use of tools, including AI, to enhance or create profile photos, “the photo must reflect your likeness.”

“Profile photos that don’t comply with our user agreement or professional community policies may be removed,” the LinkedIn spokesperson said.

DeMase noted that many job candidates remain hesitant to use an AI headshot. “A headshot is one of the few places you can inject humanity into the job search,” he said.

But with job seekers now able to provide the appearance they had access to the same studio lighting, camera, and editing team as the pros, the trend is unlikely to stop.

A recent survey found that headshot use among job seekers is the highest within the Gen Z and millennial generations. And while recruiters may say they still prefer real photos, AI headshots are becoming harder to spot, and less likely to even be reviewed by humans in the first stages of the application process. A recent study from the HR trade group SHRM found that 66% of human resource professionals are using AI to generate their job descriptions, and 44% are using the technology to review or screen applicant resumes.

Continue Reading

Trending