Broadcom said Monday that it’s agreed to produce future versions of artificial intelligence chips for Google, and signed an expanded deal with Anthropic that will give the AI startup access to about 3.5 gigawatts worth of computing capacity drawing on Google’s AI processors.

Shares of Broadcom rose 3% in extended trading.

The disclosure in a securities filing underscores the surging demand for infrastructure that can run generative AI models. Anthropic’s popularity has soared this year, with its Claude app becoming the top free U.S. app listed in Apple’s App Store in February after a dispute between the company and the Pentagon became public.

On an earnings call last month, Broadcom CEO Hock Tan said that “for Anthropic, we are off to a very good start in 2026” in providing 1 gigawatt of compute from Google’s homegrown tensor processing units (TPUs). Broadcom helps Google make its TPUs.

“For 2027, this demand is expected to surge in excess of 3 gigawatts of compute,” he said.

In a note following the earnings call, analysts at Mizuho led by Vijay Rakesh estimated that Broadcom would pick up $21 billion in AI revenue from Anthropic in 2026 and $42 billion in 2027. The filing on Monday did not contain a dollar amount.

Meanwhile, Broadcom is also collaborating with Anthropic rival OpenAI on custom silicon for AI. Both model builders currently rely heavily on graphics processing units from Nvidia through cloud providers such as Amazon, Google and Microsoft. OpenAI has also committed to drawing on six gigawatts of AMD’s GPUs, with the first gigawatt set to come in the second half of this year.

WATCH: Final Trades: Broadcom, Spotify, Applovin and Uber

AI data centers are becoming a “stress test” for insurers as rapid technological advancements and the use of increasingly complex financial structures present a unique set of challenges and opportunities for the sector.

Global spending on data centers could reach $7 trillion by 2030, according to McKinsey, and much of that spending can no longer come solely from hyperscalers. Instead, Big Tech is increasingly tapping private equity, private credit and using debt to finance the capital-intensive build-out of the facilities.

Private infrastructure data center deals were consistently above the $10 billion mark last year, according to data from Preqin. The largest deal amounted to $40 billion, with Nvidia, Microsoft, BlackRock and Elon Musk’s xAI forming part of a consortium of investors to buy Aligned Data Centers.

The fact that so much money is tied up in building, constructing, and running data centers has been a “real stress test” over the last four to five years for the major insurance companies, Tom Harper, data center leader at insurance broker Gallagher, told CNBC.

“When you put $10 to $20 billion plus in a single location, it creates capacity issues in the marketplace. The marketplace has always had an appetite for these risks because they are such high-quality builds. They’ve got cutting-edge technology, they’re AA plus plus construction locations, but the capacity — the ability to provide the insurance capacity at these locations — has been tough.”

It was nearly impossible to reasonably insure a $20 billion campus in 2023, according to Harper. In 2026, however, it’s become a weekly conversation.

Estimated spending on AI data centers has been referred to as the biggest peacetime investment project in history. Rajat Rana, partner at Quinn Emanuel Urquhart & Sullivan, told CNBC he would take it a step further and stress that this is the “largest peacetime investment project in human history, which is financed largely off balance sheet.”

Rana, who worked on structured finance litigation in the wake of the housing crisis triggered by the 2008 Financial Crash, said tracking developments in AI data center financing feels like “deja vu.”

“We’re talking about trillions of dollars, and almost going back to the same cycle where there’s almost no transparency about the financing structures — the scale is astronomical,” he said.

The AI boom is not only driving a rush in demand for the facilities, it’s also spurring rapid advancements in power generation and chips — the critical tech that the data centers house. The advancements and huge sums of money flowing into the sector pose both risks and rewards for insurers and lenders.

The M&A boom is also keeping transactional lawyers busy, with Kirkland & Ellis noting that a number of companies are forming data center specific teams, enlisting specialists across real estate, power, telecom, finance, insurance, trade, private equity and cybersecurity.

Professional services firm Marsh launched a dedicated digital infrastructure advisory group designed to help clients as contracts become increasingly complex.

Last year, Marsh also launched Nimbus, a 1-billion-euro ($1.2 billion) insurance facility for covering the construction of data centers in the U.K. and Europe. Seven months later, it expanded the facility to offer limits of up to $2.7 billion.

“Private credit can meaningfully complement banks and can support non‑hyperscale contracted offtakes,” said Alex Wolfson, senior vice president of credit specialties at Marsh Risk.

As data center loans increase, insurers who protect lenders if a borrower doesn’t pay, are starting to hit limits, Wolfson explained. Marsh is working on solutions to support lenders.

However, Quinn Emanuel’s Rana cautioned that when it comes to data centers, it’s not easy for insurance companies to fully understand the risk as financing moves off the balance sheet.

He noted that in January, four U.S. senators called on the government to investigate how Big Tech is increasingly turning to “complex and opaque debt markets to borrow staggering sums of cash.” In an open letter, the senators warned that massive debt loads could cause “destabilizing losses” for financial institutions, triggering a broader financial crisis that harms the economy.

That increased opacity in financing can lead to second-order litigation risks for downstream investors such as pension funds, insurers and asset managers invested in private credit funds who later learn they were not fully aware of concentration risk, Rana said in a note published in March.

He told CNBC that some PE funds have reached out to him with concerns about commercial leases and the valuation of properties.

Tenants are trying to negotiate the extensions of their properties and landlords are disputing the value as they look for higher prices for AI data centers.

“I’m not a doomsday guy who’s saying, hey, it’s gonna crash. My point is, whether it crashes or not, the disputes are inevitable, and we have already seen those disputes,” Rana said.

A key debate around potential cracks in financing centers on GPUs and the risk that their lifecycles may not align with the longer lifespan of the facilities that house them.

CoreWeave, which sells AI tech in the cloud, is the first company to secure GPU-backed loans, essentially using the value of the high-performance chips as collateral. Last week, the company announced it secured $8.5 billion in a first investment-grade rated GPU-backed deal. Its stock jumped 12% on the day.

While data centers typically have a decades-long lifecycle, the average lifecycle of a GPU is around seven years.

“There are different data centers that are raising debt by disclosing different life cycles to investors,” said Rana. He referred to the problem as the “GPU debt treadmill,” a phrase coined by AI commentator Dave Friedman.

“This is almost like a treadmill that these AI data centers are running on,” Rana told CNBC. Even if the financing structure is ring-fenced and backed by an investment-grade counterparty, the real risk may lie in whether an equity issue today later evolves into a credit problem over time.

“As these new chips come in, the data centers will feel pressured to raise more debt, and then they will have to build new infrastructure, and then that basically creates a billion-dollar question: how fast can you build these facilities? How fast can you raise credit?”

The cost of funding these projects is likely to continue to fuel recent growth in asset-backed securitization deals, says Harper, with greater volumes of commercial mortgage-backed securities sold to investors.

For some insurers, like Gallagher, the changing dynamics in the sector are opportunities rather than challenges. Harper said the lifecycles of GPUs have been increasing. Where things have depreciated quickly, Gallagher has had to get creative and write bespoke insurance polices with a predetermined agreement on how to value the assets.

“It would be a nightmare with the size and scope of these [facilities] to determine [the value of] each individual unit,” he said.

Harper also stressed that GPUs are interchangeable. The firm has seen operators anticipate relatively short life cycles and construct facilities that are more modular in response.

“There is a core tension in data center project finance: lenders typically want asset lives that exceed loan tenors by a comfortable margin, and the shorter useful life of GPUs challenges that assumption,” said Marsh Risk’s Wolfson.

Lenders are therefore structuring loans more cautiously to protect themselves.

It pinches here; drags there; the draping is wrong. These are some of the examples of the feedback a new crop of artificial intelligence apps might give a prospective customer trying on clothing ahead of a purchase, and in the process reduce the chances of a product being returned to a store.

Fashion retailers are increasingly turning to AI to solve the issue of rising product returns, a persistent drag on profitability and something many in the industry refer to as the industry’s “silent killer”.

A growing number of AI start-ups have emerged to provide virtual try-on technology, allowing potential customers to visualize fit and style before they buy.

While tech companies have attempted to solve online fit issues since the 2010’s, the rapid development of generative AI has finally made these applications good enough to meaningfully impact retailers’ bottom lines. 

The U.S. National Retail Federation late last year estimated that 15.8% of annual retail sales were returned in 2025, totaling $849.9 billion. For online sales, that number jumped to 19.3%. Gen Z is driving this trend, with shoppers aged 18 to 30 averaging nearly eight online returns per person last year, the NRF found.

Most returned items never make it back to the shelves and often cost the retailer more to process than the value of the refund itself. It’s a multibillion-dollar problem for the industry that’s eating directly into companies’ margins.

“Figuring out how to proactively use returns and then how to minimize them can be a meaningful driver of business and profitability,” Guggenheim Senior Managing Director Simeon Siegel told CNBC.

While fit technology will never be as good as trying something on in person, it’s a great way to bridge the gap, Siegel said. “It’s going to continue to get better, I think that’s going to continue to reduce returns.”

The primary reason for returns and abandoned shopping carts is uncertainty over fit, Ed Voyce, founder and CEO of AI startup Catches, told CNBC in an interview.

Catches has developed a platform that allows users to create a “digital twin” to try on clothes virtually with what it calls “mirror-like realism.” The application went live last month on luxury brand Amiri’s website for a select range of clothes.

Unlike other models that Voyce says “just look pretty,” the Catches platform incorporates the physics of fabric texture and how material interacts with a moving body.

“The reason we built Catches was to take advantage of a kind of confluence of technologies that is taking place right now to solve this issue effectively,” says Voyce, who founded the startup backed by LVMH’s Antoine Arnault and built on Nvidia’s CUDA platform.

While e-commerce has grown rapidly in recent years, with online shopping driving retail sales growth, the current U.S. trade policy under President Donald Trump has put a dampener on the sector which relies heavily on manufacturing in Southeast Asia. Across the price spectrum, retailers are struggling to maintain margins as costs rise and consumers become increasingly price sensitive amid inflationary pressures.

The online fast fashion player has been experimenting with virtual try-ons in partnership with deep-tech startup AIUTA, allowing prospective customers to see a piece of clothing on a range of body types, heights, and skin tones. ASOS, however, cautions that the tool is designed to give general guidance and that customers must still check size guides before purchasing. 

Shopify, meanwhile, has integrated startup Genlook’s AI virtual try-on app into its commerce platform, which it says “removes sizing doubts, boosts buyer confidence and drives higher conversion rates while reducing costly returns.” 

Tech giants like Amazon, Adobe, and Google have also created virtual try-ons in various shapes and forms, partnering with major brands to roll out the technology. 

From April 30, Google’s virtual try-on tech can be accessed directly within product search results across Google platforms, according to Google Labs’ website. 

As for Catches, it projects that its app can drive a 10% increase in conversions and a 20- to 30-times return on investment for brand partners. It focuses on luxury brands because of their higher price point. The startup hasn’t yet put a number on how much returns might decline with the use of its platform, but targets “massive reductions.”

“There are certainly companies that have absolutely seen benefits – figuring out how to quantify them is more difficult,” said Siegel. 

While the benefits are clear, the analyst cautions that AI is not a magic wand. Beyond fit, retailers are looking at AI for inventory management, customer targeting, and fraud prevention.

“All of those are really interesting use cases, as long as companies don’t abandon who they are,” Siegel says.

“What you sell is always going to be more important than how you sell, and so I just think remembering that will help dictate who wins and benefits and amplifies from AI versus who gets consumed by it.”

Oracle rose in premarket trading on Wednesday as the multinational tech conglomerate looks to cut thousands of jobs to free up cash to build AI data center infrastructure.

The software giant has started telling its 162,000-strong workforce that thousands of people will be affected in a new round of layoffs, two people familiar with the matter told CNBC on Tuesday. Its shares were last up 2.6% in early market trading on Wednesday. Oracle declined to comment on CNBC’s report.

Investors remain uneasy about the company’s hefty capital expenditure on data centers that can handle AI workloads. While shares closed up nearly 6% Tuesday, Oracle’s stock is down roughly 25% so far this year.

The company announced plans in early February to fundraise up to $50 billion during the 2025 calendar year through a mixture of debt and equity, to expand capacity for contracted cloud demand from customers, including Nvidia, Meta, OpenAI, Advanced Micro Devices and xAI.

Major AI hyperscalers Alphabet, Microsoft, Meta and Amazon have also committed to capital expenditure of nearly $700 billion to fund their AI buildouts this year, which has alarmed investors as it will reduce the companies’ free cash flow without a clear promise on near-term returns.

Job cuts at Oracle will help free up cash flow, Barclays analysts said in a note on Thursday. The investment bank said it is its overweight rating of the stock.

“Given ORCL’s existing FY26 Restructuring Plan and prior reports, we do not see today’s layoffs as being a surprise to the market, which seemed to have appreciated the cost savings potential from ORCL’s actions amidst the company’s rapid build-out of AI infrastructure capacity,” the analysts said.

Barclays also highlighted that Oracle generates less profit per employee than its competitors, with workers less productive compared to the average. The analysts expect that Oracle will triple its revenue over the next few years due to minimal headcount growth and low operating costs.

Microsoft just closed out its worst quarter on Wall Street since the 2008 financial crisis, as investors soured on the software giant’s prospects in artificial intelligence.

The company’s stock plunged 23% in the first quarter, a steeper drop than any of its tech peers or the Nasdaq, which fell 7% in the period. Microsoft bounced back a bit on Tuesday, alongside a broader market rally, with shares of the company gaining 3.3%, the biggest jump since July.

While Microsoft remains dominant in workplace productivity software and through its Windows operating system, the company is facing twin pressures to grow efficiently in AI while also building out its cloud AI infrastructure to support soaring demand.

Oil prices are surging because of the Iran war, potentially driving up costs for building and running data centers. And on the product side, Copilot, Microsoft’s AI assistant, has yet to show a lot of traction as users flock to competitive services from Google, OpenAI and Anthropic.

“Redmond is in a pickle,” wrote Ben Reitzes, an analyst at Melius Research, in a note on March 23, referring to Microsoft’s headquarters in Washington state. Reitzes, who has a hold rating on the stock, said the company has to use valuable capacity from its Azure cloud to fix Copilot, but has no choice “since Copilot is needed to maintain momentum in its most profitable and largest segment.”

Microsoft declined to comment.

Meanwhile, software stocks are getting pummeled as part of an AI-inspired “SaaSpocalypse” that has pushed names like Adobe, Atlassian and ServiceNow down more than 30% this year.

“Much of traditional SaaS is dying/in likely terminal decay,” Jason Lemkin, founder of SaaStr, wrote this week in a post on X, using the acronym for software as a service. In a blog post, he noted that earnings multiples for software trail the S&P 500.

Microsoft’s multiple hasn’t been this low since the fourth quarter of 2022, when OpenAI introduced ChatGPT, according to Capital IQ data.

Gil Luria, an analyst at DA Davidson, told CNBC that the sell-off isn’t justified, and he recommends buying shares. In the latest quarter, Microsoft reported revenue growth of almost 17%, accelerating from a year earlier.

“The dislocation in the fundamental performance of Microsoft and the stock performance of Microsoft, and the valuation of Microsoft, is the biggest it’s been in decades,” Luria said. He said he expects the company’s earnings growth to outpace the broader market this year.

“There is no stickier product in all of enterprise software than Microsoft Windows and Office,” he said.

Microsoft has been trying to build a larger revenue base from productivity software with the Microsoft 365 Copilot AI add-on, but so far, just 3% of commercial Office customers have licenses for it. Luria said he has access to 365 Copilot, but that he’s not a fan. More importantly, he said, Microsoft has pricing power with Office subscriptions. The company announced plans to raise prices in December.

With Copilot struggling to win over users, Microsoft said two weeks ago that Mustafa Suleyman, the former co-founder of AI lab DeepMind who had been running Copilot development for consumers, will focus on building AI models. Microsoft has tasked former Snap executive Jacob Andreou with leading the Copilot experience for consumers and commercial clients.

“Sure sounds like a demotion at best,” former Jane Street trader Agustin Lebron wrote on X. The change followed departures of prominent executives, including gaming chief Phil Spencer and Rajesh Jha, Microsoft’s highest-ranking productivity leader, who’s retiring.

Microsoft is still getting healthy growth out of Azure, which is second to Amazon Web Services in cloud infrastructure. Revenue in the division jumped 39% in the December quarter. Finance chief Amy Hood said in January that growth could have been in the 40s if the company had allocated all of its AI chips to Azure, rather than giving some to teams operating services such as Microsoft 365 Copilot.

Azure is benefiting from a massive backlog of business from OpenAI and Anthropic. Microsoft’s commercial remaining performance obligations at Azure more than doubled in the December quarter from a year earlier to $625 billion.

It’s a reminder that, among tech’s hyperscalers, Microsoft was viewed as an early mover in generative AI due to its 2019 investment in OpenAI and strategic partnership with the startup. But the companies no longer have an exclusive arrangement when it comes to cloud infrastructure and are now competing in a number of areas.

In February, OpenAI announced a service called Frontier that the company said “helps enterprises build, deploy, and manage AI agents that can do real work.”

Microsoft CEO Satya Nadella has been wearing a brave face, promoting the company’s AI enhancements on social media.

“It’s a lot of intense competition, but it’s not so zero-sum, as some people make it out to be,” he said in January.

Aaron Foresman, managing director of equity research at Crawford Investment Counsel, a Microsoft investor, said Nadella’s continuing presence is crucial for the company that he’s been leading since replacing Steve Ballmer in 2014.

“We’ve got a lot of trust and confidence in Satya,” Foresman said.

WATCH: Bank of America’s Tal Liani talks reinstating Microsoft as a ‘buy’

A federal judge in San Francisco granted Anthropic’s request for a preliminary injunction in its lawsuit against the Trump administration. 

Judge Rita Lin issued the ruling on Thursday, two days after lawyers for the artificial intelligence startup and the U.S. government appeared in court for a hearing. Anthropic sued the administration to try to reverse its blacklisting by the Pentagon and President Donald Trump’s directive banning federal agencies from using its Claude models.

Anthropic sought the injunction to pause those actions and prevent further monetary and reputational harm as the case unfolds. The order bars the Trump administration from implementing, applying or enforcing the president’s directive, and hampers the Pentagon’s efforts to designate Anthropic as a threat to U.S. national security. 

“Punishing Anthropic for bringing public scrutiny to the government’s contracting position is classic illegal First Amendment retaliation,” Lin wrote in the order. A final verdict in the case could still be months away. 

During Tuesday’s hearing, Lin pressed the government’s lawyers about why Anthropic was blacklisted. Her language in Thursday’s order was even sharper.

“Nothing in the governing statute supports the Orwellian notion that an American company may be branded a potential adversary and saboteur of the U.S. for expressing disagreement with the government,” she wrote.

Following the ruling, Anthropic said it’s “grateful to the court for moving swiftly.”

“While this case was necessary to protect Anthropic, our customers, and our partners, our focus remains on working productively with the government to ensure all Americans benefit from safe, reliable AI,” the company said in a statement.  

Anthropic’s suit earlier this month followed a dramatic couple weeks in Washington D.C., between the Department of Defense and one of the most valuable private companies in the world.

In a post on X in late February, Defense Secretary Pete Hegseth declared Anthropic a so-called supply chain risk, meaning that use of the company’s technology purportedly threatens U.S. national security. In early March, the DOD officially notified Anthropic about the designation via a letter.

Anthropic is the first American company to publicly be named a supply chain risk, as the designation has historically been reserved for foreign adversaries. The label requires Defense contractors, including Amazon, Microsoft, and Palantir, to certify that they do not use Claude in their work with the military. 

The Trump administration relied on two distinct designations – 10 U.S.C. § 3252 and 41 U.S.C. § 4713 – to justify the action, and they have to be challenged in two separate courts. Because of that, Anthropic has filed another lawsuit for a formal review of the Defense Department’s determination in the U.S. Court of Appeals in Washington. 

Shortly before Hegseth declared Anthropic a supply chain risk, President Donald Trump wrote a Truth Social post ordering federal agencies to “immediately cease” all use of Anthropic’s technology. He said there would be a six-month phase-out period for agencies like the DOD.

“WE will decide the fate of our Country — NOT some out-of-control, Radical Left AI company run by people who have no idea what the real World is all about,” Trump wrote.

The Trump administration’s actions surprised many officials in Washington who had come to admire and rely on Anthropic’s technology. The company was the first to deploy its models across the DOD’s classified networks, and it was championed for its ability to integrate with existing Defense contractors like Palantir. 

Anthropic signed a $200 million contract with the Pentagon in July, but as the company began negotiating Claude’s deployment on the DOD’s GenAI.mil AI platform in September, talks stalled.

The DOD wanted Anthropic to grant the Pentagon unfettered access to its models across all lawful purposes, while Anthropic wanted assurance that its technology would not be used for fully autonomous weapons or domestic mass surveillance. 

The two failed to reach an agreement, and now, the dispute will be settled in court. 

“Everyone, including Anthropic, agrees that the Department of [Defense] is free to stop using Claude and look for a more permissive AI vendor,” Lin said during the hearing Tuesday. “I don’t see that as being what this case is about. I see the question in this case as being a very different one, which is whether the government violated the law.

WATCH: Anthropic vs. Pentagon hearing

Thursday’s Opening Day may be the calm before the storm for Major League Baseball.

The league’s collective bargaining agreement with its players expires at the end of this season. Owners, with the commissioner’s backing, are almost sure to push for a salary cap (which would likely come with a salary floor to get players to the negotiating table).

MLB owners have never been able to get a cap passed by the players union. It’s unclear if the end of the 2026 season will lead to a different result, but MLB Players Association Interim Executive Director Bruce Meyer told ESPN last month he expects a lockout is “all but guaranteed.”

In addition to the CBA’s expiration, there are major shifts underway for baseball media rights. One-third of the league’s teams didn’t have local TV deals in place for this season until this week. 

Nine MLB teams – the Washington Nationals, Seattle Mariners, Milwaukee Brewers, St. Louis Cardinals, Miami Marlins, Tampa Bay Rays, Cincinnati Reds, Kansas City Royals, and Detroit Tigers – announced Wednesday their brand new MLB-operated team channels will be carried by DirecTV.

Most of those teams had previously been part of Main Street Sports (previously Diamond Sports Group), which operates FanDuel Sports Networks (previously Bally Sports). That entity has been teetering with liquidation, and the teams terminated their contracts with the company due to missed payments earlier this year.

A 10th team, the Atlanta Braves, is launching a new network called BravesVision. The Braves and Charter’s Spectrum announced a multiyear distribution agreement earlier this week. 

MLB ideally wants the rights to all 30 teams in its control by the end of the 2028 season so that it can sell the in-market local games as a national package to a streamer. That would become the modern replacement to regional sports networks, and it would likely be a new, coveted package for streaming services such as ESPN and Amazon Prime Video.

Also at the end of the 2028 season, MLB’s national media rights for all of its packages will expire, allowing the league to redistribute games to its partners and potentially select new ones. 

NBC, ESPN, Fox and a combined CBS/Turner have dominated national rights for the past few decades.

“The key in media negotiations now is having all of your rights available,” MLB Commissioner Rob Manfred told me last year. “If you have all of your content – all of your playoffs, all of your regular season – available, there will be buyers, and I’m confident there will be buyers at a higher price for us.”

Manfred has even floated the idea of expanding to 32 teams and realigning the league geographically, upending or even eliminating the American and National leagues that have existed for more than 100 years. 

And yet, baseball isn’t struggling — on the contrary. The implementation of the pitch clock in 2023 has led to shorter games, rising attendance and higher TV ratings. 

More than 50 million people in the U.S., Canada and Japan watched Game Seven of the World Series last year – the most-watched baseball game in 34 years. MLB recently wrapped up the World Baseball Classic – a global preseason tournament – which captured nearly 11 million viewers on Fox and Fox Deportes for its final game.

MLB team valuations rose 13% from last year. The average MLB team is now worth $2.95 billion, according to CNBC Sport data.

Still, the profitability of the league is in far worse shape than it is for the NFL, NBA and NHL, according to CNBC’s calculations. In 2025, MLB’s 30 teams had an EBITDA — earnings before interest, taxes, depreciation and amortization — margin of under 2%. Team average revenue was $426 million with average EBITDA of $7 million, including non-MLB ballpark events. In contrast, the comparable margin for the NFL was 20%; the NBA, 21% and the NHL, 22%, according to CNBC’s most recent valuations.

The new CBA at the end of this season could be the first significant step toward a very different MLB. But, similar to the WNBA, which announced its new CBA earlier this week, MLB must ensure negotiations to get a new labor agreement don’t jeopardize a wave of positive momentum.

U.S. District Judge Rita Lin said Tuesday that the decision by the Pentagon to blacklist Anthropic’s Claude artificial intelligence models “looks like an attempt to cripple” the company.

Anthropic appeared in San Francisco federal court on Tuesday to ask Lin to temporarily pause the Pentagon’s blacklisting and President Donald Trump’s directive banning federal government agencies from using that technology.

If the preliminary injunction is awarded, the AI startup will be able to continue doing business with government contractors and federal agencies as its lawsuit against the Trump administration plays out in court.

Without the injunction, the company has said, it could lose billions of dollars in business.

Earlier in March, the Department of Defense designated Anthropic a so-called supply chain risk, meaning that use of the company’s technology purportedly threatens U.S. national security. It was the first time an American company had been hit with that designation.

The label, if allowed to continue, will require defense contractors, including Amazon, Microsoft, and Palantir, to certify that they do not use Claude in their work with the military.

“This is something that has never been done with respect to American company,” Anthropic’s counsel Michael Mongan said during the hearing. “It is a very narrow authority. It doesn’t apply here, and it’s not a normal way to respond to the concerns that have been articulated by the other side.”

Palantir is continuing to use Claude in its work with the department as the legal battle plays out, CEO Alex Karp told CNBC on March 12. Anthropic’s model is also being used in the war with Iran.

Anthropic has argued that there is no basis to consider the company a supply chain risk. The company also said it is being unfairly retaliated against because it demanded that the DOD not use Claude for fully autonomous weapons or mass surveillance of Americans. The Pentagon insists it does not use the AI models for such purposes.

Lin said she expects to issue an order on Anthropic’s motion in the next few days.

On Monday, the judge gave lawyers for Anthropic and the government a list of questions she wants answered at the hearing.

One of those questions was: “What evidence in the record shows that Anthropic had ongoing access to or control over Claude after delivering it to the government, such that Anthropic could engage in acts of sabotage or subversion?”

In its motion seeking a preliminary injunction, Anthropic argued that such an order would prevent the company from incurring further economic and reputational harm.

“The government has infringed on Anthropic’s right to speak freely; it has disparaged the company’s good name by stigmatizing it with an unlawful designation as a national security risk; it has deprived Anthropic of government contracts and damaged its relationships with business partners in the private sector; and it has put millions, possibly billions, of dollars at risk,” the motion stated. “Absent immediate relief from this Court, those harms will continue to mount.”

The company also noted that an injunction would not require the U.S. government to use its models or prevent it from transitioning to another AI vendor. 

Before the conflict erupted in late February, Anthropic was one of the first AI companies to partner with many U.S. agencies as the government sought to rapidly upgrade its systems and capabilities with cutting-edge AI tech.

Anthropic signed a $200 million contract with the Pentagon in July and was the first AI lab to deploy its technology across the agency’s classified networks.

But as the company began negotiating Claude’s deployment on the DOD’s GenAI.mil AI platform in September, talks stalled over how the military could use the models.

The department has insisted on unfettered access to the company’s technology for all lawful purposes. 

During the hearing on Tuesday, Lin questioned if the DOD was punishing Anthropic for “acting stubbornly” in negotiations. The government’s lawyer Eric Hamilton said that the company was going beyond the normal scope of a contractor.

“Anthropic is not just acting stubbornly. It’s not just refusing to agree to contracting terms. Instead, it’s raising concerns to [DOD] about how [DOD] uses its technology in military missions,” Hamilton said. “What happens if anthropic installs a kill switch or functionality that changes how it functions? That is an unacceptable risk to [DOD].”

In February after Anthropic and the DOD failed to reach an agreement, Trump issued a Truth Social post ordering federal agencies to “immediately cease” all use of Anthropic’s technology.

“WE will decide the fate of our Country — NOT some out-of-control, Radical Left AI company run by people who have no idea what the real World is all about,” Trump wrote.

WATCH: Anthropic sues Trump administration over Pentagon blacklisting

Amazon Web Services said it was once again facing service disruptions in Bahrain on Monday, as a result of the ongoing conflict ‌in the Middle East.

“We are working closely with local authorities and prioritizing the safety of our personnel throughout our recovery efforts,” a spokesperson said in a statement shared with CNBC. 

AWS advised customers to migrate their applications to alternate AWS Regions, and said it had already helped a large number of users to do so. 

It comes after the cloud provider reported service disruption related to the Iran conflict in Bahrain and the UAE earlier in March.

In the UAE, two AWS facilities were directly struck by drones. In Bahrain, a drone strike landed in close proximity to company facilities and caused physical damage.

These previous AWS disruptions caused reported outages of apps and digital services in the UAE.

In recent weeks, Iran has continued to launch missile and drone strikes on its Middle East neighbors as part of its retaliation against Israel and the U.S.

The Women’s National Basketball Player’s Association ratified the terms of a new collective bargaining agreement Monday, calling it “transformational” and “bigger than basketball.”

The new CBA begins this season and runs through 2032.

When asked her opinion of the most important outcome from the deal, WNBPA President Nneka Ogwumike had two words: “Bank accounts.”

“Being able to have your worth tied mostly in your salary is all that we’ve been fighting for, and it’s what we were able to achieve,” Ogwumike told CNBC Sport in an interview.

The deal increases the average player salary to $583,000 in 2026 with the potential to increase to more than $1 million by 2032. The maximum salary for players will now be $1.4 million in 2026 and could grow to more than $2.4 million by 2032, based on current WNBA financial projections.

Ogwumike acknowledged the salary increases may change players’ plans for how they spend their off-seasons.

The average WNBA salary was $120,000 in 2025, spurring many players to play abroad or in other leagues, such as 3-on-3 league Unrivaled, for extra money.

“Prioritizing where you want to play is going to look a lot different now that we’ve been able to negotiate a structure, a salary structure, that is tied to the revenue of the business,” Ogwumike said.

Several WNBA players, including five-time WNBA All-Star Napheesa Collier, have expressed a loss of confidence in WNBA Commissioner Cathy Engelbert in recent months, criticizing her empathy and communication with players. Ogwumike expressed optimism that players will be able to work in tandem with Engelbert under the new CBA structure.

“I told her that we’re standing here with you, Cathy,” Ogwumike said. “We were able to come to this deal and go through the process of this deal, however bumpy or smooth it was, we got here. It’s important for her to understand that we as players are at the table with her and all WNBA leadership to have achieved something that’s incredibly historical. So, I feel like there probably isn’t a better way to represent us settling our differences and moving forward in a league that we all care about then by signing this deal.”

— CNBC’s Jessica Golden contributed to this report.