Defense Secretary Pete Hegseth’s decision to label Anthropic a “Supply-Chain Risk to National Security” on Friday resulted in more questions than answers.

“It’s all very puzzling,” Herbert Lin, a senior research scholar at Stanford University’s Center for International Security and Cooperation, told CNBC in an interview.

Anthropic is the only American company ever to be publicly named a supply chain risk, as the designation has traditionally been used against foreign adversaries. But the company hasn’t received any official declaration beyond social media posts.

A formal designation will require defense vendors and contractors to certify that they don’t use Anthropic’s models in their work with the Pentagon.

The dispute centered around how Anthropic’s artificial intelligence models could be used by the military. The Department of Defense wanted Anthropic to grant the agency unfettered access to its Claude models across all lawful purposes, while Anthropic wanted assurance that its technology would not be tapped for fully autonomous weapons or domestic mass surveillance.

With no agreement reached by Friday’s deadline, President Donald Trump directed federal agencies to “immediately cease” all use of Anthropic’s technology, and said there would be a six-month phaseout period for agencies like the DOD.

Experts told CNBC the supply chain risk designation is highly unusual, especially since the U.S. and Israel began carrying out strikes in Iran just hours later. A group of retired defense officials, policy leaders and executives wrote to Congress on Thursday, defending Anthropic and calling the Trump administration’s designation a “dangerous precedent.”

Anthropic’s models are still being used to support U.S. military operations in Iran, even after the company was blacklisted, as CNBC previously reported.

Talks between Anthropic and the DOD are now reportedly back on, according to the Financial Times, but there are still big questions hanging over the issue as of Thursday.

Why is the U.S. government still using Claude?

Stanford’s Lin doesn’t understand why the DOD is still using Anthropic’s models in sensitive settings if they pose such a threat. If the Trump administration really sees Anthropic as a risk to national security, he said, it wouldn’t make sense to phase out the models over an extended period of time.

“OK, wait a minute, they’re a really dangerous player for U.S. national security, so you’re going to use them for another six months? Huh?” Lin said. 

Michael Horowitz, a senior fellow for technology and innovation at the Council on Foreign Relations, said it’s “especially notable” that Anthropic’s models were used to support the U.S. military action in Iran. He said “there’s no clearer signal” of how much the Pentagon values the technology.

“Even in a situation where there is this intense feud between the company and the Pentagon, they are using their technology in the most important military operation that the United States is conducting,” he said. 

Transitioning away from Anthropic toward a new vendor takes time and comes at a significant cost in terms of efficiency, said Jacquelyn Schneider, a Hargrove Hoover fellow at Stanford University’s Hoover Institution.

Until recently, Anthropic was the only AI company approved to deploy its models across the agency’s classified networks. OpenAI and Elon Musk’s xAI received clearance, but their systems can’t be deployed or adopted overnight.

What’s the actual threat?

By designating Anthropic a supply chain risk, the DOD is suggesting that the company is “really bad” for U.S. national security, Lin said. But he stressed that the agency hasn’t clearly outlined what kind of threat the company poses. 

“They don’t point to any technical failing, they don’t point to any hack,” Lin said. “They say things like ‘They’re arrogant,’ and ‘We don’t want you telling the DoD what to do in some hypothetical situation that hasn’t happened yet.'”

Lin said the other punishment that Hegseth was threatening to impose on Anthropic, invoking the Defense Production Act, also contradicts the idea that the company threatens national security. 

The Defense Production Act allows the president to control domestic industries under emergency authority when it’s in the interest of national security. It could essentially compel Anthropic to let the Pentagon use its technology. 

Horowitz said he thinks the clash between Anthropic and the DOD is “masquerading” as a policy dispute. 

Months earlier, venture capitalist and White House AI and crypto czar David Sacks criticized the company for “running a sophisticated regulatory capture strategy based on fear-mongering,” after an essay published by an executive, and conservatives have repeatedly accused Anthropic of pushing “woke AI.”

Anthropic CEO Dario Amodei took a different approach than other tech executives, avoiding getting cozy with the Trump administration in its early days.

“This feels to me like a dispute that is about politics and personalities,” Horowitz said. 

Is an official designation on the way?

Anthropic hasn’t been designated a supply chain risk by any official measure, and there’s an open question as to if or when the company should expect one. Defense contractors have to decide whether they will follow Hegseth’s directive on social media or wait for more formal guidance. 

Several executives told CNBC that their companies are moving away from Anthropic’s models, and a venture capitalist said a number of portfolio companies are switching “out of an abundance of caution.” But others, including C3 AI Chairman Tom Siebel, said he doesn’t see a “need to mitigate” the technology “until it gets litigated.” 

Schneider said businesses are rational, and if they think it’s high risk to work with Anthropic, whether it’s formally declared a supply chain risk or not, they’re going to hedge and look for other partners.

“There’s all sorts of decisions that have been made within the Trump administration that, by law, require more codification,” Schneider said. “Even the example of moving from DoD to [Department of War]. That by law needs more codification, but all the contractors are using DoW.”

Even so, Samir Jain, vice president of policy at the Center for Democracy and Technology, said social media posts likely aren’t enough to actually cause a designation.

“There’s a process that the statute requires, including an actual finding that Anthropic presents national security risks if it’s part of the supply chain,” he said in an interview. “I don’t think, factually, that that predicate could possibly be met here.”

Anthropic said in a statement Friday that it will challenge “any supply chain risk designation in court.”

Does this have anything to do with the U.S. strikes on Iran?

For Schneider, the war in Iran now looms large over the spat between Anthropic and the DOD. She said she’s left wondering whether the two conflicts were happening in parallel, or if they were somehow related. 

“Obviously, you’re not going to walk away from technologies that are deeply embedded in your wartime processes right before you go to war,” Schneider said.

She said planning a military operation of that magnitude would have required “a lot of sleepless nights,” so she was surprised the DOD was willing to spend such a “remarkable amount of energy” on a public clash ahead of the initial attack.

What happens next?

As the war in Iran stretches into its sixth day, Anthropic’s path forward with the DOD remains a big mystery.  

Horowitz said he would bet that the six-month off-boarding period will become a “a locus for some re-examination” within the Pentagon, especially since members of Congress and broader public markets have shown so much interest in the dispute. 

Lin expressed a similar sentiment, and said he wouldn’t bet on Anthropic’s models being out of the DOD a year from now.

Schneider is less convinced. 

“I wish I had a more definitive thought about where this is all going to go, but everything is so unprecedented,” she said. When it comes to historical examples or analogous cases, Schneider said: “I don’t have those. It’s just super limited.”

The DOD declined to comment. Anthropic didn’t provide a comment.

WATCH: Anthropic tops $19 billion in annual revenue rate

In the aftermath of the Israeli-U.S. strikes on Iran, American allies in the Persian Gulf are hearing a sound that Ukrainian soldiers have long come to dread: the foreboding hum of the Shahed-136 ‘kamikaze’ drone. 

Originating from Iran, the Shahed has already become a fixture of modern warfare, with Tehran’s strategic partner, Russia, utilizing the technology in its years-long invasion of Ukraine. 

Now, the drones — the most advanced of which is the long-ranged Shahed-136 — have become central to Iran’s retaliation strategy against the U.S. and its regional allies, with thousands unleashed so far. 

At first glance, the Shahed is unremarkable compared with cutting-edge weapon technologies, with one analyst referring to it as “the poor man’s cruise missile.” 

But while American allies have managed to intercept the vast majority of incoming drones with the help of U.S.-provided defense systems such as the ‘Patriot’ missiles, many Shaheds still managed to hit their targets. 

The United Arab Emirates Ministry of Defence said on Tuesday that out of 941 Iranian drones detected since the start of the Iran war, 65 fell within its territory, damaging ports, airports, hotels and data centers.

Analysts say the key to their effectiveness lies in the numbers. The drones are relatively cheap and easy to mass-produce, especially compared to the sophisticated systems used to defend against them. 

Those factors make the drone ideal for swarming and overburdening aerial defenses, with each drone intercepted also representing a more valuable defense asset expended. 

“The Shahed‑136, among other unmanned aerial systems, has allowed states like Russia and Iran a cheap way to impose disproportionate costs,” said Patrycja Bazylczyk, analyst with the Missile Defense Project at the Center for Strategic and International Studies in Washington DC.

“They force adversaries to waste expensive interceptors on low‑cost drones, project power, and create a steady psychological burden on civilian populations.” 

U.S. government reports describe the Shahed-136 as a one-way attack unmanned aerial vehicle produced by Iranian entities tied to the Islamic Revolutionary Guard Corps.

Compared with ballistic missiles, the drones fly low and slow, deliver a relatively modest payload, and are limited to mostly fixed targets, said Behnam Ben Taleblu, senior director of the Iran program at the Foundation for Defense of Democracies, told CNBC.

Public estimates suggest Shahed drones can cost between $20,000 and $50,000 apiece. Ballistic and cruise missiles, by contrast, can cost millions of dollars each.

In that sense, the Shahed and its equivalents “basically serve as ‘the poor man’s cruise missile’ offering a way to strike and harass adversaries “on the cheap,” said Taleblu.

For Iran, which faces both international sanctions and limitations on acquiring advanced weapons, that cost advantage is significant.

Meanwhile, air defense systems used by Gulf states and Israel can cost between $3 million and $12 million per interceptor, according to U.S. Department of Defense budget documents.

This cost discrepancy raises a serious issue for Iran’s enemies: Air defense systems have finite numbers of defense missiles, with each target intercepted representing a valuable asset expended.

Thus, in a war of attrition, the drones could be used by Tehran to wear down air defenses, opening them up to more damaging attacks, analysts say.

“The logic is to expend drones early while preserving ballistic missiles for the long haul,” said CSIS’s Bazylczyk.

She added that Iran’s ability to sustain mass‑drone use will depend on its stockpiles, how well it can protect or restore its supply chain, and whether the U.S. and Israel can meaningfully disrupt the flow of components or production sites. 

The U.S. has long sought to disrupt Iran’s production of the Shahed-136, and recently imposed new sanctions targeting suspected component suppliers across Turkey and the UAE.

However, Russia’s production of Shahed drones shows that such systems can be manufactured at scale during wartime and amid targeted sanctions. 

U.S. officials claim Iran had launched over 2,000 drones in the conflict as of Wednesday. However, the country is understood to have large stockpiles and may be capable of producing hundreds more each week, military experts reportedly told The National newspaper.

“Gulf countries are at risk of depleting their interceptors unless they are more prudent about when it fires those interceptors,” said Joze Pelayo, a Middle East security analyst with the think tank Atlantic Council.

“The depletion is not imminent, but it remains an urgent issue,” he said, noting. However, attacks on multiple fronts by Iranian allies such as Hezbollah and the Houthis could put stockpiles at risk of being depleted within days, he added.

Some analysts have suggested that Iran has drawn from Russia’s extensive battlefield experience with the drones, including modifications such as anti-jamming antennas, electronic warfare-resistant navigation, and new warheads.

Those warheads typically carry 30 kg to 50 kg of explosives and can pack a punch, particularly when used in large swarms, with advanced versions capable of a range of up to 1,200 miles.

Michael Connell, a Middle East specialist at the Center for Naval Analyses, said that the Shahed-136 has proven so effective that the U.S. has reverse-engineered it and deployed its own version on the battlefield against Iranian targets. 

In its Iran attacks over the weekend, the U.S. Central Command confirmed that it had used such low-cost one-way attack drones modeled after the Shahed for the first time in combat. 

With drones becoming a fixture of the modern battlefield, methods for dealing with them are also evolving.

According to Taleblu from the Foundation for Defense of Democracies, Ukraine has found some success in downing drones with fighter jet cannon fire, a more sustainable deterrent than missile interceptors.

Ukraine also recently pioneered the development of cheaper mass-produced interceptors, which Kyiv claims can stop the Shahed.

Gulf states are also expected to adopt more sustainable approaches. The Pentagon and at least one Gulf government are reportedly in talks to buy the cheaper Ukrainian-made interceptors.

Meanwhile, Qatar’s Ministry of Defense says it is also using its air force jets to intercept Iranian attacks, including Shahed drones, alongside ground-based air defenses.

Electronic warfare targeting the Shahed’s GPS, as well as short-range missiles and directed-energy systems such as Israel’s Iron Beam, are also significantly cheaper to operate than traditional interceptors.

Still, analysts say Gulf states currently lack fast, high-volume anti-drone capabilities. Developing and deploying such systems will likely take years, said Atlantic Council’s Pelayo.

“Countries in the Gulf hosting U.S. bases, such as Bahrain, Kuwait, and the UAE, benefit from an extended ability to repel drone attacks through the American-operated system, but it is still not enough against mass and sustained attacks.”

Anthropic CEO Dario Amodei is back at the negotiating table with the U.S. Department of Defense after the breakdown of talks on Friday over the use of the company’s AI tools by the military, according to The Financial Times. 

Amodei is in talks with Emil Michael, under-secretary of defense for research and engineering, in a last-ditch effort to reach an agreement on the terms governing the Pentagon’s access to Anthropic’s Claude models, the Times reported, citing anonymous sources with knowledge of the matter.

Discussions fell apart Friday, with President Donald Trump directing federal agencies to stop using Anthropic’s tools, and Defense Secretary Pete Hegseth saying he would designate the company a supply-chain risk to national security.

Last week, Michael had attacked Amodei, calling him a “liar” with a “God complex,” in an X post.

Agreeing to a new contract would enable the U.S. military to continue using Anthropic’s technology, which has reportedly been utilized in Washington’s war with Iran. 

Claude became the first major model deployed in the government’s classified networks through a $200 million contract awarded by the DoD to Anthropic, but the company later sought guarantees that its tools would not be used in domestic surveillance or autonomous weapons. The Pentagon had demanded that the military be allowed to employ the technology for any lawful use.

In a Friday memo seen by FT, Amodei reportedly told staff that near the end of negotiations with the Defense Department, it had offered to accept Anthropic’s terms if they deleted a “specific phrase about ‘analysis of bulk acquired data'” — a line he said, “exactly matched this scenario we were most worried about.” 

Amodei also wrote in his note that messaging from the Pentagon and OpenAI, which struck a new deal with the Defense Department on Friday, was “just straight up lies about these issues or tries to confuse them.” 

The timing of OpenAI’s deal with the Pentagon, announced within hours of the White House decrying Anthropic, had caused a public backlash, with Claude seeing a surge of app downloads while ChatGPT reportedly saw app uninstallations surge.

OpenAI CEO Sam Altman later said that his company “shouldn’t have rushed” its deal and outlined revisions to its own safeguards with how the Defense Department can use its technology. 

In a post on X, Altman further addressed the controversy, saying: “In my conversations over the weekend, I reiterated that Anthropic should not be designated as a [supply chain risk], and that we hope the [Department of Defense] offers them the same terms we’ve agreed to.”

Anthropic was founded in 2021 by a group of former OpenAI staff and researchers, who left the firm after disagreements over its direction, with the company marketing itself as a “safety-first” alternative.

Government officials have for months criticized Anthropic for allegedly being overly concerned with AI safety.

A tech industry group, whose members include Nvidia, Google and Anthropic, had sent a letter to Hegseth on Wednesday expressing concern over his designating a U.S. company as a supply-chain risk.

The Defense Department and Anthropic did not immediately respond to a CNBC request for comment on their reported negotiations.

Broadcom CEO Hock Tan sees the artificial intelligence boom gaining so much steam that he’s projecting AI chip revenue next year “significantly in excess of $100 billion.”

After the chipmaker reported better-than-expected results for the fiscal first quarter and issued a strong forecast for the current period, Tan said on his company’s earnings call that demand is picking up from large customers that are increasingly in need of Broadcom’s help in designing custom silicon.

“We have also secured the supply chain required to achieve this,” Tan said, regarding the 2027 sales target.

AI revenue in the first quarter more than doubled from a year earlier to $8.4 billion, while total sales increased 29% to $19.3 billion. The company expects AI semiconductor revenue of $10.2 billion this quarter.

Broadcom shares popped more than 5% in extended trading on Thursday after Tan’s comments.

Chip companies like Broadcom have faced a number of headwinds in recent months, including a shortage of the high bandwidth memory crucial for custom accelerators, and capacity constraints at the most advanced levels of chip manufacturing and packaging.

Broadcom helps its customers translate their chip designs into silicon, providing back-end support before the processors are sent off to be manufactured at huge fabrication plants by companies like Taiwan Semiconductor Manufacturing Company.

It’s a role that’s fueled Broadcom’s growth as more tech giants design in-house accelerators for AI. Tan said custom AI deployment is entering its “next phase” and is expected to speed up, as the company helps six key customers design their chips. Chief among those are Google, Meta, Anthropic and OpenAI, with Fujitsu and ByteDance likely as the final two.

Google was the first to the in-house chip game in 2015, with its tensor processing units designed alongside Broadcom. Google has made its chips available to cloud customers since 2018, with key customers now including Apple and Anthropic. Broadcom expects even stronger demand from next-generation Google chips in 2027.

Meta is also reportedly in talks to use Google’s TPUs, and Broadcom assists the social media company with developing its own MTIA accelerator. Analysts have cast doubt on the future of Meta’s custom silicon program, but the “MTIA roadmap is alive and well,” Tan said on the earnings call.

During the question and answer portion of the call, Bernstein Research analyst Stacy Rasgon pushed Tan on the specific sources of the projected $100 billion in AI chip revenue. He counted 3 gigawatts of capacity at Anthropic, 3 gigawatts at Google, at least 2 gigawatts with Meta, and 1 gigawatt from OpenAI, among others. Tan said the dollars per gigawatt “vary, sometimes quite dramatically,” but that his estimates were “not far” off.

While Tan said that the AI revenue boost would come from “just chips,” Broadcom makes much more than just AI accelerators. Ben Bajarin of Creative Strategies said it includes digital signal processors, data processing units and networking switches.

It’s “everything in that bucket,” Bajarin said.

— CNBC’s Jordan Novet contributed to this report.

WATCH: Broadcom CEO on Ai revenue

Amazon‘s data center in Bahrain was targeted by Iran’s Islamic Revolutionary Guard Corps for the company’s support of the U.S. military, Iranian state media said Wednesday.

The company’s cloud computing unit said Monday that one of its facilities in Bahrain was damaged due to a nearby drone strike on Sunday. Two data centers in the United Arab Emirates were also damaged after they were “directly struck” by drones.

All of the facilities remain offline, according to the Amazon Web Services health dashboard.

The attack in Bahrain was launched “to identify the role of these centers in supporting the enemy’s military and intelligence activities,” Iran’s Fars News Agency said on Telegram.

The incidents came after joint U.S.-Israel strikes on Iran over the weekend. Iran has retaliated against Israeli and U.S. bases across the Gulf.

Amazon declined to comment.

In addition to structural damage, the data centers also experienced power disruptions and some water damage after firefighters worked to put out sparks and fire. Some popular AWS applications experienced “elevated error rates and degraded availability” due to the incident.

AWS advised cloud customers to back up their data, consider migrating their workloads to other regions and direct traffic away from Bahrain and the UAE.

AWS announced its Bahrain region in 2019, and it hosts significant workloads for governments there. The company also operates a corporate office in Bahrain that is primarily for AWS employees.

Earlier this week, Amazon instructed all of its corporate employees in the Middle East to work remotely and “follow local government guidelines” amid escalating instability in the region.

Broadcom reported better-than-expected earnings and revenue and issued a strong forecast for the current period as the chipmaker continues to benefit from the artificial intelligence boom.

Here’s how the company performed in comparison with LSEG consensus:

• Earnings per share: $2.05 adjusted vs. $2.03 estimated
• Revenue: $19.31 billion vs. $19.18 billion estimated

Revenue jumped 29% year over year during the fiscal first quarter, which ended on Feb. 1, according to a statement.

Net income increased to $7.35 billion, or $1.50 per share, from $5.50 billion, or $1.14 per share, in the same quarter a year earlier. Adjusted earnings exclude stock-based compensation and tax adjustments.

For the second quarter, Broadcom said it anticipates a 68% adjusted profit margin, higher than StreetAccount’s 66% consensus. The company said it’s looking for $22 billion in revenue, beating the $20.56 billion average estimate, according to LSEG.

Broadcom helps other companies translate their chip designs into silicon, providing intellectual property and backend technologies before they’re sent off to chip fabrication plants from companies such as Taiwan Semiconductor Manufacturing Company. It’s a role that’s gained importance as Amazon, Google, Meta and Microsoft design customized chips.

AI revenue soared 106% from a year earlier to $8.4 billion, “driven by robust demand for custom AI accelerators and AI networking,” CEO Hock Tan said in the statement.

Tan had called for a doubling of AI revenue in December. He said the company expects AI semiconductor revenue of $10.7 billion this period.

Broadcom reported $12.52 billion in revenue from semiconductor solutions, higher than the $12.25 billion that analysts polled by StreetAccount expected. During the quarter, Broadcom announced new Wi-Fi 8 chips.

For infrastructure software, Broadcom said it generated $6.80 billion in revenue, lower than StreetAccount’s $7.02 billion consensus.

Broadcom said its board authorized up to $10 billion in new share buybacks through 2026.

In December Tan said Anthropic had placed a $10 billion custom chip order. Last week U.S. Defense Secretary Pete Hegseth said the Pentagon would dub Anthropic a “supply chain risk to national security” and President Donald Trump directed government agencies to stop using Anthropic after the AI startup refused to permit uses of its technology for mass domestic surveillance or fully autonomous weapons.

As of Wednesday’s close, Broadcom shares were down 8% so far in 2026, while the S&P 500 index was flat.

Executives will discuss the results on a conference call starting at 5 p.m. ET.

This is breaking news. Please check back for updates.