Anthropic says it will sue Pentagon over supply chain risk label

[midian182]

A hot potato: Anthropic's battle with the American government has taken a new turn. The US has now officially designed the AI company as a supply chain risk. In response, Anthropic has promised to sue the Pentagon over the decision.

The clash centers on two red lines Anthropic refused to drop during negotiations with the Department of Defense: using Claude for mass domestic surveillance of Americans and for fully autonomous weapons. Anthropic says those carveouts are narrow, reasonable, and have not affected any government mission to date. The Pentagon disagreed.

A senior Pentagon official said the supply chain risk designation was "effective immediately."

In a statement published after Defense Secretary Pete Hegseth said he was directing the department to apply the designation, Anthropic called the move legally unsound, and said it would set a dangerous precedent for any US company negotiating with the government. The firm added that it would fight any such designation in court rather than cave on the two restrictions.

"Well, I fired Anthropic. Anthropic is in trouble because I fired [them] like dogs, because they shouldn't have done that," President Trump said in a new interview with Politico.

Anthropic doesn't have a history of being anti-defense. The company says it has supported US government work on classified networks since June 2024 and remains committed to national security projects. But it seems there are limits it won't cross.

Anthropic CEO Dario Amodei wrote that the practical impact of the designation is narrower than officials have implied. He said the relevant law would only allow the restriction to apply to Claude's use in Defense Department contract work, not to every commercial relationship involving contractors that also happen to do business with the military. Essentially, Anthropic believes the Pentagon is trying to stretch the law well beyond its intended scope.

The Information Technology Industry Council, whose members include several of the biggest names in US tech, warned that using emergency supply-chain powers in a procurement dispute could inject uncertainty into the market and make it harder for the government to access best-in-class tools and services.

This week, Anthropic delivered a master class in arrogance and betrayal as well as a textbook case of how not to do business with the United States Government or the Pentagon.

Our position has never wavered and will never waver: the Department of War must have full, unrestricted…

– Secretary of War Pete Hegseth (@SecWar) February 27, 2026

Anthropic says this is the first time such a measure has been publicly aimed at an American company rather than a foreign adversary or similarly high-risk entity.

"From the very beginning, this has been about one fundamental principle: the military being able to use technology for all lawful purposes," a Pentagon official said.

"The military will not allow a vendor to insert itself into the chain of command by restricting the lawful use of a critical capability and put our warfighters at risk."

Also read: AI-driven warfare is here, and the Iran strikes show how fast it's advancing

Anthropic had been in talks with the DoD recently, but the conversations came to nothing. According to a person familiar with the matter, the failure was partly due to how President Donald Trump and other members of his administration publicly berated the company.

Amodei is one of the few big tech names not to have made large donations to Trump or publicly praise him, which some believe has contributed to the current situation.

Microsoft has confirmed that it will continue working with Anthropic on non-defense-related projects. The Redmond giant will continue to embed Anthropic tech in products for clients – except for the DoD, of course.

Permalink to story:

Anthropic says it will sue Pentagon over supply chain risk label

 

[Beerfloat]

I doubt that the idea of setting a 'dangerous precedent' will sway our tweeter in chief. Firstly because he doesn't know the word and secondly because it's basically his entire modus operandi.

 

[gdavid65]

Private corp fights US Govco arguing Skynet is a bad idea... US Govco overrules..

How ****ing stupid are we?

 

[Endymio]

I doubt that the idea of setting a 'dangerous precedent' will sway our tweeter in chief.

Probably because no such "dangerous precedent" exists. If you wish to sell products to the federal government, you must allow them to be used for all lawful purposes. Imagine if Lockheed tried to tell the DoD that it must first check with them before using any of their F-35s on a mission, or DuPont telling the military that its kevlar vests could only be used in certain situations.

I realize you'll remain unswayed by facts or logic, but this has very real implications, as the recent Venezuelan raid (in which the DoD used Anthropic's products) to plan. Logistically, you have hundreds of actions -- many of which involve striking various targets -- which must occur in precise sequence. Some are targets of opportunity, and many have windows lasting only hours, minutes, or even seconds. You feed all this into a model, and it sets an optimum ordering and timing. And -- most importantly -- if the factors change during the operation, the model adjusts in real time.

Anthropic bans this as 'autonomous control' of weapon systems. Their right -- but it makes it useless for military purposes. If you need a human to second-guess every model result, what use is the model? And the possibility of Anthropic adding new, additional restrictions makes DoD dependence on their products even riskier.

Private corp fights US Govco arguing Skynet is a bad idea...

You realize "Skynet" is a fictional construct, right? There are only two options here. Either (a) we integrate AI into our military forces, or (b) we wait until China -- or some other adversary -- AI-powered military moots the point by conquering us. The DoD is estimating that, by 2040, AI will be a force multiplier larger than that from our current crop of smart weapons.

 

[Anton Longshot]

Curious to see where this will go, seeing that the current administration doesn't GAF about laws and such.
Of course Anthropic's right and Trump shouldn't be allowed to abuse his power for personal feuding.
But if the court's in Trump's pocket, oops and even if that's not the case Trump may justl ignore the verdict.
BTW the USA will soon run out of capital letters thanks to Trump who uses them up like there is no tomorrow. But then, for him there really ìs no tomorrow it seems.

 

[ScottSoapbox]

Do everything we want or we will label you a terrorist is a terrible precedent (but the logical progression) from our government.

 

[mbk34]

Logistically, you have hundreds of actions -- many of which involve striking various targets -- which must occur in precise sequence. Some are targets of opportunity, and many have windows lasting only hours, minutes, or even seconds. You feed all this into a model, and it sets an optimum ordering and timing. And -- most importantly -- if the factors change during the operation, the model adjusts in real time.

I don't see why you need specialised AI to model this, it seems like something a fairly simple program could do. You just apply scores to each target, the costs for using various weapons and the risks such as losing pilots or civilian casualties.

That doesn't get round the fact that trying to wipe out a "government" that you're conducting peace talks with is morally reprehensible. The whole attack was also illegal under international law. America is increasingly becoming more isolated and becoming better defined as a rogue state.

 

[Beerfloat]

Probably because no such "dangerous precedent" exists. If you wish to sell products to the federal government, you must allow them to be used for all lawful purposes. Imagine if Lockheed tried to tell the DoD that it must first check with them before using any of their F-35s on a mission, or DuPont telling the military that its kevlar vests could only be used in certain situations.

It's good that you include 'lawful purposes' because that is clearly not a given with this administration. Everything else you said is completely beside the point.

The contract with Anthropic was signed by the Trump administration, which accepted the restrictions. Half a year passes and one day someone decides that it is actually not okay. That can happen, of course. What would a normal administration do? Cancel the contract and/or negotiate for changes. Not Trump, though.

Trump proceeds to label Anthropic a 'supply chain risk' and decrees that not only will the Pentagon no longer do business with them, but everyone else who does business with the federal government can no longer do any business with them either. This has nothing to do with your fancy tale of reliable precision strike engineering, mind you. These other companies might be using Claude to optimize their code, or do their accounting spreadsheets, or whatever kind of menial tasks people use AI for. But it doesn't matter.

This has never been done to an American company before. It's usually solely the domain of perceived 'foreign adversaries' like the Huaweis of the world. But Trump is not a reasonable man. He sees yet another Californian company becoming a leader in its field, trounce xAI on a bunch of stuff, and yet have the audacity to hold on to just the tiniest shred of virtue. That becomes a convenient excuse for his devoted zombie army to agree that Anthropic must not just be kept out of cockpits, but utterly destroyed.

It's an over the top and outrageous response, even if you somehow do not agree with Anthropic that their 2 very simple rules might actually be reasonable (as a reminder, that's 1) Do not use our technology for domestic surveillance of Americans and 2) Do not use it for autonomous killing machines, there should always be a human making that decision).

The precedent is having a resentful and mercurial clown deciding which company can live (going by the many examples, mostly by being suitably crony capitalist and keeping the billions flowing into his family's pockets) and which company should go under. It's utter bullshit.

 

[wiyosaya]

You realize "Skynet" is a fictional construct, right? There are only two options here. Either (a) we integrate AI into our military forces, or (b) we wait until China -- or some other adversary -- AI-powered military moots the point by conquering us. The DoD is estimating that, by 2040, AI will be a force multiplier larger than that from our current crop of smart weapons.

Hmmm. Echos of Minority Report. Maybe we can prevent the future.That's the ticket.

 

[wiyosaya]

Do everything we want or we will label you a terrorist is a terrible precedent (but the logical progression) from our government.

Fortunately, for those who have decided to take matters to court, it doesn't always work.

 

[Endymio]

It's good that you include 'lawful purposes' because that is clearly not a given with this administration...

You can sell that idiocy elsewhere. The Biden and Obama Administrations both beat Trump handily for number of unconstitutional acts -- Obama in fact more than Trump, Reagan, both Bushes, and Clinton all combined.

The contract with Anthropic was signed by the Trump administration, which accepted the restrictions

Wrong again. The *first* use of Anthropic products by the DoD came through a 2024 Palantir contract, signed by the Biden Administration. Anthropic's pre-existing integration was the reason it was awarded a second contract in Jul 2025 to prototype additional capabilities ... but that contract allows the government to end usage at any time ... which it did. Nor did the contract explicitly state these restrictions; that arose only after the first actual combat mission for which Anthropic performed E2E planning.

....Trump proceeds to label Anthropic a 'supply chain risk' and decrees that not only will the Pentagon no longer do business with them, but everyone else who does business with the federal government can no longer do any business with them either.

Wrong yet again. The restriction is limited to military contractors only. And it exists for good reason. If the military depends on a product or service that itself depends on Anthropic, and Anthropic denies a lawful usage as they did here, then it's just as disruptive as if the military were using their products directly.

This has never been done to an American company before.

Right. No American company has ever yanked the rug out from under our own military in this manner. What's truly ironic is that, if Anthropic had done the same to Ukraine, crippling its war effort against Russia, you'd be shouting for their heads right now.

 

[Endymio]

I don't see why you need specialised AI to model this, it seems like something a fairly simple program could do. You just apply scores to each target, the costs for using various weapons and the risks such as losing pilots or civilian casualties.

It's considerably more complex than I laid out in my post, with sub-dependencies upon sub-sub-sub dependencies. And yes, non-ML models have assisted with this for decades. The issue is that setting up all the parameters and restrictions by hand often takes weeks or longer, by which time they've changed and changed again.

That doesn't get round the fact that trying to wipe out a "government" that you're conducting peace talks with is morally reprehensible.

Eh? The US government was not carrying out "peace talks" before we struck Iran. We attempted to convince them to give up their nuclear weapons program. They refused, and then we attacked.

The whole attack was also illegal under international law.

Not even remotely correct. The US and Israel both were responding to the threat of Iranian nuclear weapons and expanding ballistic missile capabilities. Dozens of nations and every US Administration since 1984 have designated Iran the world's preeminent state sponsor of terrorism, with 100+ attacks against US forces and US civilians in just the last 30 years alone.

 

[mbk34]

It's considerably more complex than I laid out in my post, with sub-dependencies upon sub-sub-sub dependencies. And yes, non-ML models have assisted with this for decades. The issue is that setting up all the parameters and restrictions by hand often takes weeks or longer, by which time they've changed and changed again.

So you add a hierarchical data structure to the system. Everything else remains the same. Inputs into the system would allow the position of targets to change but everything else would be fairly static. The system would just loop round calculating whether the the score for targets was enough to justify the possible costs. The advantage of a non AI system is you could then go back and see why hitting the girls school was seen as justified. With an AI system you just get a green light and no data to back it up.

Eh? The US government was not carrying out "peace talks" before we struck Iran. We attempted to convince them to give up their nuclear weapons program. They refused, and then we attacked.

My apologies, my wording was slightly wrong. You were in nuclear talks with Iran. On the 26th Feb Iran stated that significant progress had been made. On the 27th Iran said it would degrade its current stockpiles of nuclear material to “the lowest level possible” — effectively to unrefined levels. On the 28th you bombed and destroyed their government.

Not even remotely correct. The US and Israel both were responding to the threat of Iranian nuclear weapons and expanding ballistic missile capabilities. Dozens of nations and every US Administration since 1984 have designated Iran the world's preeminent state sponsor of terrorism, with 100+ attacks against US forces and US civilians in just the last 30 years alone.

Have any nuclear weapons ever been found in Iran? was there any evidence of nuclear weapons in Iran? For that matter was there any evidence of WMD in Iraq? Are you honestly trying to say that Iran is a larger state sponsor of terrorism than America or are you simply discounting anything done by the CIA because of patriotism? Just in the past month or two you guys have blown up or kidnapped the leaders of two countries! Even the guy that planned the twin towers attack was a terrorist trained by you guys.

 

[ScottSoapbox]

Fortunately, for those who have decided to take matters to court, it doesn't always work.

Court hasn't always worked. For example the US citizens that got "deported" to a prison in another country while their attorney won a judgement in court that did absolutely nothing to stop the illegal jailing or did not get them a speedy return to the US.

We are only somewhat ruled by law these days. And thus far the "outlier" "accidents" haven't resulted in much pain for the agents or agencies breaking the law so there isn't a strong incentive for the illegal actions to stop in the future.

 

[Beerfloat]

You can sell that idiocy elsewhere. The Biden and Obama Administrations both beat Trump handily for number of unconstitutional acts -- Obama in fact more than Trump, Reagan, both Bushes, and Clinton all combined.

Nice, a triple treat comment. It aims to address the 'unlawful' comment but doesn't. It is also utterly irrelevant with regards to the comment's veracity. And it's a wild claim without any evidence. You've outdone yourself here.

Wrong again. The *first* use of Anthropic products by the DoD came through a 2024 Palantir contract, signed by the Biden Administration. Anthropic's pre-existing integration was the reason it was awarded a second contract in Jul 2025 to prototype additional capabilities ... but that contract allows the government to end usage at any time ... which it did. Nor did the contract explicitly state these restrictions; that arose only after the first actual combat mission for which Anthropic performed E2E planning.

The statement was entirely correct. The Trump administration signed the contract extension on the same terms. That Anthropic had previous contracts before this one does not change that fact in any way, and you even bringing it up shows the weakness of this argument.

Wrong yet again. The restriction is limited to military contractors only. And it exists for good reason. If the military depends on a product or service that itself depends on Anthropic, and Anthropic denies a lawful usage as they did here, then it's just as disruptive as if the military were using their products directly.

It remains to be seen whether the usage was lawful, and not just because breach of contract is the opposite of that. And the hyperbole of some subcontractor not being able to use Anthropic in Excel being 'just as disruptive' as your contrived example of the Venezuelan raid would be laughable, were it not that this administration clearly did and continues to violate a number of laws in that campaign.

Right. No American company has ever yanked the rug out from under our own military in this manner. What's truly ironic is that, if Anthropic had done the same to Ukraine, crippling its war effort against Russia, you'd be shouting for their heads right now.

Nothing more than conjecture. Also, is supporting Ukraine over Russia now considered a bad thing in your circle?

 

[Anton Longshot]

So you add a hierarchical data structure to the system. Everything else remains the same. Inputs into the system would allow the position of targets to change but everything else would be fairly static. The system would just loop round calculating whether the the score for targets was enough to justify the possible costs. The advantage of a non AI system is you could then go back and see why hitting the girls school was seen as justified. With an AI system you just get a green light and no data to back it up.

My apologies, my wording was slightly wrong. You were in nuclear talks with Iran. On the 26th Feb Iran stated that significant progress had been made. On the 27th Iran said it would degrade its current stockpiles of nuclear material to “the lowest level possible” — effectively to unrefined levels. On the 28th you bombed and destroyed their government.

Have any nuclear weapons ever been found in Iran? was there any evidence of nuclear weapons in Iran? For that matter was there any evidence of WMD in Iraq? Are you honestly trying to say that Iran is a larger state sponsor of terrorism than America or are you simply discounting anything done by the CIA because of patriotism? Just in the past month or two you guys have blown up or kidnapped the leaders of two countries! Even the guy that planned the twin towers attack was a terrorist trained by you guys.

I find it interesting that you both use "we" as if you're WARFIGHTERS instead of Internet Guys.
Avoid doing the "we" thing when actual warfighters are present because they might not like it.
You may think they're part of your "we" but you are definitely not a part of their "we".
Just sayin'...

 

[mbk34]

I find it interesting that you both use "we" as if you're WARFIGHTERS instead of Internet Guys.
Avoid doing the "we" thing when actual warfighters are present because they might not like it.
You may think they're part of your "we" but you are definitely not a part of their "we".
Just sayin'...

Are you sure you're quoting the right guy? I promise you I wouldn't want to be associated with this war at all.

 

[Endymio]

It is also utterly irrelevant with regards to the comment's veracity. And it's a wild claim without any evidence.

I've posted to this forum multiple times proof of this statement. The Obama Admin was found to have acted unconstitutionally on more than 110+ occasions, including 44 for which the Supreme Court ruled unanimously 9-0 against him. The Biden Administration was well below this figure, but still vastly in excess of Trump. Nor are these facts "irrelevant" -- they demonstrate you're not concerned with legality or constitutionality ... except when you believe it advances your politics.

The statement was entirely correct. The Trump administration signed the contract extension

Wrong times two. The fixed-use contract was signed by the Biden Administration. What was signed in 2025 was not -- I repeat not -- a "contract extension". It was an entirely new contract designed to explore potential new uses for Anthropic's products. And it allows cancellation at any point in time, a fact not even Anthropic has tried to dispute.

It remains to be seen whether the usage was lawful,

No it doesn't, and wild conspiracy theories don't change that.

 

[Endymio]

My apologies, my wording was slightly wrong. You were in nuclear talks with Iran. On the 26th Feb Iran stated that significant progress had been made. On the 27th Iran said it would degrade its current stockpiles of nuclear material to “the lowest level possible” — effectively to unrefined levels.

Those talks failed. Iran refused to give up their stockpiles. They refused to surrender their enrichment capabilities, even though the US offered them free enriched nuclear fuel for commercial power reactor use forever. They refused to stop rebuilding the Arak reactor, capable of producing plutonium. And their downblending of current enriched uranium was to anywhere near "unrefined" levels (0.72%), but simply to a 3.67% level that can be easily reversed in just a few weeks time. And most importantly of all, Iran refused to stop its program of developing even longer-range ballistic missiles nor to halt its financial support of proxy terror groups throughout the Middle East.

...and then you destroyed their government.

You mean the government that slaughtered tens of thousands of its own citizens in just the most recent protests alone? The government that arrests, gang-rapes, then murders young girls for daring to appear in public without hijab? Cry me a river.

Have any nuclear weapons ever been found in Iran? was there any evidence of nuclear weapons in Iran?

Was this a joke? You don't build an weapons-grade enrichments facility hidden underneath a mountain, unless you're trying to build nuclear weapons. You don't enrich 488 kg -- enough for 20+ nuclear weapons -- unless you intend to build them.

Are you honestly trying to say that Iran is a larger state sponsor of terrorism than America

You've gone off the deep end now.

 

[user556]

George W. did indeed declare USA a terrorist state - "Fight terror with terror".

 

[Endymio]

George W. did indeed declare USA a terrorist state - "Fight terror with terror".

Why lie?

"...There is no record of President George W. Bush officially saying "we will fight terror with terror." While he pledged to use "all our resources" to "conquer this enemy" and "bring them to justice, , the phrase is often associated with criticism of his administration's "Global War on Terror" policies..."

 

[Ramzz]

Probably because no such "dangerous precedent" exists. If you wish to sell products to the federal government, you must allow them to be used for all lawful purposes. Imagine if Lockheed tried to tell the DoD that it must first check with them before using any of their F-35s on a mission, or DuPont telling the military that its kevlar vests could only be used in certain situations.

I realize you'll remain unswayed by facts or logic, but this has very real implications, as the recent Venezuelan raid (in which the DoD used Anthropic's products) to plan. Logistically, you have hundreds of actions -- many of which involve striking various targets -- which must occur in precise sequence. Some are targets of opportunity, and many have windows lasting only hours, minutes, or even seconds. You feed all this into a model, and it sets an optimum ordering and timing. And -- most importantly -- if the factors change during the operation, the model adjusts in real time.

Anthropic bans this as 'autonomous control' of weapon systems. Their right -- but it makes it useless for military purposes. If you need a human to second-guess every model result, what use is the model? And the possibility of Anthropic adding new, additional restrictions makes DoD dependence on their products even riskier.


You realize "Skynet" is a fictional construct, right? There are only two options here. Either (a) we integrate AI into our military forces, or (b) we wait until China -- or some other adversary -- AI-powered military moots the point by conquering us. The DoD is estimating that, by 2040, AI will be a force multiplier larger than that from our current crop of smart weapons.

Ok so why the need to designate a company as a supply chain risk? A company has the right to do business with their own principles. If the govt does not like that then they can just go with another vendor!

 

[Endymio]

Ok so why the need to designate a company as a supply chain risk?

Quite simply, because what a company does to the DoD, it can do to the suppliers who provide equipment, supplies, and services to the DoD.

In the case of Anthropic, this isn't some remote possibility. If Anthropic claims "moral principles" prevent it from directly supporting certain military operations, why would it not also have a problem indirectly supporting those same operations, through the civilian contract chain?

A company has the right to do business with their own principles.

Yes. So does the government. If it doesn't wish to do business with a particular vendor, it doesn't have to.

If the govt does not like that then they can just go with another vendor!

Now you're getting it. That's exactly what the federal government is doing here. Or -- to be precise -- just the US military.

 

[Puiu]

You can sell that idiocy elsewhere. The Biden and Obama Administrations both beat Trump handily for number of unconstitutional acts -- Obama in fact more than Trump, Reagan, both Bushes, and Clinton all combined.


Wrong again. The *first* use of Anthropic products by the DoD came through a 2024 Palantir contract, signed by the Biden Administration. Anthropic's pre-existing integration was the reason it was awarded a second contract in Jul 2025 to prototype additional capabilities ... but that contract allows the government to end usage at any time ... which it did. Nor did the contract explicitly state these restrictions; that arose only after the first actual combat mission for which Anthropic performed E2E planning.


Wrong yet again. The restriction is limited to military contractors only. And it exists for good reason. If the military depends on a product or service that itself depends on Anthropic, and Anthropic denies a lawful usage as they did here, then it's just as disruptive as if the military were using their products directly.


Right. No American company has ever yanked the rug out from under our own military in this manner. What's truly ironic is that, if Anthropic had done the same to Ukraine, crippling its war effort against Russia, you'd be shouting for their heads right now.

"if Anthropic had done the same to Ukraine, crippling its war effort against Russia, you'd be shouting for their heads right now" - and yet it's not what happened. talking about whereabouts is just you trying to deflect from what happened.

here's a real "whatabout" - if the democrats had done even 1% of what trump did you would be out there burning Washington DC. but somehow trump always gets a pass from you.

trump wants to be a pedophile? that's completely fine for you.
trump wants to sexually harras women? that's ok with you.
trump wants to scam ppl with a crypto scam coin? it's just "business"
trump wants to commit tax fraud? he's just smarter than the system
trump releases racist videos and comments? it's just his "personality", it's not racist when he does it
trump wants to be a fascist king? it's all in the name of the greater good, the constitution can burn.

and the list goes on and on and on and on and on and on and on and on and on and on and on and on and on.

 

[mbk34]

Those talks failed. Iran refused to give up their stockpiles. They refused to surrender their enrichment capabilities, even though the US offered them free enriched nuclear fuel for commercial power reactor use forever. They refused to stop rebuilding the Arak reactor, capable of producing plutonium. And their downblending of current enriched uranium was to anywhere near "unrefined" levels (0.72%), but simply to a 3.67% level that can be easily reversed in just a few weeks time. And most importantly of all, Iran refused to stop its program of developing even longer-range ballistic missiles nor to halt its financial support of proxy terror groups throughout the Middle East.

Perhaps it depends on which news sources you listen to? Al Jazeera will probably be biased in one direction and American news sources will be similarly biased in the other direction. I was listening to an interview with an Iranian diplomat on the Thursday evening news who was saying talks were going well and would be continuing the next week. Unfortunately you then bombed Tehran the next day which I guess stopped the talks.

I do understand that unstable countries shouldn't have nukes but Israel and America have them and, in the past year, have committed genocide in Palestine and overthrown the governments in 2 countries.

You mean the government that slaughtered tens of thousands of its own citizens in just the most recent protests alone? The government that arrests, gang-rapes, then murders young girls for daring to appear in public without hijab? Cry me a river.

I suspect the number of deaths from this new war will far outreach tens of thousands on both sides. There are abuses by governments everywhere. I understand there's been torture and sexual abuse in Guantanamo where many of the inmates haven't even been charged after 20 years.

Was this a joke? You don't build an weapons-grade enrichments facility hidden underneath a mountain, unless you're trying to build nuclear weapons. You don't enrich 488 kg -- enough for 20+ nuclear weapons -- unless you intend to build them.

The problem with not having nukes is that large countries keep invading you. Libya didn't have nukes and was invaded. North Korea has nukes but has not been invaded. Ukraine gave up it's nukes in return for protection from the America and was then invaded by Russia (and not protected by America). Perhaps Iran sees nukes as being it's only protection from you guys?

You've gone off the deep end now [regards American doing more state sponsored terrorism than Iran].

I'll give a quick list of the foreign leaderships that America has overthrown (usually by the CIA). If you want me to list all countries that America has sponsored terrorism then that's a much longer list and will take me more time to compile:
Cuba
Columbia
Italy
Nicaragua
Kosovo
Syria
Libya
Iran
Iraq
Afghanistan
Panama
Grenada
Dominican Republic
Venezuela
Guatemala
Chile
Brazil
Democratic Republic of Congo
South Vietnam
Haiti
Hawaii

There's probably more but I got bored doing the research.