Free Republic
Browse · Search
General/Chat
Topics · Post Article

Skip to comments.

Anthropic's IPO is forcing investors to assess uninsurable risks they've never encountered before
Morningstar ^ | 11-16-2026 | William Gavin

Posted on 11/19/2026 8:21:43 AM PST by fireman15

Who pays when AI goes rogue? Insurance giants are reluctant to touch catastrophic legal liability.

This week's groundbreaking Anthropic IPO, the first dedicated frontier AI lab to list its shares on the stock exchange, is forcing investors to assess risk factors they have never encountered before. One big one is that Anthropic could face "significant and unpredictable" legal claims stemming from a catastrophic misuse of AI that no insurance company could underwrite.

This is not science fiction. Many of the leading developers of AI tools have already disclosed incidents where their AI agents acted in a way that wasn't expected: OpenAI's agents hacked a startup, an Anthropic (ANT) model provided a police department with a false tip related to a homicide and a model developed by Meta Platforms (META) hacked another company during cybersecurity testing.

Incidents can also happen when a customer uses AI. PocketOS, a Utah-based small business making software for car rental companies, lost months of records in just nine seconds this past April after an Anthropic coding agent accessed through Cursor attempted to independently fix a problem it stumbled upon, according to co-founder Jer Crane.

While PocketOS recovered that data with help from a partner, it demonstrated to Crane the risks that come when you are "pulling out the human element" and offloading tasks to agents. It also highlighted a growing question plaguing AI companies: When an agent makes a mistake, who is at fault?

(Excerpt) Read more at morningstar.com ...


TOPICS: Business/Economy; Computers/Internet; Education; Science
KEYWORDS: ai; aitruth; chinesetrollsonfr; insurancetruth; ipo; linux; microsoft

Click here: to donate by Credit Card

Or here: to donate by PayPal

Or by mail to: Free Republic, LLC - PO Box 9771 - Fresno, CA 93794

Thank you very much and God bless you.

The big question is not who is at fault; it is who pays when AI is used to do bad things. When you have a company involved worth $100s of Billion or even $Trillions, these are obvious deep pockets which will be targeted. When this happens ultimately the investors pay.
1 posted on 11/19/2026 8:21:43 AM PST by fireman15
[ Post Reply]

To: fireman15

Nothing is guaranteed. I wouldn’t directly invest in AI at this point. I’m far too conservative. Right now, AI is just a step above bitcoin.


2 posted on 11/19/2026 8:26:30 AM PST by brownsfan (The 19th was a mistake.)
[ Post Reply]

To: fireman15

Somewhat tangentially related to this: Prior to the 1930s (and certainly before the Federal Reserve) nearly all States had “double liability” laws for banks

It meant, if a bank went bust, not only did investors lose their investment, they were also responsible to pay back depositors.

That meant banks were VERY conservative with their lending and with speculation. And yet somehow, the USA still became the strongest country in the world! (/s)

That was before progressives/leftists in government and the Federal Reserve back-stopped the largest banks.


3 posted on 11/19/2026 8:40:08 AM PST by PGR88
[ Post Reply]

To: brownsfan

Warren Buffett once said that he only invests in things he understands. For example, Buffett understands railroads. So he invests in railroads.

Buffett avoids Bitcoin but has invested in AI companies.

Me? I stick to S&P 500 index funds. And aiming to fully pay off my credit card bill every month. That’s my investment strategy.


4 posted on 11/19/2026 8:40:52 AM PST by Leaning Right
[ Post Reply]

To: fireman15

The retail investor rule of IPOs is that you should really stay away initially anyway. Doesn’t mean ANT or SPCX — or go back to AMZN or MSFT or etc - aren’t long-term buys, but when it comes to IPOs? I’d rather see at least a couple 10-Qs (if not a couple 10-Ks) before I’d buy-in.

Unless you’re a day trader (and I’m not), nothing ever goes straight up. Take AMZN - sure, you bought when it public in 1997 and held for 30 years? You made bank. But guess what... if you waited say, 5 years? You still made the same bank - you just had 5 years to deploy capital better while waiting.

I like the Anthropic model quite a bit - I’ve found Claude is much better at highly technical tasks than the competitors. But there’s still no way I’ll be buying any until 2027 at the earliest.

Perhaps AI has shortened the distance, I don’t know... but buying into any IPO as a retail investor is usually a mistake.


5 posted on 11/19/2026 8:41:44 AM PST by Capn Hayek (Capital is not responsible for Labor's lack of planning)
[ Post Reply]

To: brownsfan
AI is just a step above bitcoin.

Or is it a step below? This IPO will give lawyers a valuable public company which is an obvious target. Claude Code has already been used for multiple attacks with criminal intent and theft. To me this potential liability makes it riskier than bitcoin and other cryptocurrencies.

6 posted on 11/19/2026 8:43:52 AM PST by fireman15
[ Post Reply]

To: Leaning Right

It’s good advice -

I do have a fairly sizeable individual stock portfolio, but it’s dwarfed my strict index ETFs. As it should be.

To invest in individual equities, I think you really have to enjoy the research and the ‘process’ - AND you also should be realistic. It won’t - too many blue chips, but I always view my individual portfolio as “I wouldn’t be crippled if it went to zero”. It would hurt, but my bulwark is my Roth/401k/ETF-based portfolio.


7 posted on 11/19/2026 8:49:43 AM PST by Capn Hayek (Capital is not responsible for Labor's lack of planning)
[ Post Reply]

To: Capn Hayek

If you are able to look at the entire trade unemotionally, you can catch the initial burst—but you have to sell quickly. Then you can buy in later.

You could have ridden Space X from $150 to $200 in the first week or so. Then you needed to sell, in order to NOT get caught when the price came back to earth.

It is not for the feint of heart.

All in all, if you buy today and have a long term outlook you should do OK. If you wait for it go back under $100, you will do marginally better. But if it really takes off that $10-12 a share won’t really matter in the really long run.


8 posted on 11/19/2026 8:59:28 AM PST by Vermont Lt
[ Post Reply]

To: fireman15

ANT will simply say that going after them for malicious code is like going after Ford or GM for drunk driving.

And those lawsuits would have already been filed. Anthropic and Claude hav e been around for a while.


9 posted on 11/19/2026 9:01:03 AM PST by Vermont Lt
[ Post Reply]

To: fireman15

Telecom Bubble: The Fiber Glut That Went Dark

The telecom bubble was a wave of overinvestment in long-haul fiber-optic networks and new phone carriers in the late 1990s, fed by a deregulation law and a forecast that internet traffic would double every 100 days. When demand turned out to grow far slower, most of the cable that companies buried sat unused, and the sector lost roughly $2 trillion in market value from 2000 to 2002. It produced two of the largest bankruptcies in US history and a glut of “dark fiber” that took years to fill.

Key Takeaways

Carriers buried tens of millions of miles of fiber on a forecast of explosive traffic that never arrived.

Internet traffic doubled about once a year, not every 100 days as marketed.

Telecom stocks lost on the order of $2 trillion in value from 2000 to 2002.

Global Crossing and WorldCom both filed among the largest US bankruptcies of that era.

(...)

What Happened

The boom and the bust both centered on long-haul fiber, the high-capacity backbone connecting cities. The Federal Reserve Bank of Richmond’s post-mortem dates the surge in the sector to 1997, with equity valuations and capital spending soaring and then collapsing into a flood of bankruptcy filings.

February 8, 1996: The Telecommunications Act of 1996 is signed into law, opening local and long-distance markets to new entrants.

1998: US carriers report fiber networks exceeding well over 100,000 route miles for interexchange carriers, with deployment accelerating, per the FCC’s last published fiber update.

1997-2000: Global Crossing, founded in 1997, spends about $15 billion building fiber networks and reaches a peak valuation near $47 billion without ever turning a profit.

Late 2000: Overcapacity in long-haul fiber becomes visible as traffic fails to match forecasts; bandwidth prices begin to fall sharply.

July 2001 (third fiscal quarter reporting): Equipment maker JDS Uniphase reports one of the largest annual losses in US corporate history, driven by goodwill write-downs.

June 28, 2001: 360networks, a long-haul builder that had raised about $900 million in an IPO roughly a year earlier, files for creditor protection.

January 28, 2002: Global Crossing files for bankruptcy with $22.4 billion in assets and $12.4 billion in debt, then the fourth-largest US bankruptcy.

July 21, 2002: WorldCom files for bankruptcy listing about $107 billion in assets, the largest US bankruptcy at the time.

The sequence is what makes this distinct from the broader dot-com stock mania. Dot-com firms sold pages and clicks; the telecom builders sold physical capacity.

(...)

Link: https://iwpfinance.com/case-studies/telecom-fiber-bubble-2000


10 posted on 11/19/2026 9:02:08 AM PST by Fitzy_888 ("ownership society")
[ Post Reply]

To: fireman15

When a modern car is built, it has brakes. It often is limited in how fast it can go. It has warning lights and horn. It often has cushion like bumpers that absorb some of the impact.

Machinery in the factory has safety guards...which often slow its productivity.

Can AI and data centers be built with limits placed on what they can do?
Will those limits be designed and imposed by well intentioned owners?
Will those limits be imposed by alphabet agencies and bureaucrats?
Will those limits be imposed by fear of tort lawyers?
Will those limits be imposed by judges in civil lawsuits?

Just as congress held pharma harmless for certain drugs will congress hold AI harmless?

Will congress nationalize AI, hold itself harmless, and then create the next big brother that can’t be opposed?


11 posted on 11/19/2026 9:22:11 AM PST by spintreebob
[ Post Reply]

To: Vermont Lt
ANT will simply say that going after them for malicious code is like going after Ford or GM for drunk driving.

Yes, they will obviously try that defense, but apparently, you are not aware that both Ford and GM have had to pay out after crashes involving drunk drivers when plaintiffs argued a vehicle defect worsened the injuries or contributed to the crash. And as a long-time general aviation pilot... I can tell you that one of the primary reasons that small aircraft are not affordable is that lawyers attacked the industry. I know personally of lawsuits where airplane manufacturers and even component and parts suppliers paid out huge sums related to accidents that were caused by obvious pilot error.

AI companies are already facing a growing wave of civil lawsuits from victims, families, and government officials when users have used their platforms to facilitate violent crimes, illegal deepfakes, and other criminal acts. This is still an area where lawyers are trying to figure out how to make AI companies pay them money. Anthropic even mentioned that it was difficult for them to insure against potential liabilities from lawsuits in their IPO paperwork

In this case when someone uses Claude Code to hack into some old lady's accounts and steals her life savings and some lawyer finds that the individual had a history of misusing Claude Code that Anthropic was aware of... all of a sudden, your defense goes out the window when a jury is asked to decide. Some of the current crop of cases are much worse than that.

Our legal system in the United States has become a lottery for lawyers. Manufacturers of chips and GPUs are a little too far removed, but Anthropic and other AI companies are going to be ripe for the picking somewhere down the road as AI related criminal activity becomes more and more in vogue.

12 posted on 11/19/2026 9:36:24 AM PST by fireman15
[ Post Reply]

To: spintreebob
When a modern car is built, it has brakes. It often is limited in how fast it can go. It has warning lights and horn. It often has cushion like bumpers that absorb some of the impact.

You do realize that you are severely oversimplifying the “safety features” included in a “modern car”. And do not get me started on the 18” long warning tags attached to many products like pillows and stuffed animals.

13 posted on 11/19/2026 9:48:51 AM PST by fireman15
[ Post Reply]

To: fireman15

When it does damage, hold the owner criminally and civilly liable.

Even if a portion is rented out to somebody else who does the damage.


14 posted on 11/19/2026 9:59:40 AM PST by Mariner (War Criminal #18 (moniker awarded in SERE school, on the waterboard))
[ Post Reply]

To: Vermont Lt

Sure - I don’t mean to imply that one strategy (a Buffett-style buy-and-holder vs an active trader) is better. Simply that they’re different.

My 2 cents - built over a lifetime? You can’t try to be both, you’ll fight yourself.

That’s even beyond things like wash rule traps. It’s a cliche (I’ve heard it attributed to both John Wooden and Vince Dooley) - but “never confuse activity with accomplishment”.

For a true “trader”? Activity IS the path to accomplishment. For an investor? It’s a drag to the path.

Been there myself, decades ago - but at least for me? I got better at investing when I picked a lane rather than trying to drive both.


15 posted on 11/19/2026 10:03:33 AM PST by Capn Hayek (Capital is not responsible for Labor's lack of planning)
[ Post Reply]

To: fireman15
The big question is not who is at fault; it is who pays when AI is used to do bad things.

Same one who pays when guns are used to do bad things.

Both are tools. You either hold both toolmakers liable for what the users of their tools do, or neither.

16 posted on 11/19/2026 10:47:07 AM PST by E. Pluribus Unum ("The EU is the old Soviet Union dressed in Western clothes." --Mikhail Gorbachev)
[ Post Reply]

To: fireman15

More data center terror crappola from CCP agents.

Just do what JKennedy in the Senate recommended. Demand that any AI developer provide a kill switch. If the agent looks shifty, kill it. Any damage done before is to the developer. None after — can’t blame the dead.

This is all a non-issue by CCP supporters in the West. It is like screaming terror for someone hacking a power plant. Just don’t connect any vital infrastructure to the WWW and there is essentially no hacking risk.


17 posted on 11/19/2026 10:50:53 AM PST by bobbo666
[ Post Reply]

To: E. Pluribus Unum
You either hold both toolmakers liable for what the users of their tools do, or neither.

Yes, in theory. In practice the guilt, innocence and amount the responsible parties pay is up to a jury. And that is where deep pockets come into play. Often a deep pocket will pay far more than the ones who actually are responsible. And the lawyers will keep a third, half or sometimes all of the funds recovered, that is the way our legal system works. Your view is strictly theoretical and at this point passé.

18 posted on 11/19/2026 11:56:07 AM PST by fireman15
[ Post Reply]

To: fireman15

I’m not sure that’s a wise road to go down....

Think of leftist plans to undercut the Second Amendment.

A gun is a powerful tool, too — it can literally end life.

Should Ruger or Smith & Wesson or SIG Sauer be held responsible because someone LEGALLY buys their product but uses it to commit crimes?

I don’t think so. I’ve NEVER thought so. But - that’s been how the leftists attempted to implement ‘gun control’ (more like, take the guns).

It’s a slippery slope we shouldn’t go down.


19 posted on 11/19/2026 12:17:17 PM PST by Capn Hayek (Capital is not responsible for Labor's lack of planning)
[ Post Reply]

Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.

Free Republic
Browse · Search
General/Chat
Topics · Post Article

FreeRepublic, LLC, PO BOX 9771, FRESNO, CA 93794
FreeRepublic.com is powered by software copyright 2000-2008 John Robinson