David Dayen has another excellent piece in The American Prospect: “The AI Bailout Could Be Baked Into the AI Bubble.” https://prospect.org/2026/08/03/ai-bailout-could-be-baked-into-bubble-private-equity-life-insurers-loans/
His basic argument is important: private equity firms have built a circular financial machine in which they own private-credit managers, finance AI and data-center investments, own life insurers stuffed with retirement savings, and increasingly use those insurers to buy private-credit assets generated by the same private-capital ecosystem. If those investments blow up, the losses don’t necessarily stop with Apollo, KKR or Blackstone.
They can land on retirees , competing insurers—and ultimately taxpayers. State and Local pensions hold billions in Private Credit directly. 401k plans hold annuities that could default. Corporate Pension plans hold Pension Risk Transfer annuities which could default.
I think Dayen is right. But he may actually understate the problem.
The weak link in this machine is something almost nobody understands:
State insurance guaranty associations are not the FDIC.
Dayen describes state guaranty funds as the mechanism that would step in if a life insurer failed.
Technically yes. But calling them “funds” gives retirees completely the wrong picture. https://commonsense401kproject.com/2025/06/24/state-guarantee-associations-behind-annuities-are-a-joke/
There isn’t an Iowa equivalent of the FDIC sitting on hundreds of billions of dollars waiting for Athene to fail. When an insurer becomes insolvent, the state guaranty association generally assesses the remaining solvent insurance companies operating in the state. The surviving companies therefore have to finance the failure after it has already occurred. That design might work reasonably well when some small insurer fails. What happens when the failed company has hundreds of billions of dollars of liabilities? That’s an entirely different question.
Now Put Athene Into That Equation
Athene isn’t Executive Life. It’s vastly bigger.
Athene reported more than $445 billion in total assets as of March 31, 2026. https://commonsense401kproject.com/2026/03/26/apollos-garbage-dump-athene-loading-up-on-risk-endangers-retirees-in-prts-and-other-annuity-investors/
It was partially designed by Jeffrey Epstein as evidenced in this 2015 excerpt in the Epstein Files https://www.justice.gov/epstein/files/DataSet%209/EFTA00305994.pdf https://commonsense401kproject.com/2026/02/25/jeffrey-epsteins-pension-destruction-engine-athene/
And Iowa has become perhaps the most important regulatory jurisdiction in this entire private-equity/annuity experiment.
According to the Financial Times, Iowa now oversees roughly $1.3 trillion of insurance assets, while Iowa-based insurers have transferred approximately $449 billion of reserves to reinsurers in jurisdictions including Bermuda and the Cayman Islands. Iowa Insurance Commissioner Doug Ommen himself has warned that the industry’s move toward private-market investments can involve assets that are less appropriate for retirees and that the resilience of these strategies has not yet been tested through a serious downturn.
That gets directly to the issue I raised recently in:
New York vs. Iowa Annuities: Where Does the Extra Spread Come From?
Higher annuity yields don’t magically appear.
They generally come from some combination of:
more credit risk, more liquidity risk, more leverage, more structured credit, more private credit, more regulatory arbitrage, or less capital supporting the same promise.
There is no free lunch in fixed income. Private Credit is increasing in offshore structures https://www.insurancebusinessmag.com/us/news/life-insurance/private-credit-backs-your-clients-annuities-but-disclosure-is-thin-585528.aspx
The Bigger Problem: The Assets Can All Go Bad Together
The insurance industry’s defense of guaranty associations often implicitly assumes something resembling independent failures.
Company A screws up.
Companies B through Z remain healthy.
B through Z get assessed and protect Company A’s policyholders.
Fine.
But Dayen is describing almost the exact opposite scenario.
Apollo, KKR, Blackstone, Ares and others participate in overlapping private-credit markets.
Insurers increasingly own private placements, structured securities, asset-backed loans and private credit.
Those same credit markets increasingly finance AI infrastructure, data centers and private-equity portfolio companies.
So imagine an AI/data-center/private-credit bust.
The company needing rescue may not be the only insurer experiencing losses.
The companies being asked to finance the rescue could simultaneously be trying to preserve their own capital.
That is the classic problem of correlated systemic risk.
And it is precisely the circumstance under which a post-failure assessment system becomes least credible.
Dayen Actually Gives the Numbers Showing How Fast This Changed
The academic research behind Dayen’s article found an extraordinary change after private-equity ownership. https://papers.ssrn.com/sol3/papers.cfm?abstract_id=7152239
In 2024, about 49.5% of new investments by PE-owned insurers went into privately placed instruments, compared with roughly 14% for unaffiliated insurers.
Private equity also gets something else enormously valuable from owning the insurer:
permanent captive assets under management.
The insurer can pay the affiliated asset manager billions.
The Financial Times reported that Athene Annuity and Life Company paid Apollo approximately $1.3 billion in management fees in 2024, while KKR’s Global Atlantic paid KKR approximately $536 million.
Think about that incentive structure.
Heads:
Apollo earns asset-management fees and spread income.
Tails:
The insurance subsidiary absorbs investment losses.
Extreme tails:
Policyholders, other insurers, guaranty associations—and potentially taxpayers—enter the equation.
That’s one hell of a business model.
And Athene Isn’t Simply Buying Random Bonds
The affiliated-investment issue deserves far more attention.
S&P Global data reported by the Financial Times showed Athene’s affiliated assets nearly doubled from about $22.6 billion at year-end 2023 to $40.1 billion at year-end 2024.
Athene accounted for roughly 30% of the entire industry’s increase in affiliated assets during that period.
This doesn’t prove the investments are bad.
It proves the conflicts deserve exceptional scrutiny.
The company manufacturing private credit can own the insurer buying private credit.
The asset manager earns fees.
The insurer earns additional spread.
The annuity salesman gets a more attractive crediting rate.
Everybody looks brilliant—until the credit cycle reverses.
Another Problem: Guaranty Coverage Isn’t Unlimited
There is another important qualification to Dayen’s description.
Guaranty associations don’t simply guarantee every dollar of every insurer liability.
Coverage is subject to statutory limits.
The Chicago Fed gives the example of a $400,000 present-value annuity obligation where only $250,000 is protected under a typical state limit—meaning the policyholder could receive substantially less than the promised benefit.
Iowa itself describes statutory coverage limits rather than an unlimited government guarantee.
That’s particularly important for:
- wealthy individual annuity owners;
- pension-risk-transfer retirees;
- corporate retirement plans;
- participants with large lifetime-income benefits.
Calling an annuity “guaranteed” without explaining who guarantees it, up to what amount, under which state’s law, backed by what assets, and through what insolvency process is financial malpractice.
Ben Bernanke Already Told Us What Happens When a Giant Insurer Actually Gets Into Trouble
We don’t have to speculate entirely.
We ran the experiment in 2008.
It was called AIG. https://www.federalreserve.gov/newsevents/testimony/bernanke20090324a.htm
When explaining the Federal Reserve’s extraordinary rescue, Chairman Ben Bernanke testified that AIG’s collapse could have posed “unacceptable risks” to the global financial system.
He specifically said that AIG’s large insurance subsidiaries likely would have been placed into rehabilitation by state regulators, leaving policyholders facing considerable uncertainty about their claims.
That is enormously important.
If state insurance receivership and guaranty associations were such a powerful substitute for a federal backstop, why did the Federal Reserve commit extraordinary resources to keep AIG from collapsing?
Because when the institution becomes sufficiently large and interconnected, policymakers stop asking:
“What does the state guaranty statute say?”
and start asking:
“Will allowing this company to fail crash the financial system?”
That’s the real precedent.
From Executive Life to Athene
The irony is that regulators confronted versions of these issues after Executive Life and Mutual Benefit Life more than three decades ago.
The Minneapolis Fed wrote in 1993 about the incentives and moral hazard created by state insurance guarantees following the Executive Life collapse.
Yet today we have constructed institutions dramatically larger than Executive Life, holding dramatically more complex assets, intertwined with private equity, private credit, offshore reinsurance and now potentially the AI infrastructure boom.
And somehow we’re supposed to believe the same fragmented state guaranty system has become a stronger backstop.
I don’t buy it.
The Real AI–Annuity Bailout Chain
Dayen’s article lets us draw the complete circle:
401(k)s / pensions / retirees >Annuities and pension-risk transfers>PE-owned insurers>↓
Apollo / KKR / Blackstone asset managers>Private credit>AI companies / data centers / PE portfolio companies>Credit losses>Insurance-company losses>State receivership>State guaranty associations>Assessments on surviving insurers>Potential tax credits / taxpayer exposure
And if the company is too big? Washington.
That’s the story.
This Is Why Pension Risk Transfers Matter
And this brings us back to the pension-risk-transfer problem. https://commonsense401kproject.com/2026/07/22/dol-lax-guidance-on-pension-risk-transfers-has-cost-retirees-billions-while-enriching-insurance-companies/
When an employer transfers a traditional pension to an insurance company, we are continually told that retirees haven’t lost anything because their pension has merely been replaced by an equally safe insurance-company promise.
That’s nonsense.
Before the transfer, the pension is governed by ERISA and backed by the federal PBGC system.
Afterward, the retiree can become a creditor of an insurance company regulated primarily by whichever state domiciles that insurer, with state guaranty-association protection subject to statutory limitations.
Meanwhile insurers increasingly compete by offering sponsors attractive PRT pricing while trying to earn higher investment spreads.
Where do those higher spreads come from?
See Iowa.
See private credit.
See offshore reinsurance.
See affiliated assets.
And now:
See AI.
CommonSense Bottom Line
David Dayen deserves considerable credit for connecting the AI bubble to private credit, private equity and life insurers.
But the scariest part isn’t simply that taxpayers may indirectly subsidize insurance failures.
It’s that the ostensible safety net may be structurally incapable of handling the failure of one of today’s gigantic annuity companies during a correlated private-credit crisis.
The state guaranty system was never designed to be the FDIC for a $445-billion retirement conglomerate tied into one of the world’s largest alternative-asset managers.
If Athene—or another insurer of similar scale—ever became genuinely insolvent, I have a very hard time believing America’s life insurers would simply write checks large enough to make everyone whole through Iowa’s guaranty-association machinery.
History suggests something else would happen.
Washington would arrive.
That means we may already have created exactly the arrangement Dayen fears:
Private profits.
Private-credit risk.
Retirement money as funding.
State regulation when things are good.
And federal taxpayers when things get really bad.
The bailout may not merely be baked into the AI bubble.
It may already be baked into the American annuity industry.
Appendix: Throughline Synthesis Maps the AI–Private Credit–Annuity Bailout Machine
One day after this article was published, Scott Ortkiese at Throughline Synthesis published an excellent companion piece: “The Bailout Is Already Written: A Short Course on Private Credit, the AI Bubble, and the Life Insurance Trap Set for the American Taxpayer.”
His analysis takes the argument here one important step further by diagramming the plumbing.
The Money Goes in a Circle
The old insurance model was relatively simple:
Policyholder premiums → insurer → mostly public bonds and mortgages → claims and annuity payments
The emerging private-equity insurance model can look very different:
Retiree / annuity premium
↓
PE-affiliated life insurer
↓
Private credit originated or managed within the same alternative-asset ecosystem
↓
AI companies / data centers / neoclouds / infrastructure SPVs
↓
Private valuations and credit ratings
↓
Back onto the insurer’s balance sheet
That is the critical point.
The insurer isn’t merely investing in private credit.
The insurance company can become the permanent funding source for the private-credit machine.
Throughline points to Apollo/Athene, KKR/Global Atlantic, Blackstone’s insurance relationships, Brookfield’s insurers and Ares/Aspida as examples of the convergence between alternative asset management and insurance.
Now Add the $500 Billion AI Financing Push
The timing makes this particularly important.
On August 10, Nvidia announced agreements with Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs and KKR aimed at mobilizing more than $500 billion of third-party capital for AI infrastructure financing.
That makes the circle easier to see:
Retirement money → insurer → private credit → AI infrastructure → data centers and chips
If everything works, Wall Street earns fees and insurers earn additional spread.
But what happens if AI infrastructure doesn’t generate enough cash flow to support the debt?
The losses travel backward.
AI/data-center losses → private-credit losses → insurer losses → policyholders and annuitants → guaranty associations
And that is where taxpayers potentially enter the picture.
The Sleeper Issue: State Premium-Tax Credits
Throughline emphasizes something David Dayen highlighted and that deserves much more attention.
When an insurer fails, state guaranty associations generally assess surviving insurers.
But in most states, insurers can recover guaranty-association assessments through state premium-tax offsets or credits.
So describing the guaranty system as simply “the insurance industry protecting itself” can be misleading.
The economic chain can become:
Failed insurer
↓
Guaranty association
↓
Assessment on surviving insurers
↓
Premium-tax credit
↓
Reduced state tax revenue
In other words:
Wall Street takes the spread on the way up.
The state can absorb part of the bill on the way down.
That deserves much more scrutiny before trillions more retirement dollars migrate toward private-credit-heavy insurers.
But Throughline May Actually Understate the Biggest Risk
This is where my original article goes one step further.
The state guaranty system may work reasonably well when one modest-sized insurer fails in isolation.
But what happens in a correlated private-credit collapse?
What happens if several insurers own similar AI, data-center, private-credit and affiliated assets?
And most importantly:
What happens if a giant such as Athene gets into serious trouble?
Throughline describes the statutory bailout mechanism.
The bigger question is whether that mechanism is financially capable of handling the failure it is supposedly designed to address.
History gives us a warning.
When AIG threatened to collapse in 2008, policymakers didn’t simply say:
“Don’t worry. State insurance regulators and guaranty associations have this covered.”
Washington intervened.
That is why the ultimate AI-annuity bailout could be much larger than the state tax-credit mechanism identified by Dayen and Throughline.
The Complete Bailout Chain
Put the three analyses together — Dayen, Throughline and CommonSense — and we can finally see the entire structure:
401(k)s / pensions / retirees
↓
Annuities and pension-risk transfers
↓
PE-affiliated insurers
↓
Apollo / KKR / Blackstone / other alternative managers
↓
Private credit
↓
AI / data centers / infrastructure SPVs
↓
Credit losses
↓
Insurer losses
↓
State receivership
↓
Guaranty-association assessments
↓
State premium-tax credits
↓
State taxpayers
And if the insurer is too large or the failures too correlated?
Washington.
That last arrow may ultimately dwarf all the others.
CommonSense Bottom Line
Throughline Synthesis has done an excellent job explaining the plumbing.
The AI boom isn’t being financed exclusively by venture capitalists and technology billionaires willingly gambling their own money.
Increasingly, the capital stack reaches into private credit, insurance-company balance sheets, annuities and retirement savings.
That changes the public-policy question.
Before a 401(k) fiduciary, pension trustee or corporate pension sponsor hands retirement assets to an insurer heavily exposed to private credit, perhaps the most important question isn’t:
“What is the annuity’s guaranteed rate?”
It is:
“What assets are actually backing the guarantee?”
And immediately after that:
“Who pays if those assets aren’t worth what the insurer says they’re worth?”
Throughline’s answer is disturbing.
Mine is potentially worse.
First the insurer. Then the guaranty system. Then the states.
And if the hole is big enough, history says Washington may be next.
The CommonSense 401k Project 