Banks Won’t Lend Against It, Your Index Fund Just Bought It

A Repricing Evident in Current Events

Published:

When SoftBank sought a $10 billion loan against its $60 billion stake in OpenAI, banks balked. SoftBank cut its ask to $6 billion. Banks still refused.

The same banks are now collecting underwriting fees to bring OpenAI public at a valuation that implies the stake is worth the full $60 billion.

This is not a contradiction. It is a precise description of how risk moves through the financial system at the end of a cycle: refused as unanalyzable collateral by institutions that cannot establish a liquidation price for unlisted private equity — and simultaneously distributed at full stated private-market valuation to capital pools that are large, mandatory, and slower to recognize the structural gap.

This is what it looks like when collateral that cannot be priced still finds its final buyers.

The Mechanism

Financial institutions hold two roles simultaneously in transactions like the OpenAI IPO. As lenders, they assess collateral at what they would recover in a forced sale — in this case, refusing to advance even 10% of the stated private market valuation. As underwriters, they collect fees (typically 1.5–2% of the offering size) to distribute those same assets at par to public buyers.

The institution is not confused about the value. It has done the math twice, from two different desks, and arrived at two different answers — because the questions are different. The collateral desk asks: what can we recover if everything goes wrong? The underwriting desk asks: what can we sell it for today? In the SoftBank case, lenders did not calculate a lower number for the OpenAI stake. They concluded the asset was unanalyzable as collateral — a private company with no liquid market, no consensus valuation, and no price discovery mechanism a risk committee could defend. The same asset is simultaneously being prepared for public distribution — OpenAI filed a confidential S-1 on June 8; the SEC review is ongoing — at an $852 billion implied valuation anchored by its most recent private funding rounds, with no public price discovery yet established. The gap between those two outcomes is not a difference of opinion. It is a structural feature of what private valuations are and are not.

The gap is where risk travels. It moves from sellers who know the liquidation price to buyers whose reporting cycles, redemption constraints, and index mandates prevent them from acting on that information in real time.

The buyers are, in the aggregate, your retirement account.

Three Places This Is Happening Right Now

1. The IPO Queue

SpaceX completed its IPO on June 12, 2026, at a $1.77 trillion valuation — the largest in history. Anthropic filed a confidential S-1 with the SEC on June 1. OpenAI filed on June 8. Alphabet announced $80 billion in equity offerings — the largest equity capital transaction in history and its first major equity raise in two decades. Oracle, already public, launched an at-the-market equity program to fund its cloud infrastructure build-out.

J.P. Morgan, using Dealogic data, projects over $260 billion in total equity issuance arriving in 2026 — Goldman Sachs estimates $160 billion in IPO proceeds alone, which would approach the 2021 record. Historical annual US IPO and follow-on issuance capacity runs $50–150 billion outside cycle peaks; 2021 absorbed $275 billion at the prior cycle’s height, and 2026 is on track to match or exceed it.

The institutional counterargument deserves engagement: J.P. Morgan notes that $1.5 trillion in corporate buybacks expected this year could offset the equity supply, and the S&P 500’s $65 trillion market cap provides a deeper absorption pool than existed in 2021. This is a genuine demand offset, and it applies at the portfolio level.

What it does not address is the mechanism inside passive funds. When SpaceX entered the Russell 1000 and CRSP indexes on June 18, an estimated $10–16 billion in mandatory passive purchases were executed automatically. The funds buying SpaceX sold other holdings to fund the purchase. Buybacks operate on a different timeline, through different capital pools, and do not prevent this rebalancing from occurring inside index funds when new entrants arrive. The sellers — IPO investors exiting at peak private valuations — receive the exit liquidity. The buyers — index fund participants who had no choice in the matter — absorb the position.

This mechanism will repeat for each subsequent entrant. J.P. Morgan itself acknowledges that “existing cohorts of the S&P 500 — and likely the largest weightings — could see temporary selling pressure, driven by index rebalancing.” The aggregate market may absorb the supply. The individual index fund participant does not get to vote on whether they receive the allocation.

2. Private Credit Gates

Three major private credit vehicles have restricted investor withdrawals in 2026, in sequence.

Blue Owl permanently eliminated quarterly redemption windows for its $1.4 billion debt fund in February 2026 — a structural change, not a temporary measure. BlackRock’s private credit fund triggered its 5% withdrawal gate in March 2026. Blackstone’s BCRED invoked its 5% standard gate in Q2 2026 — the first time it had ever done so.

Reuters confirmed that pension funds and insurance companies are direct investors in Blue Owl’s gated fund.

When a fund gates redemptions, the positions inside it continue to be reported at book value — the price the assets were assigned at purchase, or the last internal valuation. The gate prevents investors from finding out what those assets would actually fetch in a sale. The positions become, in financial terms, zombies: still carried at full value on institutional balance sheets, impossible to exit.

For pension funds, the consequence arrives at the next annual actuarial cycle, when the shortfall between asset value and liability value is formally recognized. That recognition is scheduled for 2027 reporting, covering 2026 market conditions. By then, reallocation is no longer an option.

3. PE-Owned Life Insurance

The third channel is the least visible and the most structurally novel.

Over the past decade, major private equity firms — Apollo, KKR, Carlyle, Blackstone — acquired life insurance companies and use the policyholder premiums as permanent capital. They invest that float into their own deal flow: CLO tranches, private credit, structured products. They earn management fees on the insurance float and origination fees on the same paper the insurance company buys. The concentration of these assets in structured credit — ABS and CLOs — runs at roughly 50% of PE-owned insurer portfolios versus 25–33% for traditional insurers, per PineBridge analysis of NAIC statutory filings. That premium yield advantage allows PE-backed insurers to offer better annuity rates and attract more policyholder premiums, compounding the concentration.

The leverage gap — the difference between what the insurance company owes policyholders and what the assets would actually cover in a stress scenario — is managed offshore through captive reinsurance structures: the US insurer transfers liabilities to a captive entity, books reserve credit, and plugs the shortfall with contingent instruments. Regulators in competing captive-friendly states — Vermont, Iowa, Delaware, South Carolina, among others — have limited capacity to scrutinize this uniformly without losing business to competing domiciles. Iowa made the dynamic explicit: in May 2026, the state signed legislation establishing a new regulatory framework for life captive reinsurance companies specifically to “keep Iowa competitive with other captive insurance regulatory leaders” — effective July 1, 2026.

This mechanism moved from theoretical to confirmed on December 31, 2025, when PHL Variable Insurance — owned by a PE-backed holding company — pivoted toward formal liquidation proceedings with a confirmed $2.2 billion shortfall, two offshore captive reinsurers, and a gap that directly triggered state guaranty caps. Policyholders above the $300,000 state guarantee limit were placed under moratorium — meaning their access to funds exceeding that cap was suspended, pending liquidation proceedings. A live case, not a projected scenario.

The reason this channel has no market pricing is structural: insurance analysts at sell-side firms cover insurance companies; private equity analysts cover PE managers. No analyst routinely does both. The solvency risk of a PE-owned insurer, built on a PE-managed CLO portfolio, with PE-structured captive reinsurance, sits precisely in the gap between two analyst communities that don’t talk to each other.

The Common Thread

Each of these mechanisms shares a single feature: risk moves from capital with full information to capital with constrained mobility and delayed reporting. The risk profiles differ in kind — index inclusion delivers valuation and concentration exposure in liquid securities; private credit gating delivers structural lock-up in marked-to-model positions; PE-owned insurance delivers solvency exposure that can result in direct capital destruction — but the direction of transfer is the same in each case.

The banks assessing SoftBank’s OpenAI stake for collateral could not price it and therefore would not lend against it — not even at 10 cents on the dollar. The retirement account holders receiving that exposure through the IPO allocation do not have that constraint, and their index mandate means they cannot decline it.

The PE manager gating a private credit fund knows what the underlying assets would fetch in a forced sale. The pension LP cannot exit, cannot force a mark-to-market, and will not see the recognized shortfall until an actuary files a report in early 2027.

The PE firm that owns both the insurance company and the CLO paper knows the concentration and the captive structure. The policyholder knows their yield and their statutory guarantee amount. PHL closed that information gap for its policyholders on December 31, 2025.

The institutions managing these structures are not engaged in fraud. They are doing what their incentive structures reward: capturing fees, transferring duration and credit risk to less mobile counterparties, and operating within regulatory frameworks that were not designed for the complexity of the instruments involved. The states competing to register captive reinsurers cannot tighten unilaterally without losing business to the others. The analyst coverage gap is not coordinated; it is structural. These are systems producing outcomes no single actor intended.

Why Now

The current moment is notable because the channels are operating concurrently, not sequentially, and they share a common capital receptacle.

The $260+ billion equity issuance wave, the gated private credit funds, and the PE insurance structures land on three distinct but related balance sheets: 401(k) participants absorbing late-stage tech equity through mandatory index rebalancing; pension beneficiaries exposed to locked private credit positions their fund managers cannot exit; and life insurance policyholders whose annuity yields depended on PE-managed structured credit portfolios that were not built for the current rate environment. This is not because the system was designed to concentrate risk there. It is because those are the largest, most stable, least mobile pools of capital in the economy — which makes them the natural destination for risk that needs a long-duration home.

The convergence of recognition timelines — IPO lockup expirations in early 2027, pension actuarial cycles in 2027, PE insurance solvency stress deepening as refinancing pressure builds at current rates — means the gap between what these assets are worth and what they are carried at will close over the same window, rather than spreading out.

The SoftBank datum is useful precisely because it is so concrete. Lenders would not advance even $6 billion — 10 cents on the dollar — against SoftBank’s $60 billion stake, even after SoftBank cut its ask by 40% to meet them. The reason lenders gave was not that they had calculated a lower value: it was that they could not price unlisted private company shares as collateral at all. The asset is unanalyzable for secured lending while simultaneously being sold to public buyers at an $852 billion implied company valuation. That is not an analyst opinion or a model output. It is what two different desks determined when they attempted the same pricing exercise — one as lender, one as underwriter — and discovered the questions produce irreconcilable answers. That number is available. It is just not the number being used in the documents going to public market buyers.

This analysis describes structural patterns and their likely trajectories. It is not investment advice, financial guidance, or market timing recommendation. No content should be interpreted as suggesting specific investment actions. Readers should consult qualified financial professionals before making investment decisions.

This article draws on primary sources including Bloomberg, Reuters, SEC filings, J.P. Morgan Research, Goldman Sachs Research, NAIC filings, and PineBridge Investments analysis of NAIC statutory data. All cited events are confirmed as of June 21, 2026.

Sources

— Free to share, translate, use with attribution: D.T. Frankly (dtfrankly.com)

§