BlackRock is shopping all of BlackRock TCP Capital's remaining loans. The Street thinks the last sale cleared at 95 cents. It cleared at 90. And the reason a "5% discount" destroyed 10.4% of NAV is the single most important, least-discussed fact in private credit today.
On August 24, Bloomberg reported that BlackRock, through KBW, is pitching all $671 million of loans held by BlackRock TCP Capital to Ares Management and other credit shops. That number is not a coincidence, and almost nobody has connected it: $671 million is the exact pro-forma size of the entire remaining company as disclosed by TCPC's own president on the August 6 earnings call. This is not a portfolio trim. It is a shop of the whole balance sheet — 18 days after the CEO told an analyst the board had no "specific agenda" for its strategic review. At $4.15, TCPC trades at 0.70× pro-forma NAV of ~$5.90, which means the market is pricing the residual book at roughly 72–78 cents on management's marks. An arm's-length institutional buyer with full diligence access — Pantheon — just paid ~90 cents for the bigger, more concentrated half of the same portfolio three weeks ago. My call: buy TCPC $3.95–$4.30, target $5.00 on a deal announcement (1–3 months) and $5.10–$5.60 on full realization (6–12 months), invalidated if Q3 NAV prints below $5.35 on November 5. HIGH CONVICTION, SMALL SIZE (1.0–1.5%). Expected total return, probability-weighted including distributions: +25%.
I have written twice this month — on August 12 and August 21 — that the Pantheon transaction proved private credit marks were "good to 95 cents." That was wrong, and it was wrong in the direction that flattered my own thesis. The real print was about 90 cents.
The 95% figure is the headline price in the August 6 press release: 95% of December 31, 2025 gross fair value. But in the Q&A portion of the Q2 call, KBW's Paul Johnson asked CFO Erik Cuellar to reconcile a 5% discount with a 10.4% NAV hit. Cuellar's answer, verbatim in substance: start with the stated 5% portfolio discount, add "other customized adjustments that are done in these type of transactions," and you get "a rough effective discount of about 10%"; transaction expenses take it "up to about 10.4% of a NAV hit."
Cross-check the arithmetic: the NAV hit was $0.68 per share × 83,902,775 shares = ~$57 million on $523 million of assets sold. That is 10.9% all-in. The disclosed discount and the disclosed NAV damage only reconcile at ~90 cents, not 95.
Consequence: the observable clearing price for a diligenced, first-lien-heavy US direct lending book in August 2026 is 88–92 cents on the manager's own marks — right in line with where LP-led private credit secondaries have been pricing all year (~90% of NAV). That is the number every BDC's discount should be measured against. Not par. Not 95. Ninety.
Here is where the real analytical edge is, and it required reading Schedule III of Exhibit 10.5 in the June 30 10-Q — work first surfaced publicly by Leyla Kunimoto at Accredited Investor Insights on August 8, and which I have verified against the filing index and the earnings call.
TCPC did not sell a pile of loans. It sold the equity of a levered securitization. The sequence:
Do the equity math: pool $523M − CLO debt $406M + cash collateral $74M ≈ $191M of TCPC equity claim, roughly 36% of the gross pool. A 10% haircut on a pool that is only 36% equity-funded is a ~28% haircut on the equity. That is the entire explanation for how a "5% discount" headline turned into a 10.4% NAV amputation. Leverage magnification, disclosed only if you read the exhibits.
The market is extrapolating the −10.4% NAV hit onto the second lot. The mechanism that produced it has been extinguished. The continuation vehicle took all the CLO debt with it. The residual $671M book sits at pro-forma net leverage of ~0.4×, dropping below 0.3× once Domo repays — that is ~71% equity-funded, versus 36% for the piece that was sold.
| Per 10 points of haircut on assets | Lot 1 (sold to Pantheon) | Lot 2 (residual $671M) |
|---|---|---|
| Equity funding of the pool | ~36% | ~71% |
| Implied leverage multiplier | ~2.75× | ~1.35× |
| Hit to shareholder equity | ~28% | ~13.5% |
The second sale is roughly half as damaging per point of discount as the first one was. Investors who watched NAV fall 10.4% on a 5% headline are now demanding a 22–29% discount to underwrite the sequel. That is the mispricing.
All figures from the August 6 press release, 10-Q, and earnings call unless noted.
| NAV / share | $6.58 |
| Prior quarter / year-end 2025 | $6.72 / $7.07 |
| Net assets | $552.0M |
| Shares out | 83,902,775 |
| Portfolio FV / companies | $1.29B / 134 |
| First lien | 89.8% |
| Senior secured | 91.5% |
| Floating rate | 93.9% |
| Non-accrual (FV / cost) | 1.6% / 7.4% |
| Net leverage | 1.38× |
| Wtd avg cost of debt | 6.03% |
| NII / share | $0.22 ($0.21 adj.) |
| Total investment income | $40.0M |
| PIK share of income | 7.6% |
| Portfolio FV | $671M |
| Portfolio companies | 132 |
| Average position size | $5.1M |
| Net leverage | ~0.4× (<0.3× post-Domo) |
| Liquidity | ~$395M |
| Unfunded commitments | <$40M (was $90M) |
| NAV / share (my estimate) | ~$5.90 |
| Net assets (my estimate) | ~$495M |
| Price / pro-forma NAV | 0.70× |
| Software exposure | ~23% (~17% post-Domo) |
| Post-Q2 repayments received | $97.4M |
| Domo repayment expected Q4 | $69M at par |
Two facts about the residual book that the 22% implied haircut cannot survive contact with:
| Calculation | Value |
|---|---|
| Market cap at $4.15 | $348M |
| Pro-forma equity (my estimate) | $495M |
| Gap | $147M |
| Implied haircut on full $671M book | 21.9% |
| …excluding Domo at par ($602M) | 24.4% |
| …excluding Domo + post-Q2 repayments ($505M) | 29.1% |
| Actual arm's-length print, Aug 4, on the bigger half | ~10% |
The market is asking for a 12–19 point better price than Pantheon got, on a book that is 89.8% first lien with 1.6% non-accruals at fair value. That is either an extraordinary opportunity or a statement that management's marks are fraudulent. The DOJ is asking the second question. I'll deal with it in §8.
Let me be the one to say the unpopular thing plainly, because the sell-side won't: at $4.15, TCPC yields 16.4% on the declared $0.17 quarterly dividend, and that dividend is unpayable on the pro-forma balance sheet.
My model, built from the Q2 disclosures:
| Pro-forma quarterly P&L (my estimate) | $M | $/sh |
|---|---|---|
| Investment income: $671M × 10.5% total portfolio yield ÷ 4 | 17.6 | 0.21 |
| Interest on ~$150M cash at ~4% | 1.5 | 0.02 |
| Total income | ~19.1 | 0.23 |
| Interest expense (net debt ~$198M at 6.03%, gross-debt drag) | (4.0) | (0.05) |
| Base management fee on the smaller asset base | (2.2) | (0.03) |
| Other operating expenses | (2.6) | (0.03) |
| Net investment income | ~10.3 | ~0.12 |
| Declared dividend | 14.3 | 0.17 |
| Coverage | ~70% | |
Reported Q2 investment income was $40.0M. You cannot shrink the earning asset base by 48% and hold NII flat. The Q3 dividend of $0.17 is already declared (payable September 30, record September 16). The decision point is the Q3 print, estimated November 5. I expect a cut to $0.10–$0.12.
This is why I am revising my own invalidation level. On August 21 I set a stop on "a dividend cut below $0.12/quarter." That was analytically lazy: I am now modeling a cut to $0.10–$0.12 as the base case. A stop that triggers on your own base case is not a stop, it's a tax. The dividend cut is removed as an invalidation condition. NAV per share is the invalidation now, because NAV is the asset the thesis is actually long. Expect a 8–15% single-day drawdown on the cut headline and treat $3.60–$3.85 as an add zone, not a stop-out.
TCPC re-approved a $50 million repurchase authorization on April 29, 2026, running through April 30, 2027. In the six months to June 30 it bought 661,803 shares for $2.871 million at an average of $4.34 — and in Q2 specifically, just 156,370 shares at $3.78.
Consider what they declined to do. At $4.15 against ~$5.90 of pro-forma NAV, buying back stock is the highest-return investment available to this company by an enormous margin. Deploying the full $50M would retire ~12.2M shares and take pro-forma NAV from $5.90 to ~$6.20 — a 5.1% instant, riskless accretion. They have ~$395M of pro-forma liquidity. They used 5.7% of a buyback authorization while sitting on eight times the authorization in cash.
There are two explanations. Either management does not believe its own NAV — or the company has been inside a live M&A process and legally unable to buy stock. Note the insider record: in the first half of March 2026, four insiders bought roughly 97,000 shares on the open market at $3.65–$3.89 (August Worrell 80,000 shares across two prints, Patrick Wolfe ~8,900, Jason Mehring 6,500, CFO Erik Cuellar 1,750). Not one open-market purchase since. Nobody has sold. Insiders who were buying stopped buying — and then the whole book showed up in a Bloomberg story. My read: blackout, not disbelief. The Bloomberg leak is the confirmation.
Follow BlackRock's own economics, because that is what determines whether this ends in a sale or a stall.
Conclusion: BlackRock is being paid ~$8–10M a year to attach the name "BlackRock" to a federal valuation probe. There is no version of that trade a $11.5T asset manager wants to keep. The strategic review's disclosed options included "completing an orderly realization of portfolio assets." Nineteen days later, KBW is walking the entire book around to Ares. The board is not deciding whether to sell. It is negotiating the price. That is worth far more than a 30% discount to NAV.
I am not going to strawman this. There are four real objections.
The strongest evidence for this is the Renovo episode documented by Unicus Research: HomeRenew Buyer was carried at par at the end of September 2025, filed Chapter 7 on November 3, 2025, and was marked to zero weeks later. NAV was declared at $8.71 on November 6, 2025 alongside a $0.25 dividend; eleven weeks later NAV was $7.07 and the dividend was cut 32% to $0.17. NAV is down 53% since December 2021. That is a real and damning track record, and it is why the DOJ is involved.
My counter: the Pantheon transaction is precisely the test of that objection, and the marks passed it within ten points. A sophisticated secondaries buyer with full diligence access, a Lincoln International fairness opinion in the file, and Moelis on the other side of the table paid ~90 cents for 78 of the 132 names — and per the transaction structure, TCPC transferred roughly two-thirds of each position in substantially all of them. So the residual is not the leftovers; it is a pro-rata strip of the same credits Pantheon diligenced and bought. Adverse selection risk on lot 2 is structurally low. That is unusual and it is the whole reason this is investable.
Correct, and it deserves a discount. The residual averages $5.1M per position across 132 names, sitting alongside a Pantheon-controlled CV holding the majority of the same loans. A buyer of lot 2 gets little control. Offsetting it: the CV's manager is TCPC's adviser, so the two holders are administered by the same team, and a rival buying lot 2 acquires an economic strip without needing agency. I score this as worth 2–5 points of additional discount, i.e., lot 2 clears at 85–90 cents, not 90–95.
It is, and this is the honest headwind. Fitch's U.S. private credit default rate hit a record 6.1% on a trailing-twelve-month basis through July 2026, with median non-accruals across the 20 largest BDCs at 2.8%, up 80bp quarter-over-quarter. Morningstar/PitchBook LCD data puts top-10 BDC non-accruals at 3.95% at cost, but 5.95% on an adjusted basis that counts the whole debt stack of any borrower with one impaired tranche — 101 borrowers, and $772M of BDC interest income at risk, or 204bp of yield on a $38B cash-interest base. Lincoln International finds PIK on 11% of tracked private credit portfolios, with the "bad PIK" cohort's LTV having risen from 49% to 86%.
But the deterioration argument cuts both ways for TCPC specifically. Sector stress is a mark-to-model problem for BDCs that must hold their loans. TCPC is the one BDC that is selling — at 0.3× leverage, with $395M of liquidity, into a secondaries market where buyers still have capacity. As Unicus put it: the first credible seller sets the clearing price while there is still buyer capacity; the last seller discovers that "NAV" stands for No Asset Value. TCPC is going first. That is a feature.
This is the real tail risk and the reason for small sizing. If SDNY brings a valuation-fraud action, every mark in the book becomes untradeable, the sale process stalls, and buyers reprice for legal uncertainty rather than credit. TCPC printed a 52-week low of $3.08 on June 25, 2026 — at that price and today's pro-forma balance sheet you are at 0.52× NAV. Assume a $3.00–$3.40 floor in that scenario, roughly 18–28% downside. I size for it rather than pretend it away.
| Scenario | Prob. | Measurable trigger | Price | 12-mo total return |
|---|---|---|---|---|
| Bull — whole-book sale or BDC merger at ≥92c | 30% | 8-K with a definitive agreement by Q1 2027; or a stock merger with a larger BDC at ≥0.90× pro-forma NAV; Domo closes at par | $5.40–$5.75 | +44% |
| Base — sale at 88–92c, dividend cut to $0.11, orderly wind-down over 12–18 months | 45% | Q3 print (est. Nov 5) shows NAV $5.60–$5.95 and a dividend cut; process confirmed but not signed | $5.00–$5.20 | +34% |
| Bear — DOJ escalates or residual marks crack | 25% | Q3 NAV below $5.35; or an SDNY enforcement action; or the board announces redeployment of the $395M into new originations | $3.20–$3.60 | −11% |
| Probability-weighted expected total return | +25.6% | |||
Total returns assume a $4.15 entry and include ~$0.44 of distributions in bull/base and ~$0.30 in bear.
Three approaches, all anchored on pro-forma NAV of ~$5.90:
| Method | Assumption | Value / share |
|---|---|---|
| Liquidation at the observed print | Residual clears at 90c; less 1.5–2.5% wind-down and advisory costs | $4.95–$5.10 |
| Liquidation at the headline print | Residual clears at 95c; Domo at par credited | $5.40–$5.55 |
| Governance-discounted strip sale | Residual clears at 85c for minority-strip control discount | $4.65–$4.80 |
| Plus retained 5% CV stub | ~$9.5M of value not in the residual book | +$0.11 |
| My fair value range | Weighted, net of costs | $5.00–$5.30 |
Point estimate: $5.15. That is 24% above the current price, and it assumes the buyer extracts the same discount Pantheon did on a book that is now half as levered.
Ares is the named counterparty being shown the book. The structural point is bigger than one deal: the winners of a 6.1% default-rate cycle are the platforms buying assets at 88–92 cents with fresh capital, not the vehicles marking legacy books. But I will not pay up. ARES at $140.39 is 63× forward earnings with revenue growing 38.5% and ROE of 14.5% — a great franchise at a price that already assumes the franchise compounds. It is 25% off its $186.85 September high and 47% above its $95.80 March low.
The analytical framework from §3 generalizes, and it is the most useful thing in this note. What matters is not a BDC's discount to NAV — it is how much of the pool the equity actually funds. A BDC that finances itself by retaining the equity tranche of its own securitizations converts a 5-point move in loan marks into a 15–20 point move in your NAV. Screen for it and avoid it.
Buried in the Q2 disclosure: TCPC's new investments in Q2 went out at a 9.4% weighted average yield, while its exits came off at 10.9%. Read that again. The market that is defaulting at a record 6.1% is writing new paper roughly 150bp cheaper than the paper it is exiting. Private credit is not repricing risk; it is compressing spreads into a rising default rate because the capital formation machine cannot stop deploying. Consequence: the 2026 vintage will be the worst in the asset class's history, and it is being written right now. The trade is to own the platforms that can buy 2021–2023 paper at 90 cents and to avoid anyone whose earnings depend on originating 2026 paper at 9.4%.
Robert Miller of the University of South Dakota Law School published a paper this month, summarized today on Columbia Law School's Blue Sky Blog (full paper), arguing that private credit's problem is liquidity and information, not bank-style fragility. His central line is the best sentence written on this asset class this year: "A manager can update estimates every day, but unless inputs are observable and trade data is available, the result remains an internal model, not price discovery." His prescription is a deeper secondary market — transferable documentation, common identifiers, consistent reporting, settlement conventions, an LSTA-style coordinating body.
The TCPC whole-book shop is that prescription arriving by accident. Each print — Pantheon at ~90c, Blue Owl's $1.4B institutional sale at 99.7% of par in February, this residual at whatever it clears — installs a rung on a real price curve. Consequence: dispersion, not collapse. Once observable prints exist, a 23% equity discount on a 1.6×-levered BDC with 7% non-accruals at cost is correct, and a 30% discount on a 0.3×-levered book in an active sale process is wrong. The sector-wide "BDCs are cheap" trade dies; the specific-situation trade begins.
TCPC just demonstrated that the retained equity of a middle-market CLO — $130M at par against a $536M pool, issued in May 2026 — can be transferred within ten weeks at a price implying roughly a 28% loss on that equity. Consequence: every 2026-vintage private credit CLO equity piece on a BDC balance sheet is worth materially less than carried, and CLO liability spreads for middle-market collateral should widen. That removes the cheapest marginal leverage in the system. Downstream: BDC ROEs fall structurally, originations slow, and new-money loan spreads finally widen — resolving the anomaly in §13.1, but in 2027, not 2026.
An $8–10M pro-forma annual fee against $11.5 trillion of AUM is 0.03% of revenue — BLK at $1,176 is not a short on this. But the precedent matters: if the world's largest asset manager walks away from a public BDC because the fee doesn't cover the litigation, the assumption embedded in every alternative manager's multiple — that private credit AUM is sticky, permanent, high-margin capital — takes a real hit. It is not theoretical: BlackRock's ~$26B HPS Corporate Lending Fund faced 13.3% redemption requests earlier this year, and both BlackRock and Blackstone cut private credit fund values by roughly 5% in Q1. Regime levels: ARES below $120 or APO below $118 means the market is repricing fee-related-earnings multiples, not just credit. ARES above $160 means the narrative is repaired.
On August 21 I opened a SPECULATIVE long on TCPC at $3.85–$4.15 with a $5.00 target, on the thesis that the strategic review would end in capital return rather than redeployment. The Bloomberg report confirms that thesis three days later: the whole book is in market via KBW. Accordingly:
This is analysis, not investment advice. I hold no position in any security mentioned. Prices as of 16:07 UTC, August 25, 2026. Pro-forma NAV, NII, and liquidation values in this note are my own estimates derived from company disclosures and are labeled as such; they are not company guidance. TCPC is a ~$350M market-cap security with wide spreads — position size accordingly.