Why a 2009 supply-chain framework is the right diagnostic for 2026

Igor Zaks, CFA, President, Tenzor Ltd. May 2026

The week the framework went on the record

PE-backed SaaS distress: a layering problem — Tenzor white paper, May 2026.

In the last days of April 2026, two senior figures running the largest PE-backed SaaS distress cycle in history said things in public that should have surprised no one — and yet did.

On April 28, Kort Schnabel, CEO of Ares Capital, walked his Q1 2026 earnings call through three Clearlake-owned software writedowns. Cornerstone OnDemand was cut to 73¢ on the dollar. Symplr to the mid-70s. DigiCert junior debt to 75¢. ^[1] Per Bloomberg’s reporting, Cornerstone EBITDA had grown 8% in 2025 and DigiCert’s by 16%. ^[1] Schnabel’s framing on the call:

Not everything is what it seems. A lot of it isn’t really performance-related. — Kort Schnabel, CEO Ares Capital, Q1 2026 earnings call [2]

Two days later, Orlando Bravo of Thoma Bravo, speaking at a Bloomberg event in Miami, confirmed what reporting had been pointing at since early April: Thoma Bravo would not inject fresh equity into Medallia. Lenders would take the keys. The wipeout — approximately $5 billion of sponsor and co-investor equity against a 2021 take-private valuation of $6.4 billion — would be borne. ^[3] Bravo’s framing:

We could do it — kick the can down the road another five years, pretend like it never happened. But we have a big fiduciary duty to our investors. — Orlando Bravo, Thoma Bravo, April 30, 2026 [3]

Two voices. The senior credit underwriter and the senior sponsor. Different sides of the table. And in the same week, both of them said — in different vocabulary — the same thing: the current writedowns and equity wipeouts are not primarily about operating performance. They are about something structural underneath the leverage. In our view, that structural thing is a layering problem. The diagnostic for it has been available since 2009. It just hasn’t been applied to software.

The 2009 framework

In a 2009 piece for International Corporate Rescue, Igor Zaks (then writing as Igor Zax) argued that the post-Lehman financial crisis would force what he called a “deplatformisation” of supply chains. ^[4] The argument was Coasean. Following Ronald Coase’s foundational 1937 paper, firm boundaries are determined by transaction costs: when transaction costs rise, vertical integration becomes more economic and the boundaries of the firm expand. ^[5]

In the pre-2008 model, electronics and consumer-goods OEMs operated as platform companies — what GaveKal had described in 2005 as firms that “produce nowhere and sell everywhere.” ^[6] The platform layer captured most of the value chain’s margin: roughly 12% EBITDA at the OEM, 3–5% at the contract manufacturer (EMS), and approximately 1% at the distributor. Moody’s rated the platform companies around A2; the layers around them were rated five or six notches lower at Ba1–Ba2. ^[4]

That stack worked when financing for the middle layers was cheap and reliable. In 2008–09 it stopped working. Credit insurance withdrew from the contract-manufacturer / distributor middle. Inventory financing tightened. The middle cracked first. The platform position turned out to be far more fragile than its A-rated credit suggested — because the platform’s economics depended on those middle layers continuing to function.

The argument was simple and is invariant to the trigger. When transaction costs rise — whatever the cause — firm boundaries expand and the most economic response is vertical integration. In 2008–09 the trigger was credit. In the geopolitical-chokepoint M&A cycle now developing, the trigger is physical-supply-chain risk. In the cycle we are working in directly, the trigger is twofold: leverage at peak-vintage software multiples, and AI-driven repricing of the orchestration layer. The framework is the same. The stack is the same. The question is which layer of the stack each company actually sits in today.

PE-backed SaaS distress: applying the 2009 framework

A decade ago, vertical SaaS occupied the platform layer in enterprise software. It sat between cloud infrastructure (its supplier), system integrators (its channel), and the enterprise customer — and captured the workflow value. The high entry multiples paid by PE in 2021–2022 reflected that position.

That position has shifted. The orchestration role has moved up the stack. Foundation-model providers — Anthropic, OpenAI, Google, Perplexity — and the hyperscaler suites embedding them (Microsoft 365 Copilot, Google Workspace, Salesforce Agentforce) now sit at the top of the value chain. They produce nowhere — they don’t own end-user workflow — but they sell everywhere, in the GaveKal sense, through API access and embedded distribution. They capture the orchestration margin.

Many PE-backed SaaS companies that traded at platform multiples in 2021 are no longer at the platform layer. The categories most exposed are seat-based, repeatable, narrow-workflow products that can be substituted by agentic capability sitting one layer up: observability, CPQ, contact-center, corporate learning, document automation, identity. These are now structurally closer to the contract manufacturers and distributors of 2009 than to the platforms they once were. Bloomberg Intelligence reported that approximately $17.7 billion of US tech loans dropped to distressed pricing in a single four-week stretch in early 2026, bringing the total tech distressed pile to roughly $46.9 billion. ^[7] The category is being repriced as a layer, not as a collection of individual companies.

Schnabel’s own framing — “a lot of it isn’t really performance-related” — is exactly this point in different vocabulary. The market is now pricing structural-position risk independent of operating performance, and getting ahead of the writedowns by sponsor-side recognition (as Bravo did with Medallia) is the responsible move.

Conversely, where vertical SaaS still occupies the platform layer — where it encodes industry-specific workflow that competitors cannot replicate — the position remains defensible. Veeva Systems is the cleanest available example: 120%+ net revenue retention sustained for years against life-sciences workflow that is not orchestration-replaceable. ^[8] Notably, Bravo’s own Thoma Bravo portfolio is split along this exact line. Medallia, acquired in 2021 at $6.4B, is being released to creditors with a roughly $5B equity wipeout. Proofpoint, acquired in the same year at $12.3B, has by Bravo’s own April 2026 disclosure grown EBITDA from approximately $250 million at acquisition to over $1 billion currently. ^[9] Same sponsor. Same vintage. Two opposite outcomes — driven by which layer of the stack each company actually occupies.

Five questions we ask in our work

In our practice, we approach distressed and pre-distressed software credit through five operational diligence questions. The framework is industry-agnostic — it traces directly to our receivables-finance and trade-credit work — and it travels because the structural questions are the same.

1. How far has the headline metric drifted from cash?

In a receivables-finance situation, invoiced volume keeps rising while the cash that should follow on a 30-, 60- or 90-day cycle does not arrive on time, or arrives only because new advances are being used to repay old ones. In a stressed PE-owned SaaS portco, ARR can keep being reported while the cash collected against that ARR softens. Deferred-revenue balances stretch, days-sales-outstanding drifts, churn shows up first in the contraction line and only later in the headline retention metric. Our diligence approach pulls bookings, billings, and cash collections separately and reconciles all three.

2. Is concentration hidden inside aggregate growth?

Receivables-finance failures often involved exposures that looked diversified at the program level and turned out to be highly concentrated at the obligor level — the Carriox / HPS Investment Partners case (more than $400 million of allegedly fictional collateral, reported by the Wall Street Journal in February 2026 ^[10]) is the cleanest recent example. SaaS books carry the same shape of risk in different form: a handful of customers driving net new ARR, or a single end-market carrying most of the renewal risk. We work cohort-level rather than portfolio-level, and concentration is a separate analytical lens.

3. How much operating deterioration is liquidity management postponing?

PIMCO reported that distressed exchanges accounted for 11 of 20 USD high-yield default events in Q1 2026 — a continuation of a pattern that has held for roughly a decade. ^[11] Liability-management exercises, dividend recaps, super-priority refinancings, and drop-down structures: each can be legitimate, and each can also be a way of postponing recognition rather than addressing the underlying issue. Anthology’s 2024 super-priority refinancing bought time and reorganized the capital stack but did not address the operating decline; the company filed Chapter 11 in September 2025 with $1.625 billion of funded debt. ^[12] We assess what the LME is actually fixing, and what it is deferring.

4. Where are the control gaps?

This is the question on which the parallel between trade-finance frauds and PE-backed SaaS distress is strongest. In Carriox, the failure point was customer-confirmation control on a remote obligor base. In Anthology, the public record points to integration cost overruns, mispriced fixed-cost implementation contracts, and overlapping product portfolios after a rapid roll-up. ^[12] In the PowerSchool MDL data-breach litigation, a federal court in March 2026 expressly allowed plaintiffs to proceed against Bain Capital on theories that sponsor-directed offshoring of cybersecurity functions caused the breach — an unprecedented ruling that has direct implications for sponsor-level operational diligence. ^[13] Different industries, similar shape: a control-environment weakness opened up by a structural cost-out, off-shoring, or M&A roll-up decision and not visible in the headline financials until it is too late. Our operational due diligence approach treats control-environment assessment as a first-class workstream, not a side-check.

5. How are stakeholder incentives shaping when the issue gets recognized?

PIK toggles in unitranche loans. Dividend recaps that increase the debt stack on a thesis. Drop-down liability-management exercises like the one Vista executed at Pluralsight in mid-2024. Distressed exchanges that reprioritize non-consenting first-lien holders, like the two Quest Software exchanges in 2025. Each of these is a tool, and each can be used responsibly. The diligence question is whether, in any specific situation, the tool is being used to fix the operating issue or to delay the moment when a covenant test or refinancing forces recognition. By the time recognition is forced, the real loss has often been crystallizing for some time. Bravo’s quote on Medallia is the responsible counterexample — recognizing rather than deferring — and it is rarer than the public record makes it look.

How we work this

Tenzor was built around exactly this kind of diagnostic work. Our practice has its origins in working capital optimization, receivables finance operational due diligence, and corporate restructuring across multiple cycles — including the 2008–09 cycle that produced the framework above; the IBRC distressed-real-estate workout in Ireland; and most recently the wave of receivables-finance fraud cases (Greensill, First Brands, Tricolor, Stenn, Carriox / HPS) that produced our public commentary in the Wall Street Journal, Bloomberg, BCR News, and the Sarachek Turnaround Take podcast in March 2026.

The translation from receivables-finance ODD into PE-backed SaaS is shorter than it looks. The operating questions — cash conversion, customer concentration, billings quality, control environment, working-capital composition — are the same. The vocabulary changes (ARR replaces invoiced volume, deferred revenue replaces accounts payable, renewal cohort replaces obligor concentration), but the diagnostic discipline is one and the same.

In the current environment we work in three engagement formats:

Independent operational due diligence ahead of covenant-relief discussions, refinancings, or LMEs. Typically a three- to six-week engagement covering the five questions above plus situation-specific extensions. Counterparties on the engaging side have included direct-lending vehicles, ad-hoc creditor groups, and PE LPs.
Pre-restructuring billings-quality and working-capital diagnostics where the immediate question is the gap between reported metrics and cash reality. Smaller in scope, faster — typically two to three weeks — and intended as input into a larger restructuring or workout discussion. Our work on the Greensill, First Brands and Carriox cases over the last 12 months has refined the methodology specifically for receivables-backed and recurring-revenue books.
100-day post-handover and turnaround support for lender groups or distressed funds that have taken control of an operating company. The structural framework above is most useful here: knowing which layer of the stack the company actually sits in determines whether the turnaround plan should focus on operating recovery, defensive consolidation, or controlled wind-down.

In each format, we work small and senior. Engagements are led personally by Igor Zaks, with project teams scaled to the situation. Senior-led, single-counterparty engagements are the model and have been since Tenzor was founded.

What we expect over the next 12 months

Three implications follow from the framework, and from what we are seeing in our current work:

For BDCs and direct-lending vehicles holding 2021–2022-vintage software unitranche paper. The relevant question for any borrower approaching a covenant test, maturity, or LME is no longer whether ARR has grown, but whether the cash collected against that ARR has tracked the reported metric — and which layer of the stack the company sits in. The Schnabel quote at the top of this paper is the public version of a private conversation that is now happening across multiple credit shops.

For sponsors managing PE-backed software portcos at high leverage. The Bravo quote is the public version of a discipline that the rest of the cycle has been postponing. Liquidity buys time, not solutions. The work has to start with the operating model, the renewal cohort, and the layer of the stack — not the cap-stack. Sponsors that recognize early have access to a wider range of outcomes (controlled handover, partial recovery, strategic sale) than sponsors that recognize late.

For lenders’ counsel and ad-hoc creditor groups. Operational and billings-quality diligence is increasingly the missing input in covenant-relief, LME, and pre-restructuring discussions. The Pluralsight, Quest Software, Anthology, and now Medallia and Cornerstone / Symplr / DigiCert situations all share a common pattern: by the time the legal and capital-structure work is fully developed, the operational picture is materially different from what either side initially assumed.

Every cycle produces a small number of cases that are not what they appear to be on the surface. Our experience across the 2008–09 cycle, the post-Greensill receivables-finance cycle, and now the PE-backed SaaS cycle is that the diagnostic discipline travels. The vocabulary changes; the questions do not.

If you are working a specific situation that fits any of the patterns above, we are happy to discuss it directly. We work from Toronto and engage globally.

A note on sources and limitations

Public reporting on private credit, liability-management exercises, and PE-owned company financials is necessarily incomplete: BDC-level marks and quarterly disclosures provide partial visibility, news reporting fills important gaps, and individual transactions remain private. The framework and observations in this paper draw on public reporting (Bloomberg, the Wall Street Journal, S&P Global Market Intelligence, PIMCO research, court filings, and SEC disclosures) supplemented by the published work cited at the back. Specific situations may differ from the patterns described, and engagement-level diligence is the only reliable substitute for the public-record reading.

References

[1] Davide Scigliuzzo and Laura Benitez, “Ares Slashes Value of Three Clearlake-Owned Software Companies,” Bloomberg, April 28, 2026.

[2] Ares Capital Corporation, Q1 2026 earnings call transcript, April 28, 2026.

[3] Preeti Singh, “Thoma Bravo Refuses to Inject Fresh Cash Into Ailing Medallia,” Bloomberg, April 30, 2026.

[4] Igor Zax, “Distressed M&A: Some Strategic and Financial Trends and Considerations,” International Corporate Rescue, Vol. 6, Issue 2 (Chase Cambria, 2009).

[5] Ronald H. Coase, “The Nature of the Firm,” Economica, New Series, Vol. 4, No. 16 (November 1937), pp. 386–405.

[6] Anatole Kaletsky, Charles Gave, Louis-Vincent Gave, Our Brave New World (Editions GaveKal, 2005), p. 6.

[7] Bloomberg Intelligence, “Distressed Software Loans Swell by $18 Billion in Span of Weeks,” Bloomberg, February 4, 2026.

[8] Veeva Systems, Form 10-K disclosures (FY2024 and FY2025); multiple sell-side reports confirming sustained net revenue retention above 120% over the prior eight fiscal years.

[9] Orlando Bravo, public remarks at Bloomberg event in Miami, April 30, 2026, as reported in Singh, Bloomberg (cited above).

[10] Wall Street Journal, “How Fake Invoices Duped BlackRock Unit Into a $400 Million Loan,” February 2026 (Tenzor commentary cited).

[11] PIMCO, “The Credit Market Lens: Differing Signals in BDCs, and Orderly Defaults in High Yield,” April 2026.

[12] Anthology, Inc., First Day Declaration in support of Chapter 11 cases, U.S. Bankruptcy Court for the Southern District of Texas, September 29, 2025.

[13] In re PowerSchool Holdings, Inc. and PowerSchool Group, LLC Customer Security Breach Litigation, S.D. Cal. (MDL); ruling on motions to dismiss, March 18, 2026. See also Womble Bond Dickinson commentary, April 2026.

The PE-Backed SaaS Distress Cycle: A Layering Problem