Investors in public business development companies are sending a clear message. As of October 1, 2026, the BDCs tracked by Raymond James traded at an average of 76% of net asset value and a median of 69%. Among venture lenders, some traded at premiums while others traded at steep discounts. Investors see real risk. The question is whether they are measuring it correctly.
Over more than three decades of investing in growth companies—first as a venture capitalist and later as a lender—I have learned that venture debt is not a single trade. Each loan has its own borrower and terms. Each lender has a different balance sheet. Two firms can both report meaningful software exposure and still own portfolios that behave very differently when conditions tighten.
Sector labels are useful starting points, not underwriting conclusions. That does not mean every venture lender is undervalued. It means investors cannot know which discounts are justified without looking at the loans behind them.
The concern around software is justified. J.P. Morgan estimates that software represents about 21% of public BDC portfolios, while Moody’s puts the median exposure at roughly a quarter. AI could reduce paid seats and weaken renewal rates, pressuring margins and the enterprise value supporting some loans.
Even those estimates rest on labels that are not applied consistently. BDCs do not share a common industry taxonomy, so a business one lender files as software may appear under business services, IT services or high tech at another. “Software” covers an enormous range of companies. A platform embedded in a customer’s workflow, especially one built on proprietary data, does not necessarily carry the same risk as a generic application. The right question is what the software does and how difficult it is to replace. The company’s ability to produce durable cash flow matters just as much.
Redemption pressure at nontraded BDCs and other semiliquid private-credit vehicles has added to the anxiety. Those requests can limit a manager’s flexibility, particularly when a fund offers periodic liquidity while holding loans that may take years to mature. But redemption pressure at the fund level does not mean every loan in the portfolio is in trouble.
Investors also remember the easy-money years. When too much capital chased too few deals, some lenders accepted looser terms and weaker covenants. Others assumed another equity round would always be available. When capital became scarcer, those decisions mattered.
That history calls for scrutiny, not blanket conclusions. A later-stage company with meaningful revenue and a manageable burn rate is different from one whose repayment depends on the next financing round. A credible path to cash flow and the strength of the equity cushion widen that difference.
Recent data make that distinction more important. PitchBook recorded $64.7 billion of U.S. venture-debt issuance in the first half of 2026, nearly matching the full-year 2025 total. But only 280 loans produced that volume, and a single $20 billion refinancing for SpaceX accounted for almost one-third of it. The headline reflects a divided market in which a small group of large, established borrowers is absorbing much of the capital while many traditional software companies face harder questions about growth and durability. In a market this divided, sector-level numbers say even less about any single lender’s book.
For investors, the analysis begins with whether the borrower has a durable business and enough liquidity to reach sustainable cash flow. Debt can help a sound company get there, but it cannot fix a business model that does not work.
Structure comes next. Seniority and collateral determine a lender’s position in a restructuring. Loan-to-value shows the cushion beneath the debt; maturity and covenants determine when a lender can act. A covenant does not guarantee repayment, but it can bring a lender and borrower to the table before the company runs out of cash. Regular financial reporting and performance tracking can reveal deterioration before a covenant breach, allowing lenders to engage while a borrower still has options.
The lender matters as well. Investors should look at portfolio concentration, non-accruals, payment-in-kind income, and realized losses. Leverage and funding maturities affect flexibility, while dividend coverage shapes shareholder returns. Good loans can still sit inside a poorly financed vehicle.
Investors should also assess confidence in net asset value (NAV). A discount can reflect both expected losses and skepticism about existing marks. For private loans, examine how valuations are established, the role of independent third parties and how prior marks compare with realized exit values.
At Runway, we believe our public disclosures show the kinds of distinctions investors can examine. As of June 30, 2026, 96% of our loan portfolio was senior secured. Software represented 27% of the debt portfolio. Of that exposure, 93% was first lien, 94% had financial covenants, and 94% was sponsor-backed. Those borrowers span nine end markets, including logistics, utilities, insurance and government/defense. Our cumulative net loss rate since inception in December 2016 was 1.72%.1
Those figures do not mean the portfolio is immune to a credit cycle. They show what we think investors are being asked to value. Another lender may report very different numbers, and that is the point.
There is a real risk here. AI may reduce enterprise software spending faster than some borrowers can adapt. Seniority and covenants can help, but they cannot create value that no longer exists. Some of today’s discounts may prove rational.
The evidence so far, however, does not show a broad collapse in software credit. A July 2026 analysis from the Bank for International Settlements found that less than 1% of BDC software loans were behind on payments. Moody’s has observed that software maturities do not rise more meaningfully until 2028 and 2029, making AI more of a monitoring issue in the near term than a current driver of widespread losses. Those findings call for vigilance rather than complacency, since credit metrics look backward and payment-in-kind structures can delay the moment when stress becomes visible.
The market is right to demand a risk premium from venture lenders. But a sector label cannot tell investors which discounts are deserved. Venture debt is a loan-by-loan business held inside a portfolio and a balance sheet. Sentiment can set the share price today. Over time, value will depend on borrower performance and how well the lender underwrites, monitors and values the loans.
1 Past performance is not indicative of future performance or outcomes. The article is provided by Runway Growth Capital for informational and educational purposes only. The views and opinions expressed herein are those of the author(s) as of the date of publication and are subject to change without notice. Historical portfolio data and results are not indicative of future performance or outcomes. Cumulative net loss rate is calculated as net debt and equity losses divided by total commitments since inception. The views expressed are not intended to be, and should not be construed as, investment, legal, tax, accounting, or other professional advice. Certain statements contained in this article reflect the author’s opinions, interpretations, expectations, or judgments regarding market, economic, industry, and investment-related matters. Such opinions are inherently subjective and may differ from the views of other market participants. There can be no assurance that any opinions, expectations, or forecasts discussed herein will be realized.
——–
Author:
David Spreng is the Founder, Co-CEO, and Co-CIO of Runway Growth Capital and chairs the firm’s investment committee. With 30 years of experience as a venture capitalist and growth debt lender, he has helped form and develop nearly 50 technology companies.
Website: https://runwaygrowth.com/