Golub Capital BDC, Inc. Q2 FY2026 earnings call

David Golub discussed private credit trends, including shift from borrower-friendly to lender-friendly market. Tim Topicz detailed credit performance, investment income yield, borrowing costs, and earnings drivers. Robert Tuchscherer talked about investment activity, with GBDC participating in limited new originations, selective underwriting, and focus on core middle market. Software portfolio makes up ~26% of GBDC's portfolio, with 95% in top ratings and 8% facing elevated AI disruption risk. Chris Ericson covered financial results, balance sheet, and debt funding structure.

GBDC had a small loss for the quarter, about 1% of NAV, primarily due to mark-to-market fair value write-downs. Adjusted NII per share for the quarter was $0.34, corresponding to an annualized adjusted NII return on equity of 9.5%. Nonaccruals remain low. Approximately 89% of GBDC's investment portfolio at fair value is in highest performing internal rating categories. Investment income yield was 9.7% annualized, down 30 basis points sequentially due to lower SOFR. Borrowing costs declined 20 basis points to 5.2% annualized. Adjusted NII per share covered the $0.33 per share base distribution. Credit spread widening drove majority of net realized and unrealized losses. NAV per share declined to $14.35. Net debt to equity was 1.24x. Company repurchased 2.2 million shares in the quarter.

David Golub predicted private credit is in a Darwinian moment, credit stress to continue, and market to become more lender-friendly, with wider spreads creating medium- and long-term benefits. Confident Golub Capital and GBDC will be winners.

Mark-to-market fair value write-downs from spread widening, AI and software concerns, continued credit stress with subset of companies needing restructuring, macro factors like Middle East situation impacting M&A, and redemption pressure on nontraded BDCs.

Q: Just one on the software loan side of the portfolio. And you talked about the new AI risk framework, and I think it's about 8% of the investments being at risk there. Wonder if you could just talk a little bit more about the -- some of the characteristics that underlie some of those investments, commonalities there? And what sorts of mitigation could you see being performed over time on those types of investments?

A: Sure. Thanks, Ken. I'll start, and Rob, maybe you can add to what I'm going to say when I'm done. So for some context, we started investing in software at a time when almost no other lenders did. So the idea of being a lender to this space at a time that's contrarian is, for us, not uncomfortable. In some ways, all of the noise that you're hearing right now about risks in software is good for us because we understand the difference between good software credits and bad software credits, and it means less competition. What we're seeing in the marketplace right now is many pure lenders who had started to get into software lending in the last few years want to be able to report to their shareholders how they're reducing their software exposure. So they're literally not participating in marketplace opportunities for new loans. So that's just some context for you. The exercise that Rob talked about involved looking at our portfolio from the standpoint of degree of AI disruption risk. And he correctly said that 8% of the roughly 25% of our portfolio that's in software. So it's roughly 2% of our overall portfolio is in a category of elevated AI risk. That doesn't mean we think we're going to lose money on these loans. They could be low leverage, they could be near maturity. There are a lot of other factors that go into whether we're going to see elevated risk of credit loss in these loans. But this is a very important rating system from the standpoint of both evaluating new loans and helping us figure out from a monitoring perspective, what should be our goals with those borrowers. So for example, it would be reasonable to conclude that if we see elevated risk of AI disruption, we're going to want to reduce exposure or we're going to want to get paid for the exposure that we're taking. We may want to increase pricing. We may want to increase equity cushions. We may want to take other steps that reduce risk. So what kinds of companies fall in this category. The most significant element of the category are companies that are involved in providing tools that enable others who are writing code to do so more effectively. This has historically been a significant category of software companies. It's not a category that we've historically been attracted to, but we do have a couple of exposures that fall in this category. I'd say that's the largest component of the group. Rob, if you want to add more color, please do.Q: Just one on the software loan side of the portfolio. And you talked about the new AI risk framework, and I think it's about 8% of the investments being at risk there. Wonder if you could just talk a little bit more about the -- some of the characteristics that underlie some of those investments, commonalities there? And what sorts of mitigation could you see being performed over time on those types of investments?

A: Sure. Thanks, Ken. I'll start, and Rob, maybe you can add to what I'm going to say when I'm done. So for some context, we started investing in software at a time when almost no other lenders did. So the idea of being a lender to this space at a time that's contrarian is, for us, not uncomfortable. In some ways, all of the noise that you're hearing right now about risks in software is good for us because we understand the difference between good software credits and bad software credits, and it means less competition. What we're seeing in the marketplace right now is many pure lenders who had started to get into software lending in the last few years want to be able to report to their shareholders how they're reducing their software exposure. So they're literally not participating in marketplace opportunities for new loans. So that's just some context for you. The exercise that Rob talked about involved looking at our portfolio from the standpoint of degree of AI disruption risk. And he correctly said that 8% of the roughly 25% of our portfolio that's in software. So it's roughly 2% of our overall portfolio is in a category of elevated AI risk. That doesn't mean we think we're going to lose money on these loans. They could be low leverage, they could be near maturity. There are a lot of other factors that go into whether we're going to see elevated risk of credit loss in these loans. But this is a very important rating system from the standpoint of both evaluating new loans and helping us figure out from a monitoring perspective, what should be our goals with those borrowers. So for example, it would be reasonable to conclude that if we see elevated risk of AI disruption, we're going to want to reduce exposure or we're going to want to get paid for the exposure that we're taking. We may want to increase pricing. We may want to increase equity cushions. We may want to take other steps that reduce risk. So what kinds of companies fall in this category. The most significant element of the category are companies that are involved in providing tools that enable others who are writing code to do so more effectively. This has historically been a significant category of software companies. It's not a category that we've historically been attracted to, but we do have a couple of exposures that fall in this category. I'd say that's the largest component of the group. Rob, if you want to add more color, please do.Q: Just one on the software loan side of the portfolio. And you talked about the new AI risk framework, and I think it's about 8% of the investments being at risk there. Wonder if you could just talk a little bit more about the -- some of the characteristics that underlie some of those investments, commonalities there? And what sorts of mitigation could you see being performed over time on those types of investments?

A: Sure. Thanks, Ken. I'll start, and Rob, maybe you can add to what I'm going to say when I'm done. So for some context, we started investing in software at a time when almost no other lenders did. So the idea of being a lender to this space at a time that's contrarian is, for us, not uncomfortable. In some ways, all of the noise that you're hearing right now about risks in software is good for us because we understand the difference between good software credits and bad software credits, and it means less competition. What we're seeing in the marketplace right now is many pure lenders who had started to get into software lending in the last few years want to be able to report to their shareholders how they're reducing their software exposure. So they're literally not participating in marketplace opportunities for new loans. So that's just some context for you. The exercise that Rob talked about involved looking at our portfolio from the standpoint of degree of AI disruption risk. And he correctly said that 8% of the roughly 25% of our portfolio that's in software. So it's roughly 2% of our overall portfolio is in a category of elevated AI risk. That doesn't mean we think we're going to lose money on these loans. They could be low leverage, they could be near maturity. There are a lot of other factors that go into whether we're going to see elevated risk of credit loss in these loans. But this is a very important rating system from the standpoint of both evaluating new loans and helping us figure out from a monitoring perspective, what should be our goals with those borrowers. So for example, it would be reasonable to conclude that if we see elevated risk of AI disruption, we're going to want to reduce exposure or we're going to want to get paid for the exposure that we're taking. We may want to increase pricing. We may want to increase equity cushions. We may want to take other steps that reduce risk. So what kinds of companies fall in this category. The most significant element of the category are companies that are involved in providing tools that enable others who are writing code to do so more effectively. This has historically been a significant category of software companies. It's not a category that we've historically been attracted to, but we do have a couple of exposures that fall in this category. I'd say that's the largest component of the group. Rob, if you want to add more color, please do.