Golub Capital (GBDC) Q2 2026 Earnings Transcript

Golub Capital BDC reported NAV per share of $14.35, down from fair value markdowns and $0.52 per share of net realized and unrealized losses, while adjusted NII was $0.34 per share and fully covered the $0.33 dividend. Portfolio nonaccruals stayed low at 1.4% of investments at fair value, but management flagged continuing credit stress, wider spreads, and softness from lower SOFR, which reduced investment income yield to 9.7% and borrowing costs to 5.2%. The company also highlighted AI-related scrutiny in software, with 8% of its software portfolio deemed elevated AI disruption risk, while continuing share repurchases at a 16% NAV discount.

Analysis

GBDC is in a classic late-cycle BDC setup where accounting noise is masking a mixed fundamental picture: the equity is being hit by spread-driven marks while the cash earnings engine is only barely holding the payout. The second-order implication is that management now has a stronger incentive to keep originations selective and continue buying back stock rather than force balance-sheet growth, which should mechanically support per-share NAV but cap near-term loan growth. That makes the stock behave less like a clean credit beta proxy and more like a capital-allocation story.

The market is likely underestimating how much the borrower mix matters versus headline nonaccruals. If roughly 70% of unrealized losses sit in credits still behaving inside underwriting tolerance, a meaningful portion of the NAV drawdown can reverse with no “credit event” required; that creates a lagging positive catalyst over the next 1-3 quarters if spreads stabilize. The real risk is not current nonaccruals, but whether the stress broadens enough to convert today’s paper marks into realized losses just as refinancing walls and rate resets collide with weaker end-demand.

The software-AI framing is more important than it sounds: by publicly tagging a small slice of software exposure as higher disruption risk, management is signaling willingness to reprice or shrink exposure before performance deteriorates. That should improve underwriting discipline across the sector, but it also means lenders with less differentiated software credit selection will likely have to compete on spread or step back, pushing better assets toward the strongest platforms. In other words, GBDC may gain share in the best deals while the middle of the market gets priced for more risk.

Consensus appears too anchored on the dividend as the main variable. The more important variable is whether buybacks plus stable marks can keep NAV per share from compounding downward while lower base rates compress spread income; if that stabilizes, the stock can rerate even without earnings growth. If spreads widen further or credit stress becomes realized, however, the cushion from repurchases disappears quickly, and the shares likely re-trade back to a deeper discount to NAV.

Key Decisions for Investors

• Long GBDC common only on pullbacks into a wider discount-to-NAV entry zone; target a 3-6 month hold for mean reversion if unrealized marks reverse, but size modestly because the downside is still tied to realized credit losses.
• Pair trade: long GBDC / short a higher-leverage, more cyclical listed BDC with weaker nonaccrual trends and less buyback capacity; thesis is that GBDC’s capital allocation and underwriting should outperform in a stressed tape over the next 1-2 quarters.
• Sell near-dated upside calls against a starter long in GBDC if you already own the shares; the stock is more likely to grind on NAV and dividend uncertainty than to rerate sharply higher in the next quarter.
• Watch for confirmation in the next earnings cycle: if unrealized losses narrow and buybacks continue at a meaningful discount, add to the position; if realized losses tick up, exit before the market prices a dividend reset risk.
• Avoid chasing software lenders with materially weaker underwriting disclosure; GBDC’s AI-risk framework suggests the best platform lenders may win share while undisciplined competitors are forced to reduce exposure or misprice risk.