The Five-Year Clears 5%. Business Activity Runs Its Hottest in Five Years. The Nasdaq Sets Records and the Russell Still Falls.

The best economic data in five years was the worst news of the week for anything that has to borrow. S&P Global's flash September readings landed on Wednesday at 57.0 for manufacturing against a 53.7 consensus and 58.7 for services against 55.8, pushing the composite to 58.4 and a five-year high on a surge in new orders. The bond market took it from there. The five-year Treasury yield rose about 20 basis points to 5.03%, above 5% for the first time since 2007 and past the 4.99% peak of the last tightening cycle, after a $70 billion auction drew the weakest demand since 2006. The ten-year topped 5.12% intraday and finished the week at 5.18%. The thirty-year reached 5.39%, its highest since July 2004, and a Treasury buyback of up to $6 billion in long bonds on the same day failed to slow any of it. Every point on the curve except the two-year now trades above 5%, and futures price four more hikes over the next twelve months. Equities split hard. The Nasdaq set consecutive record closes early in the week and finished up 2.1%, the S&P 500 rose 1.2% to 7,743.41, and the Dow added 0.3% to 51,828.62 to snap a three-week losing streak. The Russell 2000 fell 0.8% to 2,837.55, the only major index lower. Utilities and real estate led Wednesday's decline, energy was the only sector higher, and crude fell for five straight sessions on reports that Iran had asked to return to the June framework, with WTI settling Friday at $92.41.

Analysis

The five-year is the number this library should read first, and not the ten-year that made the headlines. A micro-cap does not fund itself for thirty years. It borrows on a revolver, a term loan, or a convertible with three to five years on it, and the five-year Treasury is the floor under all three. That floor moved 20 basis points in a single session to 5.03% and printed above 5% for the first time since 2007. Every financing conversation in this universe now starts from a risk-free rate that has not existed in the working lifetime of most of the executives having it.

What makes the week unusual is that nothing went wrong. Manufacturing at 57.0 and services at 58.7 describe an economy accelerating, with new orders leading. In an ordinary cycle that is the setup small caps wait for, because domestic demand is where they earn. This is not an ordinary cycle. Strength now reads as more hikes, more hikes read as a higher discount rate, and the higher discount rate falls hardest on companies whose value sits in cash flows several years out. The tape said so precisely: the Nasdaq rose 2.1% on the week and set records while the Russell 2000 fell 0.8%. That is a 2.9 point gap in four sessions, wider than the 1.6 point gap in #035 and the 2.2 point gap in #036. Three consecutive weeks of the same trade, each one wider.

The split has a logic worth stating plainly rather than dismissing. The companies being bought are funding capital spending from operating cash flow and are not sensitive to the cost of borrowing, because they are not borrowing. The companies being sold are the ones for whom the next raise is a real event on a real calendar. That is a rational sort by balance sheet, not a mood, and it will not reverse because small caps look cheap. It reverses when the rate path stops moving, or when a company demonstrates it does not need the market. The one relief in the week came from energy: crude fell five sessions running on reports that Iran asked to return to the June framework, WTI settled at $92.41, and the diesel price that has been taxing every freight line in this library finally has a reason to ease. A proposal in Congress to ban U.S. diesel exports cuts the other way, and the industry response was immediate.

Coverage ledger: this is the third straight week in which the balance sheet, not the business, decided the outcome, which is the case for how this desk screens. UTMD holds $87.5M in cash and investments against no debt and ESP holds a $137.1M backlog against no debt; at a 5% five-year, not needing the market is itself the return. IIIN is debt-free but sells into construction, where project financing is now priced off the same curve, so it carries the rate through its customers rather than its balance sheet. NEOV is the clearest test in the library: demand from the Grid Modernization Act is real, and the ramp was funded with roughly $20M of equity in FY2026, which is exactly the profile the tape spent the week selling. BTBT sits on the side that was bought, though for the AI build rather than the mining, and the distinction matters more each week. LWAY and KRT get the first genuine cost relief in a month if diesel follows crude down. GLMD and ACON keep the cash floors that make the rate a question rather than a deadline. Full dossiers in the research library.

Movers