Roteiro macroeconômico do Goldman Sachs para o segundo semestre: a competição por capital determinará os rumos do mercado.

Goldman Sachs macro traders, in their latest H2 roadmap, point out that the market's mainstream narrative on AI is shifting from a "software cycle" to a "capex cycle," a transition that will profoundly reshape the interest rate market landscape.

According to a co-authored report by Goldman Sachs macro traders Cosimo Codacci-Pisanelli and Rikin Shah, AI infrastructure construction is physically closer to railway building than to the software expansion of the dot-com bubble era. This implies that massive capital expenditures require sustained financing, with both the private and public sectors competing for increasingly scarce funds, which will substantively cap the rebound potential in the belly and long end of developed market curves.

Meanwhile, the softer US June nonfarm payroll data has cooled expectations for a July rate hike, but Goldman believes a rate hike remains "an option, not a necessity," with inflation data being the ultimate deciding factor.

This roadmap has direct implications for bond investors: Goldman advises selling into rallies on the long end of the curve rather than chasing gains, arguing that fiscal supply pressures have gained a "second leg" from private sector capex financing needs, both vying for the same shrinking pool of duration buyers.

AI narrative framework flip: From software cycle to railway-like capex

The core thesis of Goldman's report is that markets have long characterized AI as a software cycle, supporting optimistic expectations for interest rate trends—believing that AI will ultimately lead to lower inflation, reduced computing costs, improved corporate margins, and even a lower neutral rate. Goldman suggests this "destination" assessment may not be far off but seriously underestimates the cost of the "journey" itself.

The report points out that the software cycle can depress equilibrium rates because its expansion requires little capital; in stark contrast, an infrastructure boom pushes rates higher by competing for scarce capital. Current capex forecasts for hyperscale data centers are being revised upward, and financing pressures have clearly spilled over from private balance sheets—credit issuance is large and still expanding, while equity financing channels have also been activated.

The report also warns that if these capex levels materialize, physical bottlenecks such as electricity, power grids, skilled construction labor, and cooling systems will pose potential headwinds to inflation normalization. Goldman argues that the burden of proof on whether and how quickly productivity gains will materialize now falls on the market bulls.

Equity market rotation already signals: ROIC concerns arise

Goldman sees structural rotation in equity markets as the clearest market signal. The report notes that capital flows from AI capex "spenders" to "beneficiaries" suggest the stock market has begun pricing in "return on invested capital" concerns, rather than merely chasing narratives.

The report acknowledges that this issue is still undecided, recommending placing more weight on known certainties—i.e., the reality of intensified capital competition. Goldman concludes that AI may still deliver the disinflation and lower neutral rates implied by the software cycle, but before reaching that destination, massive physical construction must be financed, and that funding must come from somewhere.

This logic directly points to the interest rate market: the rebound potential in the belly and long end of developed market curves is constrained, with fiscal supply pressures now joined by the private sector, both competing for the same shrinking pool of duration buyers. Goldman's trading recommendation: sell into rallies on the long end of the curve.

July rate hike expectations cool: Soft nonfarm payrolls buy time for the Fed

On the Fed's policy path, Goldman believes the June nonfarm payroll report has "deflated" expectations for a July rate hike. The report shows that the three-month average nonfarm payroll increase fell from 188,000 to 111,000; the unemployment rate edged down, but alongside a declining labor force participation rate, the real significance is limited. Additionally, June saw negative employment growth in the hotel sector, with revised data showing healthcare services as almost the sole true source of job creation.

Goldman notes that overall employment data is soft, consistent with other labor market indicators—weak job openings, a declining labor market differential, sluggish hiring intentions, and no rebound in key wage growth.

Regarding Fed officials' statements, Fed Chair Warsh's remarks at the Sintra conference were cautious, with Goldman likening his style to "Greenspan-era communication." Goldman views his overall stance as slightly dovish—Warsh described inflation risks as having declined, and his comments on AI focused on long-term supply-side effects rather than the capital cost issues Goldman emphasizes.

Goldman maintains its base case: the Fed could raise rates, but it's an option. Energy price trends are moving in the right direction, and PCE methodology adjustments will help lower sequential inflation readings. A deeper question is whether a rate hike is even the right tool to address a capex-driven cycle—the rate-sensitive areas that rate hikes can truly affect (e.g., housing) are already under significant pressure. Goldman expects that even if the Fed initiates rate hikes, there is no rationale for a sustained tightening cycle, with no more than 2 to 3 hikes.

H2 core theme: Micro driving macro again

In its summary, Goldman states that market focus is returning to AI and capital costs, with micro again becoming the driver of macro trends, but the narrative framework is now entirely different from the start of the year. The destination may still be the picture painted by the software cycle—lower inflation, margin improvement, lower neutral rates—but the current path is a capex cycle with railway-like physical scale that must be financed.

The competition for capital between the private and public sectors, credit and equity, is the core narrative for H2 and the fundamental reason for the limited rebound potential in the belly and long end of developed market curves. In the US short end, nonfarm data has weakened the case for a July hike; a rate hike remains optional, inflation data will determine September's direction, and whether rate hikes are suitable for a capex-driven cycle remains an open question.

Risk Warning and Disclaimer

        Market risk exists; invest with caution. This article does not constitute personal investment advice and does not consider the specific investment objectives, financial situation, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article are suitable for their specific circumstances. Investment decisions based on this content are at your own risk.