AI Momentum Portfolio Plunges 12% in a Single Week, Goldman Sachs Pivots: Gold and Copper Mining Stocks Become New Favorites

Publicado a 2026-08-24Actualizado a 2026-08-24

Resumen

Goldman Sachs believes that the expectation of a weaker dollar, driven by the US Treasury's expansion of its bond buyback program, will fuel the next round of gains for gold mining stocks.

The golden age of AI trading is quietly drawing to a close, with signals of market rotation becoming increasingly clear.

In a recent report, Goldman Sachs senior trader Natasha Tiwana warned of a fundamental shift in the momentum structure of the AI theme—semiconductors and the AI composite sector are moving from the long to the short side of the momentum factor, with the software sector taking over as the largest weight in short-term momentum.

Simultaneously, Goldman Sachs advises investors to shift their focus to previously overlooked areas: European and Japanese bank stocks, gold miners and copper mining stocks for exposure to hard assets, and hedges related to French political risks.

The backdrop for this strategic pivot is a severe deleveraging move the market experienced in August 2026. The Goldman Sachs High Beta Momentum portfolio (GSPRHIMO) fell 12% in a week, and the AI long-short hedge portfolio dropped 10% over five days. Goldman Sachs points out that in 2026, the number of days with momentum factor declines exceeding 5% has already surpassed the total for the past five years, forcing the market to seek broader diversification beyond the AI narrative.

AI Trading Enters a Phase of Fine-Tuned Operations

Goldman Sachs explicitly states that AI trading is not over, but its composition, momentum characteristics, and safety margins are being rewritten in real-time.

The report indicates that the market debate is rapidly shifting from "who are the winners of massive capital expenditures" to "who are the winners of large-scale AI adoption," driven by the continuous decline in computing costs. In this context, Goldman Sachs recommends a more tactical approach to investing in AI beneficiaries, focusing on opportunities where price and earnings per share (EPS) show significant divergence, with the storage chip (GSTMTMEM) and data center (GSTMTDAT) sectors offering the most compelling value.

Nvidia's Q2 earnings report is the most significant near-term catalyst, followed by support from the intensive industry conference season in September. On the AI application front, the collaboration announcement between MRNA and MRK this week sparked widespread market interest in healthcare AI. The breadth of positive EPS revisions for related beneficiaries (GSXGHDDD) has notably increased, broadening investor enthusiasm for the AI drug discovery space.

Quiet Reshuffle of Momentum Factor: Software Replaces Semiconductors

The structural reorganization within the momentum factor is one of the most noteworthy signals of this market move.

Goldman Sachs data shows that the overlap between 12-month winners and 3-month winners has fallen to a multi-year low, while the overlap between 12-month winners and 3-month losers is near a historical high. Specifically, the software sector has become the largest weight in the short-term momentum long portfolio (GSXUHMO3), while semiconductors and the AI composite sector have moved into the short side (GSXULMO3).

Goldman Sachs believes this reshuffling of long-short structures reflects a deep market reassessment of leadership in the next phase, leading to sustained high factor volatility. In this environment, demand for sophisticated hedging tools has risen noticeably, with investors preferring factor-based hedge portfolios over simple index tools to manage exposures.

European and Japanese Bank Stocks: Interest Rate Tailwind Meets Valuation Discount

Under the "broadening" allocation logic, Goldman Sachs lists European and Japanese bank stocks as core recommendations.

The report notes that over the past month, the market has unwound its pricing for Fed rate hikes, while rate expectations in Europe and Japan have diverged from the US. In this backdrop, the European Banks Index (SX7E) and Japanese bank stock basket (GSXAJMEB) benefit from net interest margin (NII) improvement in a "higher for longer" rate environment, alongside sustained positive non-interest income fundamentals—strong fee income growth, improving efficiency ratios, and significant shareholder return potential.

Data shows Japanese bank stocks have outperformed the TOPIX index and Japanese semiconductor stocks over the past three months, with lower volatility. European bank stocks trade at an aggregate ~15% discount to their US counterparts. Within Europe, Goldman Sachs' top value pick is Greek bank stocks—their valuations are converging towards eurozone core country peers but still offer a ~10% discount, with additional upside potential from M&A. Notably, positioning in European bank stocks is at its lowest in nearly two years, opening a window for contrarian positioning.

Gold Miners and Copper Mining Stocks: Clear Catch-Up Logic for Hard Assets

Goldman Sachs is also positive on the hard assets sector, providing specific valuation support.

The gold miners basket (GSXGOLDM) has risen 32% month-to-date but remains ~12% below its all-time high, with a forward P/E of just 11x, a ~20% discount to its five-year average. Goldman Sachs believes expectations of a weaker US dollar driven by the US Treasury's expanded bond buyback program will fuel the next leg up for gold miners. Additionally, Goldman Sachs' derivatives team notes rising demand for gold as a year-end hedge, with a double-digit option expiring December 2026 (paying out if gold miners rally >5% and S&P 500 falls >4%) priced around 6% (midpoint 4.25%).

Regarding copper mining stocks, while copper prices hit a record high this week, the copper mining basket (GSXGCOPP) has lagged the commodity itself and its solid fundamental earnings performance since February, primarily due to sentiment spillover from escalating Middle East tensions and AI sector selling. Goldman Sachs argues that tight supply-demand microstructures are supportive, and industrial metals exposure will directly benefit once the Fed turns dovish.

French Political Risk: An Underpriced Tail Exposure

Goldman Sachs also highlights French political risk as a potential market disturbance that is being overlooked.

The spread between French and German government bonds (OAT-Bund) has widened consistently since early June, reaching a recent high of ~85 basis points, yet French domestic equities (GSXEFRDO) have been largely unaffected so far. Goldman Sachs warns this calm may be unsustainable, as French domestic stock valuations are at the 90th percentile of their five-year history, implying almost no election uncertainty premium is priced in.

Goldman Sachs points out that with a dense political calendar ahead (including the French Employers' Federation summer meeting on Aug 27 and Justice Minister Darmanin's summer gathering on Aug 30), headline risks could heat up as early as next week. Historical data shows French domestic stocks' sensitivity to credit spreads jumps significantly during periods of political uncertainty, reacting far more sensitively to domestic risks than the CAC index.

September to be Key Window Testing Rotation Sustainability

Goldman Sachs concludes that all current signals point in the same direction: the market is being forced to move beyond the singular AI narrative towards broader diversification.

The quiet rotation in momentum towards software, structural buying in European and Japanese banks (especially the residual discount in Greek banks), still undervalued gold miners and lagging copper mining stocks, along with severely underpriced French political risk premium, together sketch a picture of deep market restructuring.

Goldman Sachs emphasizes that the "easy money" phase of AI trading is over. The only remaining alpha lies in tactically bottom-fishing select stocks where the divergence between price and EPS trajectory has become excessive. The September earnings season and conference calendar will ultimately judge whether this adjustment is merely a healthy position clean-out or the start of a more lasting market style shift.

Lecturas Relacionadas

Besu Patches Vulnerabilities in 5 Components: What Node Operators Need to Know

The Ethereum client Besu, developed by the Hyperledger community, has patched five security vulnerabilities discovered by blockchain security firm CertiK. These vulnerabilities, detailed in four security advisories on August 14, were all addressed in version 26.7.1, an urgent security update initially released on July 27. The intentional delay between the patch release and the public disclosure of details gave node operators a crucial window to update. JiaLiang Chang, CertiK's Director of Security Engineering, explained this "patch first, details later" model provides defenders a time advantage, allowing them to identify affected systems, test the update, and coordinate deployments—particularly important for institutional or permissioned blockchain networks requiring formal change management. The vulnerabilities, found through CertiK's "Chain Scan" attack methodology, involved issues in block announcement handling, consensus proposal buffering, WebSocket subscription limits, and JSON-RPC filter creation. If exploited, they could have allowed an attacker to exhaust a node's memory or thread resources, compromising its availability and the consensus process. Chang highlighted that while the open-source ecosystem is moving toward more formalized security testing (like differential fuzzing and bug bounty programs), coverage remains uneven. Testing often focuses more on protocol compliance than on continuous resource exhaustion, race conditions, or deployment-specific failures. He emphasized that third-party research remains vital for uncovering attack vectors beyond routine development, advocating for a mature, cumulative security model combining continuous integration, multi-node attack testing, independent audits, and regression testing for each confirmed vulnerability.

cryptonews.ruHace 20 min(s)

Besu Patches Vulnerabilities in 5 Components: What Node Operators Need to Know

cryptonews.ruHace 20 min(s)

Hack of Term Finance: Attacker Withdraws $8.5 Million Through Governance System Vulnerability

Decentralized lending protocol Term Finance lost $8.5 million due to an exploit of its governance mechanism, not a smart contract hack. The attacker manipulated the low voter turnout in the Meta Vaults system. By depositing just 0.5 ETH, they received a token (tmvETH) representing a share in the vault. Most users did not convert this token into a separate voting token (gtmvETH). The attacker did, gaining control of 90.66% of the issued voting power despite owning only 0.017% of the vault's capital. The attacker created a legitimate governance proposal to disable the 7-day withdrawal timelock and add their own contract to receive user funds. The proposal, publicly viewable for about 145 hours, passed due to a quorum requirement of only 5% and simple majority rule. After voting ended on August 23, 2026, the proposal was executed, allowing the immediate theft of 2,841 WETH and 1.68 million USDC (later swapped for DAI) from several vaults. The incident highlights systemic risks in decentralized governance when voting rights are separated from economic stake and low participation thresholds exist. Security mechanisms like timelocks were ineffective as the attacker could disable them from within the proposal itself. In response, Term Labs irrevocably closed all Meta Vaults and revoked DAO roles. The core lending protocol remained unaffected. The attack underscores the need for governance designs where security parameters are protected from modification through ordinary proposals.

cryptonews.ruHace 21 min(s)

Hack of Term Finance: Attacker Withdraws $8.5 Million Through Governance System Vulnerability

cryptonews.ruHace 21 min(s)

Trading

Spot
活动图片