Copper Prices Suddenly Reverse After Surging Higher, Shadow of Rate Hikes Looms: Can the $6.45/lb Rebound Still Be Chased?

Published on 2026-09-16Last updated on 2026-09-16

Abstract

Soaring energy prices have reignited inflation concerns, while a strengthening dollar and the approaching Federal Reserve meeting are putting pressure on copper prices. LME copper fell to $14,024.5 per ton, while SHFE copper also declined. Seasonal low-price procurement has shown some recovery, but overall demand remains primarily rigid.

Stronger U.S. Dollar Weighs on Copper Prices

Overnight, LME copper opened at $14,085 per ton, hitting the session high at the open before trending lower throughout the day. It fell to a low of $13,958 and ultimately closed at $14,024.5, down 1.42%. Trading volume was 27,000 lots, while open interest decreased by 1,215 lots to 274,000 lots, indicating long positions were actively being reduced.

The SHFE copper main 2610 contract also faced pressure. It opened at 106,900 yuan per ton, reached a high of 107,020 yuan and a low of 106,650 yuan, and settled at 106,920 yuan. Open interest decreased by 3,283 lots. The synchronous decline in both domestic and international copper prices suggests the pressure primarily stems from macro-oriented fund flows, rather than short-term fluctuations in a single market.

Rising Energy Prices Reignite Inflation Concerns

With the Federal Reserve's policy meeting approaching and Middle East geopolitical conflicts pushing energy prices higher, market worries about a resurgence in inflation have returned. This has provided support for the U.S. dollar, making dollar-denominated copper more expensive for buyers using other currencies.

Heightened expectations for interest rate hikes also increase costs for corporate inventory holding and trade financing. As a capital-intensive commodity, copper is highly sensitive to changes in interest rates and the dollar. The expectation of tighter macro policies prompted long positions established at earlier highs to take profits.

Demand Recovery Remains Largely Necessity-Driven

On the fundamental side, the traditional peak consumption season and the price pullback have spurred some bargain-hunting, leading to a recovery in downstream demand compared to earlier periods. However, procurement remains primarily necessity-driven, with companies not showing a strong willingness for large-scale chase-buying or inventory replenishment.

This suggests copper prices may oscillate in the short term between macro pressure and low-level procurement. If the dollar continues to strengthen and rate hike expectations intensify, the room for any rebound will be limited. Only a significant improvement in end-user orders and proactive inventory replenishment can allow demand-side factors to re-emerge as the dominant force driving sustained price increases.

Related Reads

"Strictest Ever" Wet-Lab Prediction Benchmark Released: Maxed-Out Reasoning and 5x Compute Still Won't Cut It

UniPat AI has launched Science Prediction Universe (SPU), a new benchmark that evaluates whether AI models can predict the outcomes of real wet-lab experiments before they are conducted. The benchmark covers 116 tasks and 4,207 wet-lab experiments across chemistry, materials science, biology, and medicine, drawing on recent papers and unpublished collaborations with institutions such as Stanford, Cornell, and Harvard. Models are given eight hours, an H20 GPU, internet access, and scientific software, and must submit a ranked prediction of experimental results, which is then scored against hidden ground-truth data using the Spearman correlation coefficient. To prevent memorization, evaluation runs in isolated, offline containers and screens outputs for paper identifiers. Results show that even leading models like Opus 5.5 and GPT-6 Astra remain far from reliable prediction: the best average score was 0.486, and only about half of tasks exceeded the random baseline. The study also found that increasing reasoning effort fivefold barely improved accuracy, since models often settled on an answer early and spent the rest of the time on minor adjustments. Likewise, more tool use and GPU calls did not reliably help. Performance varied sharply by measurement target, and API costs ranged 300-fold across configurations. SPU aims to give researchers and model developers a unified way to measure and improve AI's scientific judgment before investing in costly experiments.

marsbit2h ago

"Strictest Ever" Wet-Lab Prediction Benchmark Released: Maxed-Out Reasoning and 5x Compute Still Won't Cut It

marsbit2h ago

From Harshly Criticizing Stablecoins to Building an L1: Airwallex's Jack Zhang Isn't "Converting" — the Battle for the Settlement Layer Has Begun

In June 2025, Airwallex CEO Jack Zhang publicly questioned stablecoins, arguing that fiat conversion costs outweighed their cross-border payment benefits and that crypto had failed to solve real-world problems for 15 years. Yet just 17 months later, in November 2026, Zhang emerged as co-founder of Alta, a blockchain project that raised $80 million (a $47.5 million Series A led by Index Ventures plus a $32.5 million seed round). Rather than a change of heart, the move reflects commercial necessity: a battle over the stablecoin settlement layer is underway, and payment giants that stay out risk becoming mere interfaces atop someone else's chain. Alta is not a general-purpose public chain but a compliant Layer 1 built for banks and licensed payment institutions. Co-founder Loong Wang, previously of Ren Protocol, says existing chains like Ethereum and Solana cannot simultaneously meet institutional demands for privacy, compliance, integration with local financial systems, and liquidity. Alta uses recursive zero-knowledge proofs to enable on-chain compliance checks without exposing identities or transaction amounts, and builds in programmable compliance modules, native identity authorization, and anti-airdrop protections. Co-founder Catherine Porter previously led Meta's Libra (Diem) partnerships. Airwallex itself is Alta's key advantage: with $287 billion in annualized volume, $1.3 billion in revenue, 200 countries, and 90 currencies, it will channel enterprise clients onto the new chain. Zhang's apparent reversal is consistent: he criticized stablecoins as cheap remittance tools, but now bets on tokenized fiat settlement infrastructure. Competitive pressure comes from Stripe, which acquired Bridge for $1.1 billion, bought Privy, and incubated the Tempo chain (raising $500 million at a $50 billion valuation, with OpenAI, Shopify, Deutsche Bank, and Visa as partners). Circle's Arc, Fireblocks, and Canton Network are also competing. With stablecoin market cap exceeding $320 billion and 2025 on-chain volume surpassing $33 trillion, the real contest is over who controls minting, clearing, and fiat redemption. Success will depend less on technology than on three factors: genuine adoption by traditional financial institutions, compliance adaptation across jurisdictions, and efficient on- and off-ramps to real banking systems. The paradigm has shifted: instead of disrupting banks, fintech giants now aim to build infrastructure that lets traditional finance voluntarily move on-chain. Pragmatism, not ideology, now dominates the race.

marsbit4h ago

From Harshly Criticizing Stablecoins to Building an L1: Airwallex's Jack Zhang Isn't "Converting" — the Battle for the Settlement Layer Has Begun

marsbit4h ago

"Everyone is too conservative with AI!" Head of Claude Code: My approach is more ambitious

Felix Rieseberg, who leads Claude Code's desktop app, argues that most people use AI too conservatively and should set more ambitious goals. His central trick: set unreasonably aggressive targets and insist they're achievable—for example, instead of cutting website load time from 10 seconds to 5, aim for 0.2 seconds, and when the model protests, keep telling it the problem is solvable and to try again. He organizes ambition into four categories: speed (doing the same work faster), height (raising the abstraction level to focus on desired outcomes rather than individual tasks), depth (working like top experts, using their real tools, data, and methods), and breadth (drawing on ideas from other fields). For height, he suggests asking what you ultimately want to happen and what the goal behind the goal is—e.g., instead of fixing single crashes, require the app to never crash. For a quarterly business review, move from filling in a template to understanding what changed and why, identifying the most urgent decisions, and redesigning the review process itself. For depth, he advises insisting on the tools top professionals actually use (databases and statistical models rather than Markdown tables), setting "unreasonable" targets, challenging your own assumptions against raw data, hunting for new knowledge, and using external validation rather than letting the model grade itself. A useful prompt asks the model what top experts would do that you haven't requested, what raw materials and tools they'd use, and what besides the model and you could prove the result correct. For breadth, he recommends replacing nouns with verbs (not "make a PPT" but "in 30 minutes, convince busy people of an explanation and drive a decision under uncertainty"), assembling cross-disciplinary teams, and using multi-agent setups—while avoiding mere role-play in favor of adopting other fields' actual methods. He closes with self-efficacy research: people given clear, challenging goals outperform those told to do their best, and believing difficult things are possible—as with the four-minute mile—unlocks achievement. In an era of rapid capability gains, updating our beliefs about what's possible matters more than ever.

marsbit4h ago

"Everyone is too conservative with AI!" Head of Claude Code: My approach is more ambitious

marsbit4h ago

Trading

Spot
活动图片