Treasury yields hit multidecade highs as AI-linked technology shares show resilience
publishers 40articles 60first reported 23 Sep, 23:47 UTCdeveloping since 23 Sep · 9 editions
A selloff in U.S. and European government bonds deepened Thursday, pushing long-term Treasury yields to multidecade highs and weighing on stocks. CNBC reported that the benchmark 10-year U.S. Treasury yield touched 5.342%, its highest level in 24 years, while the 30-year yield reached 5.683% before both retreated from their morning peaks.
U.S. stock futures were little changed to modestly higher after a losing session driven by rising yields, with Nasdaq futures leading early gains. Technology shares tied to artificial intelligence remained comparatively firm: Alphabet rose in premarket trading after Google released its Gemini 4 Argon AI model, while Broadcom gained modestly following a Reuters report that it agreed to lend Anthropic as much as $42 billion for AI infrastructure. Akamai shares also rose on news of an Anthropic deal, according to TradingView.
Micron shares declined before the opening despite strong quarterly results and the company’s forecast that AI-driven memory supply constraints would become more pronounced in 2027 and 2028. CNBC attributed the weakness to current-quarter gross-margin guidance affected by higher-cost inventory associated with employee incentive compensation.
Oil prices also rose, while higher Treasury yields increased investor interest in bonds offering returns around 5%, MarketWatch reported.
HOW THIS STORY WAS MADE
Written from 8 articles, 1 with the publisher's own text; 7 independent newsrooms once syndicated copies count as one; this version written 16 h after the record first saw the story; the editor kept it as written.
- Publishers
- 40
- Source articles
- 60
- Given to the writer
- 8, 1 with the article's own text
- Independent newsrooms
- 7
- This version written
- 16 h after the record first saw the story
- Second model (editor)
- kept as written
- Publication gate
- passed
- Human review
- none
Written by a language model from the sources above, then checked by a second model that may only cut, attribute or correct. How it works →