Cerebras launched the CS-4 system to directly compete with NVIDIA, claiming its inference speed is up to 30 times faster than GPU servers.

1. Cerebras unveiled its new rack-mounted system, the CS-4, on Tuesday, claiming that its token generation capacity in single-user scenarios is 30 times that of traditional GPU servers; 2. The CEO of Cerebras stated that the CS-4 can achieve industry-leading speeds on cutting-edge large models, fundamentally reshaping industry paradigms; 3. Since its listing in May, the company's stock performance has been underwhelming.

cls.cn22minutes ago

Will the AI boom end like the 19th-century railway bubble?

Technology analyst Thompson warns that while AI is not a false premise, the capital chain may break before returns materialize. He draws an analogy to the 19th-century railway bubble, noting that funding constraints are more urgent than issues of electricity and computing power. Financing has progressed from burning through free cash flow to tapping debt markets, now approaching the limits of equity financing. If the current AI bubble bursts, its most likely legacy will be enhanced power infrastructure.

wallstreetcn11:34

Elon Musk Warns Nvidia-Backed $105 Billion Ohio Data Center for OpenAI Will Take 'Much Longer' to Come Online Than Expected

Benzinga11:29

Is panic over U.S. Treasury sell-offs beginning to spill over? AI darlings plunge in unison as Wall Street grows jittery...

① On Tuesday, although long-term U.S. Treasuries halted their sell-off and rebounded as the New York trading session began, the pervasive sense of panic among Wall Street traders clearly failed to dissipate... ② The wave of selling triggered by the sharp plunge in U.S. Treasuries is likely to continue driving up borrowing costs, which could simultaneously place heavy pressure on U.S. households, corporate operations, financial markets, and the federal budget.

cls.cn08:53

U.S. Market Close | Concerns over AI debt financing intensify, leading to a third consecutive day of declines for major indices; the chip index plunges 6%. The pharmaceutical and high-dividend sectors rotate into strength, with Johnson & Johnson hitting a

The S&P 500 fell 0.67%, the Dow Jones Industrial Average dropped 0.22%, and the Nasdaq Composite declined 1.33%, marking the third consecutive trading session of losses for the three major indices. The Philadelphia Semiconductor Index plunged 6%, with memory and optical communication chips leading the decline, while the energy and healthcare sectors bucked the trend to lead gains, with Exxon Mobil rising 2.58%. The yield on the 10-year U.S. Treasury note closed at 4.7060%, and the 30-year bond yield settled at 5.2858%, retreating approximately 5 basis points from their intraday highs.

wallstreetcn06:54

Top 20 by Trading Volume | AI infrastructure stocks retreat broadly, with SanDisk and LITE both dropping over 9%; Apple bucked the trend with a 1.5% gain as it prepares to lower European App Store fees; Meta’s landmark lawsuit officially goes to trial

Top U.S. Stock by Trading Volume on Tuesday: Closed down 7.02% with a trading volume of $34.784 billion. On Tuesday, the three major U.S. stock indices closed lower collectively. As market concerns over persistent inflation mounted, sovereign bond yields surged to multi-decade highs. A pullback in the semiconductor sector weighed on the broader market. Significant declines were seen in memory storage, optical communications, and AI cloud services sectors: SanDisk, SK Hynix, and Seagate Technology fell more than 9%; Western Digital and Micron Technology dropped over 7%; Coherent declined more than 12%, Lumentum fell over 9%, and Corning dropped more than 7%; CoreWeave slid over 12%, NEBIUS fell more than 7%, and Applied Optoelectronics dropped over 8%.

Futu News05:42