AI Chip Market

Key takeaways include: Retail earnings are scrutinized ahead of Black Friday, Broadcom challenges Nvidia in AI chips, James Comey’s case was dismissed, Amazon invests heavily in AI infrastructure for the government via AWS and competes with SpaceX’s Starlink with “Leo,” NHL team values rise with the Toronto Maple Leafs leading, a Microsoft data center project in Wisconsin was abandoned due to community opposition.

TSMC reported a strong Q3, driven by the thriving AI chip market. Profit surged 39.1% year-over-year, with revenue exceeding expectations at NT$989.92 billion. Net income also surpassed estimates, reaching NT$452.3 billion. The company’s HPC division, fueled by AI and 5G infrastructure demands, was a key growth driver. Advanced chips using 7-nanometer or smaller technologies comprised 74% of wafer revenue, highlighting TSMC’s technological lead in miniaturization. TSMC’s performance underscores its critical role in the global semiconductor supply chain and the AI revolution.

AMD is challenging Nvidia’s AI chip dominance with a potentially multi-billion dollar GPU deal with OpenAI. This partnership aims to diversify OpenAI’s supplier base and mitigate antitrust concerns surrounding Nvidia’s 90% market share. AMD’s stock surged following the announcement, reflecting investor confidence. The deal is expected to generate significant revenue for AMD and positions the company as a stronger competitor in the AI hardware market, attracting regulatory attention amid concerns about market concentration and potential tariffs.

Marvell Technology (MRVL) shares fell 15% after a weaker-than-expected data center revenue report and cautious Q3 outlook, despite exceeding EPS estimates. While Q2 revenue grew 58% YOY, driven by AI demand, Q3 revenue guidance of $2.06 billion fell short of expectations. Investors are seeking clarity on customer engagements, and Bank of America downgraded MRVL’s stock, citing concerns about near-term AI growth. The company anticipates a stronger Q4, attributing Q3’s flatness to “nonlinear growth” in custom AI chips.