Asian Chip Stocks Extend Losses on Funding Worry, DeepSeek Model
THE SO WHAT
A cheaper, more efficient DeepSeek model plus funding anxiety is forcing a rethink of perpetual up-and-to-the-right AI chip demand. Infra buyers should use this window to renegotiate capacity commitments and pricing—while model efficiency is improving faster than most 2024–2025 plans assumed.
READ THE SOURCE
MORE FROM THE WIRE
Deep & Emerging TechSamsung, SK Hynix Slide Causes $137 Billion Hit to Korea Pension
When a national pension fund takes a ₩182 trillion ($137 billion) mark from chip volatility, it underlines how concentrated AI infrastructure risk has become in a few equities. Operators relying on long-term cheap memory and logic should model scenarios where capex slows or pricing power returns to suppliers faster than expected.
Deep & Emerging TechAnalysis: 795 accounts funded by ADI Predictstreet's own fee-collecting wallet drove 99% of the FIFA-partnered prediction market's activity since late August
When 99% of a prediction market’s volume traces back to the platform’s own fee wallet, you don’t have liquidity—you have a marketing artifact. If you’re building in on-chain markets or partnering with branded “prediction” products, assume you must independently verify user distribution and funding sources before treating any volume as real demand.
Deep & Emerging TechEurope's three biggest industrial companies are cashing in on the AI boom, as Schneider and Siemens expand further into software, while ABB focuses on equipment
The European industrial giants are drawing a clear line: Schneider and Siemens are racing to own the software and orchestration layer, while ABB leans into hardware and equipment. If you sell into factories or infrastructure, expect tighter, more opinionated ecosystems—your integration and data access strategy now matters as much as your feature set.
Deep & Emerging TechSources: Nvidia asked Samsung, SK Hynix, and Micron to make base dies for its custom NVHBM memory on TSMC's 3nm process; Samsung agreed, its first use of TSMC
When Samsung is willing to fab on TSMC for Nvidia’s NVHBM base dies, memory and foundry loyalties are yielding to AI demand gravity. For anyone building on Nvidia’s roadmap, assume tighter coupling between HBM design and leading-edge logic nodes—and more fragility in the memory supply chain.