bubble boi
Twitter

bubble boi

by bubbleboi

377 tweets

head of risk @thru_xyz.
Investment Summary
Updated 51 minutes ago
Summary of insights from content in the last 30 days

AI Compute And Packaging

AI demand is running into foundry, packaging, and power constraints, creating openings for alternative capacity and advanced packaging. The bullish capacity story is balanced by concerns that AI financing and future supply growth could pressure economics.

  • Intel (INTC): Sold-out Xeons through 2027, strong 14A interest, and a potential advanced-packaging ramp from late 2027 support its role as a capacity alternative.
  • NVIDIA (NVDA): Bank of America’s $350 price objective and roughly 70% FY28/CY27 sales-growth outlook contrast with financing concerns and emerging inference-chip competition.
  • TSMC (TSM): CoWoS and co-packaged-optics constraints could redirect opportunities toward Intel EMIB, though Intel’s projected market dominance remains a thesis, not a certainty.
  • Qualcomm (QCOM): AWS custom-silicon collaboration and AI200/AI250 launches broaden its data-center ambitions; accelerator commercialization is slated for 2026–27.

Memory And Storage

AI memory demand is lifting shipment and pricing expectations, but skeptics warn that memory makers may be near cyclical peak earnings. Flash is also emerging as a potential lower-cost complement to DRAM for AI workloads.

  • Micron (MU) and SK Hynix (SKHY): Citi sees revenue and bit shipments growing through 2031, while unusual short-dated MU call activity signals near-term catalyst speculation.
  • Kioxia (285A): CXL-connected NAND could expand AI memory capacity and improve performance, supporting the flash-over-DRAM thesis.
  • Samsung (005930.KS) and EWY: Foundry turnaround expectations and broader Korean semiconductor exposure offer additional ways to participate in AI demand.

AI Infrastructure

Cloud and accelerator demand is supporting higher rental prices and ambitious custom-chip programs, but capital intensity and execution remain material risks.

  • Nebius (NBIS): GPU rental rates rise October 1; deliberately short contracts may preserve pricing power as demand visibility extends toward early 2029.
  • Oracle (ORCL): Cloud infrastructure sales jumped 121% to $7.4B, reinforcing its position as an AI infrastructure beneficiary, alongside customer-concentration and project-delay concerns.
  • Amazon (AMZN): Its multi-year AWS silicon partnership with Qualcomm targets AI inference, with first Qualcomm data-center silicon expected in 2028.

Energy And Cyclicals

Potential Saudi delivery disruptions have sharpened the energy-supply risk, while some market participants see diesel and jet fuel as more direct exposure than crude alone.

  • Heating oil / diesel (HO): Highlighted as a direct way to express disruption risk; reports of canceled September and October Saudi oil deliveries could push crude above $120 if confirmed.
  • Jet fuel (JETF): Also cited as a targeted beneficiary of refinery disruption, though the thesis depends on sustained supply impacts.

AI-generated summary. Not investment advice. Learn more.

Ask about bubble boiAnswers are grounded in this source's posts from the last 30 days.

Latest Content

377 posts

Frequently asked about @bubbleboi

What does @bubbleboi talk about on Kazuha?

Kazuha indexes 377 posts from @bubbleboi, with AI-extracted insights covering 228 distinct assets (stocks, ETFs, cryptocurrencies, and other investable assets).

Which assets does @bubbleboi cover the most?

@bubbleboi's most-discussed assets on Kazuha are NVDA, INTC, TSM, MU, GOOGL. See the "Top assets covered" section above for the full breakdown with sentiment.

Is @bubbleboi bullish or bearish right now?

Mostly bullish. In the last 30 days, @bubbleboi had 69 bullish, 38 bearish, and 14 neutral takes across all assets they discussed (per AI-extracted sentiment scoring on Kazuha).

Where does Kazuha get @bubbleboi's insights?

@bubbleboi's publicly available content (podcast episodes, YouTube videos, or X/Twitter posts) is transcribed and analyzed by an LLM that extracts the assets discussed and the speaker's sentiment toward each one. Each insight links back to the original source.