SMCI · Super Micro Computer, Inc.

Bloomberg Businessweek Daily discusses: (1) escalation risk around Iran/Hormuz with Trump threatening strikes on energy targets near Tehran if Iran attacks shipping; implications for oil prices and inflation; (2) expected new US tariffs Friday; (3) OpenAI “accidental hack” of Hugging Face framed as less alarming; (4) AI’s impact on Auto/Aviation/Defense and an “industrial revolution” narrative; (5) market mentions of chip stocks, Tesla, Alphabet, Super Micro, plus AT&T and Nike.

US equity futures are down ahead of Alphabet earnings amid broader big-tech caution/rotation. Brent crude is above $95 (highest in ~6 weeks) as US/Iran downplay talks. Trump signals a policy push to force generic drug manufacturing onshore via a proposed 100% import duty. Japan’s yen hits a four-decade low; Bank of Japan considers faster rate hikes. Mentions of AI/data center investment and ‘AI winners,’ plus early movers: Super Micro surges while IT is weak and drugmakers face pressure.

Geopolitical escalation risk in the Middle East (Iran/Red Sea) is supporting oil prices and can spill into defense, shipping, and inflation expectations. Separately, tech momentum persists (AI hardware demand cited via SMCI), and industrial aerospace cycle commentary (GE). Policy risks include potential new tariffs aimed at generic drug manufacturers. Japan yen weakness and South Korea market controls are notable for FX/EM positioning but are less directly tradable from this snippet alone.

TSMC frames AI compute growth as increasingly constrained by power/thermal limits (“power wall”), arguing that continued AI proliferation depends on energy-efficiency innovations across the semiconductor ecosystem. This is a high-level narrative piece without specific product, capex, guidance, timelines, or quantified financial impact for any company beyond broad industry trends.

Artificial Analysis reports that Kimi K3 (Moonshot) ranks #2 on the AA-Briefcase agentic knowledge-work benchmark (behind “Fable 5”) but is expensive to run—costing more than “Opus 4.8” while taking ~1 hour per task on average. Moonshot released Kimi K3 last week; it is described as a 2.8T-parameter model and scores 57 on an Artificial Analysis metric.

Bloomberg “The Close” episode framed a late-day market narrative around (1) a rebound gathering pace in chipmakers/AI spend, (2) the idea that value stocks and financials may be underappreciated beneficiaries of AI capex, (3) company-specific updates including Amazon Business scale, GM raising outlook despite tariffs, and (4) notable movers/laggards (Danaher, Schwab, Super Micro) plus a near-term Tesla earnings preview. The source is light on hard numbers, so actionability is mainly thematic/sec

A new Chinese open-source model ("Kimi K3") reportedly triggered a sharp selloff in AI/tech names by raising fears that China can rapidly close the model-capability gap via distillation/IP copying. The episode frames the key debate as: (1) are model labs’ moats eroding due to open source/cheap replication, and (2) regardless of who leads in models, does demand for compute/infrastructure (GPUs, networking, data-center buildout, hyperscalers) continue to win over the long term. The piece leans tow

Bloomberg TV segment highlights perceived acceleration in China’s AI model progress (Alibaba, Moonshot, 01.AI) and a narrative shift from AI infrastructure spending toward AI applications, plus a separate geopolitical/oil inflation-risk segment (US strikes Iran) and an India private space milestone (Skyroot orbital launch). Actionability is moderate because it’s thematic without hard datapoints, but it supports relative-positioning trades: China AI/app software beneficiaries vs US infra names if

Kimi.ai (Moonshot) says its Kimi K3 demand over the last 48 hours is near capacity limits; to protect existing subscribers it is temporarily pausing new subscriptions. This is a datapoint of strong AI inference demand but also highlights near-term GPU/compute scarcity and potential revenue throttling for AI app providers without enough capacity.

Post argues a “mega bull case” for AI infrastructure would materialize if inference market share shifts from high-margin frontier labs toward cheaper models (open-source or closed), improving end-customer ROI by increasing “intelligence per $”. This implies higher inference adoption/volume and thus stronger demand for compute/networking infrastructure. No explicit tickers mentioned; implications are thematic across AI infra supply chain.

BIS (U.S. Commerce) guidance reiterates/clarifies that a license is required to export “advanced computing items” to entities headquartered in Country Group D:5 or Macau (including when the receiving entity is located outside those jurisdictions, or where the ultimate parent is headquartered there). This raises compliance friction and potential shipment restrictions for high-end AI/advanced compute chips and related systems, increasing downside risk to U.S. semiconductor vendors’ China-adjacent

Arthur Mensch argues enterprises should use open‑source AI models because closed model providers increasingly impose data retention, creating vendor leverage and lock‑in risk. Implication: accelerating enterprise demand for open/portable model stacks, private deployment, and compute/inference infrastructure; relative pressure on “closed, proprietary API-only” model economics (mostly private companies).