WhatLLM

What LLM? Any LLM. For AI & Semis.

[WhatLLM @WhatLLM]

Anthropic has reportedly filed confidentially for an IPO. Founded in 2021 by former OpenAI executives Dario and Daniela Amodei, it is riding the AI boom. There is no market without AI. The Goldman Sachs TMT AI Index is up +42% YTD, nearly 4x the S&P 500’s +10.7%. $DRAM $MU $ARM

[WhatLLM @WhatLLM]

The constellation of logos behind Jensen Huang highlights the deep integration of the global #AI and #SEMI supply chain. From chips to software, leading brands form a powerful ecosystem moat. From @Brand #BrandAI’s perspective, this is a masterclass in ecosystem branding.

[WhatLLM @WhatLLM]

NEW: I was reading the $SIVE earnings call transcript and somehow ended up reading $AAOI and $AXTI too. I had to check the numbers twice because I couldn’t believe it. Either investors know something management doesn’t, or management is selling while telling a different story.

[Brand @Brand]

$SIVE: During the earnings call, Sivers CEO Vickram Vathulya and CFO Heine Thorsgaard appeared to selectively address investor questions in real time, raising concerns about transparency. This alone does not prove wrongdoing, but BrandAI’s long-term tracking of distressed and controversial companies suggests that controlled communications and limited media engagement are recurring warning signs. Executive behavior like this can undermine corporate credibility, weaken brand trust, and ultimately pressure valuation.

[WhatLLM @WhatLLM]

$SIVE 799M pipeline is a bloated illusion with obscured metrics. Management hides vital yield data while the company faces a criminal insider trading probe by ECA prosecutor Jonas Myrdal. Sivers $SIVE admits prior financial restatements and warns of weak Q2 revenues ahead.

[WhatLLM @WhatLLM]

$ORCL LARRY ELLISON: MOST LLMS ARE HEADING TOWARD COMMODITY STATUS. THEY FEED ON THE SAME PUBLIC INTERNET AND ARE CONVERGING. THE NEXT AI WAR WON’T BE WON BY BETTER MODELS. IT WILL BE WON BY WHO OWNS DATA NOBODY ELSE CAN ACCESS. PROPRIETARY DATASETS MAY BE THE LAST REAL MOAT.#LLM

[WhatLLM @WhatLLM]

Memory supercycle = $NVDA? Would be weird if it was flat/declining. One important note is memory often scales log(n) to n while compute usually scales n to n^2. Basically compute demand grows faster. Compute scales exponentially, but memory capacity bottlenecks throughput.

[WhatLLM @WhatLLM]

$SIVE Earnings Bearish! 3 Unresolved Risks:
1. Weak Fundamentals: Revenue fell 22% and losses widened.
2. Hype vs. Reality: AI and the NASDAQ listing fueled the rally, but execution remains unproven.
3. Exit Risk: A future market-maker exit could trigger significant downside.

[WhatLLM @WhatLLM]

AI and semi leaders remain strong into the close.
$ARM continues to lead as a core AI CPU architecture play, showing strong relative strength versus the broader market.
$SMCI is extending its rally on robust AI server demand, supported by strong volume and momentum.

[WhatLLM @WhatLLM]

NEW: HUAWEI’s Tau (τ) Scaling Law proposes a new semiconductor roadmap targeting 14 Å (1.4 nm)-level transistor density by 2031. The τ Law is a bold attempt to bypass lithography bottlenecks, but overcoming 3D physics and building an independent ecosystem remain major challenges.

[WhatLLM @WhatLLM]

NEW: Anthropic CEO Dario Amodei says software engineering could be automated within 12 months. AI is shifting from hiring teams to deploying prompts. If you still use tools like Claude LLM for basic tasks, you’re underusing a powerful. Misuse wastes time, money, and opportunity.

[WhatLLM @WhatLLM]

AI bubble rivals historic peaks, with market concentration near 40%. Yet this cycle reflects a structural shift: unlike 1999 speculation, it’s backed by real productivity gains. If AI moves from capital absorption to value creation, the bubble could become a durable foundation.

[WhatLLM @WhatLLM]

Here are the top 5 stocks I hold in the NASDAQ-100 that have more than doubled so far in 2026: Sandisk $SNDK +523%🟢, Intel $INTC +225%🟢, ARM $ARM +180%🟢, Micron $MU +163%🟢, and AMD $AMD +118%🟢. Wall Street is set to welcome three AI giants: OpenAI, SpaceX, and Anthropic.

[WhatLLM @WhatLLM]

LLMs dominate AI, replacing fragmented tools with general intelligence from text. Leaders include OpenAI, Google, and Anthropic, alongside rivals. This race favors tech giants, and most models fade. Outputs remain probabilistic, requiring verification.

[WhatLLM @WhatLLM]

Donald Trump at a Suffern, New York rally: “Micron Technology is great.”Is this what $MU whales were waiting for? Trump highlighted Micron’s major investment in Clay, NY as part of the U.S. chip and AI push. Another $MU “buy signal” right before a sharp, insider-style move. #AI

[WhatLLM @WhatLLM]

UPDATE: the AI bubble math doesn’t add up. Anthropic spends $3 to make $1, excluding major costs. Microsoft invested $300B capex for ~$18B AI revenue. Startups like OpenAI dominate cloud backlogs. Enterprises burn budgets with little ROI. This experiment risks overcapacity…

[WhatLLM @WhatLLM]

This is insane: The pure-play memory ETF, ⁠DRAM⁠, has amassed $9.67 billion in AUM and $8.6 billion in inflows since its April 6 launch. It has taken in assets faster than any ETF in history, breaking the record previously held by the Bitcoin ETF, IBIT⁠.
$DRAM $SNDK $MU $SOXX

[WhatLLM @WhatLLM]

@WhatLLM Q2 2026 rankings show OpenAI’s GPT-5.5 leading reasoning, Anthropic’s Claude 4.7 excelling in coding, and Google’s Gemini 3.1 dominating multimodal. xAI’s Grok 4.3 leads real-time intelligence, while DeepSeek V4 disrupts markets with top-tier performance at lower cost.