gavinsbaker
I publish concise, market-focused commentary on AI infrastructure and semiconductor implications. Emphasis on inference economics, demand dynamics, and how these trends affect public-market leaders such as NVDA, TSM, and AVGO.
Past bets that played out
Notable calls emphasize that AI inference economics appear robust: posts argue that Anthropic is generating very high (~80%) gross margins on inference, undermining last year’s ‘circular demand’ thesis. Another highlighted post asserts that ‘orbital compute’ has already been used successfully for both training and inference, citing NVDA as the sole explicit public-market reference. None of these posts include explicit trade recommendations, position sizing, timing, or catalyst details.
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.
Post argues that the prior (last year’s) notion that AI demand is “circular” is no longer defensible, citing Anthropic allegedly generating ~80% gross margins on inference—implying inference economics are strong enough to sustain/expand demand rather than being self-limiting.
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.
What this channel is watching now
Focus areas: AI compute and inference economics, demand sustainability, and semiconductor supply-side implications. Top-mentioned tickers: NVDA (2 mentions), ANTHROPIC (1), TSM (1), AVGO (1).
Latest videos and market context
No recent videos are listed.
Gavin Baker @GavinSBaker 28m New Pareto Frontier: Grok 4.5, SWE-1.7 and Opus 5. Intelligence per $ will be the only m...
Post argues a new “Pareto Frontier” of AI models (Grok 4.5, SWE-1.7, Opus 5; later K3) and claims “intelligence per $” will be the dominant metric over time—implying competition shifts toward cost/performance efficiency and away from pure capability bragging rights. No explicit public tickers or trade instructions.
Gavin Baker @GavinSBaker 21m Awesome to have Jensen on X! Jensen Huang @JensenHuang 43m For my first post, I’m sharin...
Post is mainly a positive signal/narrative around Nvidia (Jensen Huang) promoting open AI models as important for innovation, safety, and national “sovereignty.” It’s not a direct product/earnings/catalyst call, but it reinforces Nvidia’s positioning as a central platform provider for broad AI adoption.
Gavin Baker @GavinSBaker Jul 21 American open source FTW! Nice job @poolsideai Crazy metrics for 118b parameters. I e...
Post praises PoolsideAI’s 118B-parameter open-source model and expects it to be on the “Pareto frontier” (implying strong efficiency/quality tradeoff). No public-market tickers or tradable instruments explicitly mentioned; largely narrative/technology context.
Pinned Gavin Baker @GavinSBaker 26m Kimi K3 may be an important inflection point for AI. Potentially negative for Ant...
Post claims Kimi K3 could be an AI inflection point, potentially negative for Anthropic/OpenAI, and broadly positive for “essentially every other company,” with the bigger catalyst being an open-source frontier model (“Sputnik moment”). No public tickers are directly mentioned; implications to public markets are indirect via OpenAI/Anthropic partners and broader AI commoditization.
Proof-backed call history
Published short-form market commentary and responses on platform X. Recent posts mix technical claims about AI compute and brief social interactions; most contain commentary rather than explicit investment instructions.
Gavin Baker @GavinSBaker 28m New Pareto Frontier: Grok 4.5, SWE-1.7 and Opus 5. Intelligence per $ will be the only m... Gavin Baker @GavinSBaker 28m New Pareto Frontier: Grok 4.5, SWE-1.7 and Opus 5. Intelligence per $ will be the only metric that matters over time. K3 probably joins the frontier onc...
Post praises PoolsideAI’s 118B-parameter open-source model and expects it to be on the “Pareto frontier” (implying strong efficiency/quality tradeoff). No public-market tickers or tradable instruments explicitly mentioned; largely narrative/technology context.
...nned Gavin Baker @GavinSBaker 26m Kimi K3 may be an important inflection point for AI. Potentially negative for Ant... Pinned Gavin Baker @GavinSBaker 26m Kimi K3 may be an important inflection point for AI. Potentially negative for Anthropic and OpenAI while being net positive for essentially every other company in the world. I mean that very literally. Although the real “Sputnik moment” would be an open-source frontier model that Show more Pinned Gavin Baker @GavinSBaker 26m Kimi K3 may be
Pinned Gavin Baker @GavinSBaker 26m Kimi K3 may be an important inflection point for AI. Potentially negative for Ant... Pinned Gavin Baker @GavinSBaker 26m Kimi K3 may be an important inflection point for AI. Potentially negative for Anthropic and OpenAI while being net positive for essentially every other company in the world. I mean that very literally. Although the real “Sputnik moment” would be an open-source frontier model that Show more Pinned Gavin Baker @GavinSBaker 26m Kimi...
Post claims Kimi K3 could be an AI inflection point, potentially negative for Anthropic/OpenAI, and broadly positive for “essentially every other company,” with the bigger catalyst being an open-source frontier model (“Sputnik moment”). No public tickers are directly mentioned; implications to public markets are indirect via OpenAI/Anthropic partners and broader AI commoditization.
Post claims Kimi K3 could be an AI inflection point, potentially negative for Anthropic/OpenAI, and broadly positive for “essentially every other company,” with the bigger catalyst being an open-source frontier model (“Sputnik moment”). No public tickers are directly mentioned; implications to public markets are indirect via OpenAI/Anthropic partners and broader AI commoditization.
Post claims Kimi K3 could be an AI inflection point, potentially negative for Anthropic/OpenAI, and broadly positive for “essentially every other company,” with the bigger catalyst being an open-source frontier model (“Sputnik moment”). No public tickers are directly mentioned; implications to public markets are indirect via OpenAI/Anthropic partners and broader AI commoditization.
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.
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.
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.
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.
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.
About this channel
Contributor on X covering AI infrastructure, inference economics, and semiconductor market implications. Content is analytical and often challenges prevailing narratives about demand dynamics in AI.
@gavinsbaker
Most recognized assets
Unlock the full track record
Follow @gavinsbaker on X for concise, debate-focused commentary about AI compute, inference economics, and semiconductor implications.
10 more thesis calls are available after sign-up.