Keep pulling the thread on Y Combinator.
Google is now pursuing the development of space-based data centers.
The company Gamma announced it reached $100 million in ARR with a team of only 50 employees.
The market is seeing increased competition in the AI chip space, with customers buying from AMD and using Google's TPUs, reducing NVIDIA's sole dominance.
Series B-level AI startups are building model orchestration layers to abstract away specific LLM providers, allowing them to swap models in and out.
The founders of Zephyr Fusion determined that building a fusion reactor in space is viable based on their analysis of the physics and mathematical models.
Y Combinator-backed startup Zephyr Fusion, founded by former national lab engineers, is developing a space-based fusion energy system.
Anthropic has become the most used LLM API among applicants for the Y Combinator Winter 2026 batch, surpassing OpenAI.
A company named StarCloud, which proposed building data centers in space, launched in summer 2024 and was initially ridiculed.
The speaker believes the current over-investment in AI compute is analogous to the telecom bubble, which created an abundance of cheap bandwidth that enabled companies like YouTube to exist.
A trend in 2024 saw AI startups reaching $1 million ARR and raising Series A funding with teams as small as just the founders.
The supply chain for jet engines used for data center power generation is backed up, with lead times of two to three years for new orders.
The speaker asserts that AI scaling laws are log-linear, requiring exponential increases in compute for linear performance gains, which acts as a natural brake on a 'fast takeoff' scenario.