We Went All In: The Part That Isn’t on YouTube
Six off-stage conversations test what power, politics and financing mean for the AI buildout.
Independent research for investors and institutions
We connect models, infrastructure, and company economics to identify the forces shaping long-term investment.
Original analysis, research tools, and a broader view of the AI economy.Latest research
Six off-stage conversations test what power, politics and financing mean for the AI buildout.
Read best config against best config, SemiAnalysis's own dashboard gives NVIDIA 9.7x outperformance on the TPU it just crowned.
The reported Hugging Face deal makes NVIDIA the Compute Patron of open source AI.
Latest interview · Full conversation
Ben Pouladian joins ETN’s Luke Knight and Ronan Chambers to discuss NVIDIA, open models, cost per task, and AI’s next inflection in the lab.
15 minutes · Recorded conversation
More interviews & videos ↗· BEP Research on YouTube · Watch on YouTube ↗
Compute Market Brief · Public sample
Explore the reported Nebius increase, inspect the dated sources, and calculate the budget sensitivity for your GPU-hours.
One idea, made explorable
Follow the experimental loop through nine supply-chain layers, from model design to real-world measurement, and explore the companies behind them.
Explore the AI Biolab ↗Illustrative preview · opens the interactive toolThe perspective behind BEP
BEP Research is led by Ben Pouladian, an electrical engineer, entrepreneur, and investor who co-founded Deco Enterprises and led it through its 2019 exit. BEP connects developments at model companies such as OpenAI and Anthropic with their implications for chips, compute, power, and capital—building a broader view of where the industry is going. Specialist collaborators include Rui “Rick” Xie, who contributes memory-systems engineering and standards analysis.
Meet the people and explore the work ↗Where we’ll be
Stay close to the research
Get BEP’s analysis of AI infrastructure and its economics in your inbox.