Keep pulling the thread on United States.
Anthropic and OpenAI are adding more revenue per month than established technology companies like Meta, Google, or Microsoft.
David predicts that the combined revenue run rate of Anthropic and OpenAI could reach $200 billion by the end of the current year.
The threshold for a top 1% venture capital exit has increased from $10 billion in the 2020-2024 period to $32 billion currently, with the exit of Wiz setting the new benchmark.
The host predicts that the threshold for a top 1% venture capital exit could surpass $100 billion by September, driven by potential future exits of OpenAI and Anthropic.
David predicts that the valuation of any of the three large anticipated AI IPOs will likely be larger than the combined $1 trillion market capitalization of all VC-backed IPOs from the last six years.
New data center capacity at scale is unlikely to be available until late 2028 or early 2029 due to massive supply constraints.
The total market capitalization of all venture-backed IPOs over the last six years is slightly over $1 trillion.
The current diffusion of AI technology into the real economy is estimated to be less than 5%.
The buildout of data centers in the United States is estimated to be approximately one year behind schedule.
Leading large language models in China are estimated to be about six months behind their U.S. counterparts in capability but are 10 times cheaper.
The per-token cost for comparable AI models is decreasing by more than 10x year-over-year.
The cost to distill a large language model is estimated to be approximately 2% of its original pre-training cost.