Sacks says DeepSeek moved perceptions of China's AI lag from six-to-twelve months to three-to-six months.
“if you had asked most people in the industry a few weeks ago how far behind is China on AI models, they would say six to twelve months.”
Sacks says people would have been surprised a Chinese company released the second reasoning model after OpenAI.
“if you had said to people a few weeks ago that the second company to release a reasoning model along the lines of o one would be a Chinese company, I think people would have been surprised by that.”
Sacks estimates 90% of AI revenue goes to OpenAI, giving them dominant consumer position.
“And if you look at AI revenue, I think something like 90% of it is going to OpenAI. I mean, they have the dominant position among consumers.”
Sacks states AI represents 40% of US GDP growth and 80% of stock gains this year.
“40% of US GDP growth this year is AI and AI companies have accounted for 80% of the gain in US stocks this year.”
Sacks estimates each gigawatt data center requires roughly 50 billion dollars of investment in chips, land, and power.
“Here's a rule of thumb is that each gigawatt data center is about $50,000,000,000 of investment, roughly. Between the chips, the land power shell, it's about 50,000,000,000 per gigawatt.”
Sacks argues AI will increase demand for knowledge workers by making code cheaper, following Jevons Paradox.
“So the amount of software generated in the economy was limited by that. I think it's gonna increase massively now because the cost of generating code is coming down so much.”
Sacks contrasts limited AI copilots within individual SaaS tools with cross-platform AI tools like Claude that span multiple products.
“All these SaaS products are rolling out like AI copilots inside their tools. And some of them work pretty well, but they're limited to playing in that sandbox.”
Sacks argues established SaaS products have millions of bug reports and enterprise testing that AI-generated code cannot replace.
“Think about how many bug reports have been filed on Salesforce's code base over the last twenty five years, maybe millions of them. That system has been tested across thousands of large customers and enterprises.”
Sacks claims Anthropic has the fastest growing revenue run rate at scale ever seen based on public data.
“Third, Anthropic's revenue run rate is based on what I can tell and what's been publicly released is the fastest growing revenue run rate at scale that I think we've ever seen.”
Sacks predicts Anthropic may regret supporting data center nimbyism now that they need to build their own.
“But now that they're, I think, gonna have to move into the game of building their own data centers, they might regret salting the earth for data centers all over the country.”
Sacks says Anthropic grew 10x annually to $30B ARR versus OpenAI's 3-4x, projecting $80-100B by year-end.
“So they went from, let's call it 1 to 10,000,000,000 of ARR last year. And by the end of Q1 this year, they were already at 30,000,000,000 of again, let's call it their revenue.”
Sacks says Anthropic grew from $10B to over $80B ARR and will hit $100B forecast early.
“Started the year at 10. It had forecast a 100,000,000,000 as exit ARR for the year, and most people said that that would be impossible to achieve.”