Gavin Beck Claims July AI Sell-off Was Purely Market Panic
Gavin Beck, a tech investor managing $5 billion in assets, traveled to Silicon Valley after the severe sell-off in the AI sector in July, believing that the sell-off was entirely due to market panic, with fundamentals and all quantitative indicators accelerating.
He pointed out that there are no negative quantitative indicators in Silicon Valley; the availability of GPUs, rental prices, DRAM spot prices, and token growth are all accelerating, with the renewal price for B200 clusters increasing by 50%-60% within seven months. Open-source models are transferring profit margins rather than reducing computing power demand; NVIDIA's forward P/E ratio is at a ten-year low; long-term memory agreements cannot be easily torn; NVIDIA is rapidly building its cloud business through credit packaging; data centers are creating blue-collar jobs; orbital computing is becoming increasingly real; and under spot re-pricing, cloud giants have sufficient operating cash flow to be self-sufficient.
This is a public judgment from a top tech investor on AI infrastructure, with funds and attention flowing towards the infrastructure layer and spot data. AI-related stocks are under pressure from sentiment and liquidity, while computing power and supply chain fundamentals benefit from accelerating signals.
Source: Public Information
ABAB AI Insight
Beck has long focused on public market technology and AI infrastructure investments, and this on-site research in Silicon Valley addresses the 40%-60% decline in the sector in July. He emphasized that the public market does not see the private data of OpenAI and Anthropic, leading to information asymmetry and extreme volatility.
On the capital path, by stress-testing the supply chain and spot data, he confirmed that profits are shifting from the model layer to GPUs, memory, and electricity. The motivation is to identify buying opportunities that diverge from fundamentals during panic, while also warning of regulatory and public relations risks.
Similar to the reversal path after the 2022 tech stock adjustment, the current AI infrastructure is in a phase where emotional compression coexists with data acceleration.
Essentially, this is a transfer of pricing power: public market sentiment pricing is disconnected from private computing power data, and the mechanism is to re-anchor the value of infrastructure spot and long-term agreements to real consumption.
ABAB News · Law of Cognition
Emotional compression for a month, while fundamentals are accelerating
Open-source takes away profit margins, yet computing power demand is even greater
Spot prices are the true results of stress testing