Meta Inflates AI Bubble
Welcome to MktContext! I am a professional money manager, trader, and investor who has been timing and beating the market for over a decade. We specialize in predicting market direction by studying the economy and market signals. Join 13,000 subscribers at MktContext.com for our weekly deep dives and analysis!
Last week we discussed Meta’s decision to rent out excess computing capacity. Our view was the market had misread the announcement; Meta would continue to spend profligately on AI infrastructure.
This week, Meta clarified their plans to double computing capacity in 2027. They are also designing their own in-house AI chip, aimed at reducing reliance on Nvidia and AMD. It’s a heavy toll that will ensure their spending habit continues to increase by the hundreds of billions.
The strategic importance cannot be understated. As Meta owns more of the chip stack, AI training and inference gets cheaper over time. Meta’s massive scale means they can build data centers at much lower cost than the industry average.
That leads to their latest advanced AI coding model Muse Spark 1.1, which is not only comparable to the best that OpenAI or Anthropic have to offer, but also costs one-quarter of the price.
This is a big improvement for Meta. Recall they had previously spent a shocking $14B to poach Alexandr Wang and offered multi-hundred-million-dollar pay packages to top AI engineers… all to produce an extremely underwhelming language model last year (Llama 4):
No doubt this is an improvement in Meta’s fundamental story, as the stock price confirms. But take a step back from the trees to see the forest: Meta is another in a long line of US and foreign competitors making AI models. SpaceXAI already joined the fray with their blockbuster IPO and 50% model price cut. Now, Meta is cutting it another 50%. The industry is hurtling toward commoditization wherein prices are a race to the bottom and differentiation is non-existent.
Commodity producers (think oil companies and mining companies) typically are not valued highly by the market.
The longer term implications will become clearer when OpenAI and Anthropic come to market with their own IPOs (slated for early 2027). It will be harder and harder for them to justify their exorbitant spending, from a profitability perspective. For example, the market already doubts OpenAI’s revenue-generating capabilities.
The AI bubble continues to sow the seeds of its own demise.
Get the rest of this article, including our portfolios and trades, at MktContext.com!
Join 13,000+ investors who are timing the market!