Morgan Stanley has a view that sharply cuts its cumulative shipment forecast for AMD’s AI GPUs in 2026–2030. The story is that it was reduced from roughly 14.1 million units to around 8 million, a drop of a little over 40%; Nvidia’s outlook, by contrast, was left relatively unchanged. The reasons are basically power, supporting infrastructure, and ecosystem constraints.
This kind of research note may not immediately become reality, but it’s a very good sentiment catalyst for the week: the market was already wondering whether everyone can really share in the AI feast.
AMD is still emphasizing strong demand and the need to expand capacity. The sell side is cutting estimates while the company is talking it up, so in the short term it’s easy for people to argue. For me, the key is not to treat a single broker model as gospel.
Names with more stable power and stronger positions in the supply chain tend to hold up better against volatility; players further down the chain tend to be more volatile. Position sizing matters more than guessing who’s right.
When sell-side models and company statements clash, stock prices often follow sentiment first. Your job isn’t to pick a side—it’s to control the weight of each position.
The market is reassessing how many companies in the AI supply chain can actually share in the feast. Those counted out will be more volatile—that’s normal.
Treat expectation gaps as a risk-management issue, not a matter of pride or winning and losing.
Trading on expectation gaps can be thrilling, but it can also easily turn into a battle of egos. Position sizing is the only antidote.
Ultimately, today’s market action looks more like an attempt at sentiment recovery than a trend reversal. You can take part if you want, but come prepared with a stop-loss and a position-size limit.
The above is for reference only and does not constitute investment advice.
Follow me, haha