VC10X Pulse - AI Slowdown? - What it really means

For the last few years, AI has been defined by acceleration, bigger models, more GPUs, more data centers, and increasingly massive capital investment. Now we are hearing more about an AI slowdown, but what exactly is slowing? Is it model development? Is it AI infrastructure spending, revenue growth, or simply investor expectations? Because these are very different things. Let's find out. AI spending isn't actually slowing yet. Global AI spending is expected to reach around $2.7 trillion in 2026, up roughly 50% year-over-year according to Gartner. And Nvidia's latest numbers tell a similar story. Nvidia generated $96.2 billion in quarterly revenue, up 106%, while data center revenue reached $89 billion, up 117%. So we are clearly not looking at collapsing AI demand. The real question is, can AI continue growing fast enough to justify the enormous amount of capital being deployed? Now, AI infrastructure spending has become enormous. Hyperscalers are investing hundreds of billions of dollars into data centers, GPUs, networking, and power. And increasingly, that investment involves debt and other forms of financing. So the question isn't simply whether AI demand is growing, it's whether the returns on that capital will be high enough. If AI infrastructure grows 40%, but investors were expecting 70%, the technology can still be growing rapidly while AI stocks and valuations come under pressure. That's the difference between growth slowing and the AI thesis breaking. A slowdown could create winners and losers. A more selective capital environment doesn't affect everyone equally. Companies like Nvidia and major hyperscalers have enormous cash flows and access to capital. But highly diversified infrastructure companies and startups that require massive amounts of funding could face much more pressure. Capital doesn't necessarily disappear, it gets more selective. And investors start asking, what is the revenue? What are the margins? How much is the infrastructure actually being utilized? And what's the return on invested capital? Frontier AI could also slow without AI adoption slowing. There's another important distinction. Some AI leaders are now discussing whether the pace of frontier model development should slow so that safety measures can keep up. But slow model progress doesn't necessarily mean slower AI adoption. The next phase could be less about how much bigger can we make the model and more about how much economic value can we generate from the models we already have. That means inference, AI agents, enterprise adoption and automation become increasingly important. Now to the bigger question. I don't think the evidence points to an AI collapse. AI spending and infrastructure demand remain extraordinarily strong but the standard for investment is changing. The market is moving from AI will transform everything to show me the returns and that's probably the most important shift to watch because AI doesn't have to stop growing for the AI trade to slow down. It simply has to grow more slowly than the expectations already priced into the market. That's all for this week. We'll be back next week with another story from the front lines of global