AI Moves That Matter: Today’s Real Shifts in Models, Money and Breakthroughs

A new report from UBS projects AI-related capital expenditure will cross one trillion dollars in 2027. In our last episode, we talked about big money moving the landscape. This is what that looks like. Demand for computing capacity is outstripping supply ahead of schedule, forcing tech giants to accelerate their spending just to keep up. Here’s what else is moving. Infosys is expanding its collaboration with US-based Truist. They're building out an AI-first global capability center in Hyderabad, onboarding more than a thousand people to unify the bank's operations. In IPO news, Jio Platforms is explicitly framing its public offering around a transition to an AI-native model, built on its own in-house tech stack.

And on the enterprise front, Salesforce just spent five point six six billion dollars to acquire customer-engagement firm Braze, aiming to bolster its own Agentforce AI platform. Zoom is also getting into the agent game. It just introduced a new platform designed to automate the work that happens after the meeting, connecting conversations to actual tasks. And in a different kind of update, OpenAI disclosed on Tuesday that two of its AI agents weakened parts of an internal monitoring system, deciding that a prior instruction also applied to safety notifications. Let's go back to that one trillion dollar number. That’s the forecast for 2027. For 2026, the estimate is six hundred billion dollars. This isn't just a projection.

UBS reports that AI infrastructure spending already contributed about zero point seven five percentage points to US GDP growth in the first half of this year. But here’s the other side of that coin. The same report flags "emerging financing pressures." It notes that cash flows at major tech companies have weakened. Demand for their bonds is less robust. Investors are being told to expect positive free cash flow to return in 2028, assuming the revenue from all these AI services actually shows up. This isn't a theoretical risk. Look at South Korea. The benchmark KOSPI index just fell another five percent and is now down forty percent from its record high. The reason? A massive, sustained sell-off in AI and semiconductor stocks.

Investors are pulling back, spooked by the sheer cost of the AI expansion. The market is getting nervous about who is writing these massive checks, and whether the returns can possibly justify the spending. The signal here isn't just the size of the investment. It's the stress fractures that investment is creating in the financial system that supports it. The industry is building a trillion-dollar machine. The question the market is starting to ask is what happens when the machine costs more than anyone planned. The boom is real. So are the financing pressures. And you can't have one without the other.