AI momentum is now carrying the global stock market higher

Global momentum stocks are on track for their strongest two-month outperformance on record, powered by the AI infrastructure trade. The rally is broader than chips, but higher rates and rising power costs could make the next phase harder to chase. Global momentum stocks are running harder than at any point in the modern data set, and AI is the reason investors are still willing to chase them. The AI trade has moved beyond a handful of chip names. It is now pulling global momentum strategies, power companies, data center owners, networking suppliers and large-cap growth indexes into the same current. That is what makes this rally different from the narrower tech runs investors have seen before. According to Bloomberg data published by Business Standard on May 25, MSCI's global momentum gauge has beaten the MSCI All Country World Index by 17 percentage points since the end of March, putting it on track for its strongest two-month outperformance in records going back to 1991. That is not a small rotation. It is a market telling you that AI is no longer being treated as a product cycle. It is being priced as an economic cycle. The easy explanation is Nvidia, and there is truth in that. The company reported record first-quarter fiscal 2027 revenue of $81.6 billion, up 85% from a year earlier, with data center revenue rising 92% to $75.2 billion. It also guided for roughly $91 billion in second-quarter revenue. Those numbers gave investors fresh evidence that demand for AI compute is still outrunning the doubts around valuation, export controls and capital intensity. But the more important point is that Nvidia is no longer carrying the story alone. Advanced Micro Devices jumped to an all-time high earlier in May after forecasting stronger revenue on demand for data center chips. Corning rallied after expanding work with Nvidia on optical connectivity products for AI data centers. Hut 8 surged after signing a 15-year lease tied to a Texas data center campus. These are not all the same business, but they are being valued through the same lens: AI needs physical infrastructure, and someone has to build, power, connect and finance it. Previous technology booms often depended on a small group of consumer platforms or software names delivering enough growth to justify the whole theme. This one is spreading because the bottlenecks are spreading. GPUs matter, but so do memory, networking, cooling, fiber, buildings, substations and reliable electricity. The market has started to understand that every new model, agent and enterprise AI workflow eventually becomes a demand signal for concrete, copper, silicon and power. That is why the winners now include more than the obvious semiconductor leaders. Data center real estate investment trusts such as Equinix and Digital Realty sit closer to the physical demand for capacity. Power equipment companies such as Eaton and Vertiv are tied to the harder question of how facilities stay cool and online. Utilities and grid-exposed names are being pulled into the discussion because the most advanced AI infrastructure is useless without firm electricity. Japan shows how far the trade has traveled. The Nikkei has pushed into fresh record territory during the same cycle, helped by large technology and semiconductor-related shares. In the U.S., large-cap growth indexes have also been making new highs. Europe has not been left out either, as investors look for industrial suppliers and power-linked companies that can participate in the buildout without depending on one software winner. This breadth is encouraging, but it also changes the risk. When a theme spreads across regions and sectors, it can look more durable. It can also become harder to find clean entry points. A chipmaker with explosive earnings is one kind of bet. A regulated utility being re-rated because of future data center load is another. The first is about growth. The second is about rates, capital spending, permitting and whether regulators let companies earn a fair return on the infrastructure they must build. Rates still matter The biggest threat to the momentum trade is not that AI demand disappears overnight. The more realistic risk is that the price paid for that demand gets too high at the same time interest-rate uncertainty refuses to go away. Long-duration growth stocks are sensitive to yields, and infrastructure-heavy AI beneficiaries need capital. Higher rates make both stories harder. There is also an inflation angle that investors cannot ignore. Allianz Research recently estimated that data center investment grew 32% in 2025 and is set to rise another 75% in 2026, with pressure building on exposed utilities as electricity demand catches up with the AI buildout. That is useful for companies selling power equipment and grid capacity. It is less comfortable for households, regulators and businesses that have to absorb the cost. For founders, the message is practical. Public markets are rewarding AI infrastructure and anything that can credibly attach itself to the compute buildout, but that window will not stay equally open forever. If valuations continue to move before private companies reach scale, late entrants may find that investors want proof of revenue, contracts and margins much sooner than they did during the first AI software rush. For investors, the better question is not whether AI is real. The market has already answered that. The better question is which part of the chain still has earnings power that has not been fully priced. Chips have the clearest demand, but much of that optimism is visible. Power, cooling, networking and data center services may still offer room, but they come with slower execution and less glamorous risks. This is what a structural trade looks like when it grows up. It stops being just a bet on the company with the best chip and becomes a test of the entire economy's ability to fund, permit and operate a new layer of infrastructure. The next phase will be less forgiving. Momentum can keep running, but from here, the market will ask every AI-linked company a harder question: where is the cash flow, and how much of it is already in the price?