Why AI Infrastructure Stocks Keep Climbing

The spending number driving the rally

The clearest signal behind the climb in AI infrastructure stocks is not a product launch or a benchmark result. It is a capital expenditure number, and it is enormous. The five largest US cloud providers, Amazon, Microsoft, Alphabet, Meta and Oracle, are on track to spend between $650bn and $690bn on capital expenditure in 2026, close to double what they spent the year before. Almost all of it is pointed at one thing: the compute, data centres, networking and power needed to train and run artificial intelligence. When a small group of the most profitable companies on earth commit that much cash in a single year, the firms that supply the underlying hardware feel it directly in their order books. That flow of money is the engine under the whole theme.

Why it matters right now

This is not speculative capital chasing a demonstration. It is contracted, budgeted spend from companies with the balance sheets to fund it out of cash flow. Widen the lens and the numbers grow further: industry trackers put total 2026 capital expenditure across the fourteen largest data centre operators close to $750bn, and JPMorgan expects the cycle to push beyond $1.1trn in 2027. AI infrastructure has again been named among the standout thematic categories in the 2026 outlooks from the major banks and asset managers. What makes this cycle different from earlier technology booms is that the earnings are turning up to support the story rather than lagging years behind it.

The second reason it matters is breadth. For a UK investor, the temptation is to reduce the theme to a single famous American chip designer. The reality is a long supply chain. The money reaches graphics processors and custom silicon, but also high-bandwidth memory, networking switches and optics, cooling systems, electrical equipment, and the power generation feeding the data centres themselves. Data centre electricity demand is now large enough to strain grids, and forecasters expect it to keep rising steeply through the decade. Several of the businesses in these links are companies most private investors have never examined closely, and they are where a good number of AI infrastructure stocks now sit. That is precisely the gap a consistent, scored screen is built to close.

What puts a stock in the AI infrastructure basket

Openbook does not place a company in the AI infrastructure basket because a news headline connects it to the theme. Every stock is measured against the same factor model, and the theme is defined by the pattern that model produces. On the Reward side, Growth carries the heaviest weight, then Momentum, then Profitability, then Valuation. Companies that qualify here typically score strongly on Growth, because rising capex is feeding genuine revenue expansion, and on Momentum, because their shares have been trending well ahead of the wider market for some time.

Profitability is where the model does its most useful sorting. It separates the established suppliers earning hard margins on real products from the earlier-stage names still consuming cash in the hope of future scale. A company growing fast with strong profitability may carry a tag like Growth Machine, while one growing fast without profits can surface as Profitless Growth, and the two deserve very different treatment even inside the same theme. The factor most likely to flash a warning is Valuation, which carries the smallest Reward weight at fifteen percent. Many of these shares trade at rich multiples of forward earnings, so a strong overall Reward Score can sit right next to a weak Valuation sub-score. That is not a contradiction to ignore. It is the model telling you the market has already priced in a lot of the growth. The theme runs across the whole hardware stack, with compute names like Nvidia and AMD, custom silicon and networking from Broadcom, foundry capacity from TSMC and high-bandwidth memory from Micron — covered in depth in our Micron stock analysis — all scored on the same factors rather than grouped together by narrative. You can filter the entire market on those measures using the Openbook stock screener, or browse the wider semiconductor collection for the chipmakers beyond the AI build-out.

How to weigh the risks

The risk here is best described as moderate, and the word is doing real work. There are genuine earnings and genuine demand underneath the theme, which sets it apart from a pure sentiment trade. But the valuations are stretched, and that cuts both ways. The most discussed danger is that the capex boom outruns the revenue AI actually produces, leaving expensive, half-used infrastructure and disappointed shareholders. The comparison analysts reach for most often is the fibre-optic overbuild of the late 1990s, when demand was real but arrived years later than the spending assumed. There is also concentration risk. A large share of the spending comes from a handful of buyers, so any single hyperscaler trimming its budget can ripple through the entire supply chain at once. Physical limits matter too. Power availability and grid connection are now a real bottleneck on how quickly data centres can be built, which can cap the growth the valuations depend on.

On the Openbook Risk Score, this profile tends to show up as elevated Volatility, the factor with the largest risk weight, because momentum-led shares move sharply in both directions. None of this means the theme is broken. It means these shares are priced for a great deal to go right, and the Risk scores reflect that honestly. For this kind of theme, position sizing and a clear view of what you are actually paying for matter far more than they do in steadier corners of the market.

See the scored list

The businesses behind the AI buildout are easy to name and genuinely hard to compare on a like-for-like basis. That is exactly where a single consistent scoring model earns its place. See the live, scored list of AI infrastructure stocks, each rated on the same Reward and Risk factors, on the Openbook Discover page. It is the quickest way to move from a theme you have read about to the specific companies that currently qualify under Openbook's criteria, with their growth, momentum, profitability and valuation reads set out side by side so you can judge them on the same terms.

Openbook is an educational research platform. We do not give personalised recommendations or investment advice.