UBS Top AI Stock Picks: Amazon, Palantir & More

Global data-center investment is accelerating at a pace not seen since the early cloud era. UBS analysts, publishing a sector note in September 2026, identified 14 stocks across the technology, media, and telecommunications landscape positioned to capture the next phase of that buildout. The UBS AI stock picks include marquee names like Amazon and Palantir alongside a dozen less-discussed plays — a signal that the bank sees AI infrastructure as a broad sectoral opportunity, not merely a semiconductor story.

Why UBS Is Doubling Down on AI Infrastructure

The core argument behind the UBS AI stock picks is timing. UBS analysts characterize the current moment as a "second act" — suggesting the first wave of capital deployment has crested, but the underlying infrastructure build is far from finished. That framing matters. First-act trades in AI were dominated by GPU suppliers and hyperscalers racing to provision raw compute. The second act is about the wider ecosystem: power infrastructure, networking, software platforms, and the connectivity pipes that carry AI-generated workloads.

Research firms including IDC have projected global data-center capital expenditure to run well above $400 billion annually through the late 2020s, driven substantially by AI workload demands. That spending does not flow to a single sector. It touches real-estate developers building facilities, utilities supplying power, networking hardware vendors, and cloud platforms selling capacity downstream. UBS's list of 14 names appears to reflect that breadth.

The "still early" characterization from UBS carries analytical weight. Enterprise AI adoption curves — measured by the share of large companies running AI-native applications at scale — remain in early innings by most independent estimates. The implication: the buildout is a sustained capex cycle, not a one-time event, and investors may have a longer runway than headline market moves suggest.

UBS's 14 Top AI Stock Picks Explained

Amazon and Palantir anchor the published list, but the UBS AI stock picks extend to 12 additional names spread across the TMT spectrum. Amazon's inclusion is straightforward: Amazon Web Services remains the largest cloud infrastructure provider globally, and AWS growth is increasingly tied to AI-related workloads — large-language-model training, inference services, and enterprise AI toolkits. Amazon management has publicly guided toward sustained hyperscale capital expenditure, a signal UBS analysts likely treated as confirmation the company intends to hold its infrastructure lead.

Palantir's presence reflects a different thesis. Where Amazon is an infrastructure provider, Palantir is an AI software and analytics platform with both government and commercial contract exposure. The company has expanded its Artificial Intelligence Platform into commercial enterprise accounts while maintaining established defense and intelligence agency relationships. That dual-track growth — federal contracts providing recurring-revenue stability, commercial AIP deals offering a higher-growth vector — may explain why UBS flagged it alongside hyperscaler plays.

The remaining 12 picks are not individually itemized in UBS's published summary, but the bank's stated focus on TMT broadly suggests the list includes names across cloud networking, data infrastructure, media distribution, and telecommunications. Each presumably connects to either the supply side of the data-center buildout or the demand side — companies whose products become more valuable as AI capacity scales.

Technology, Media, and Telecommunications: The Three Pillars of the AI Trade

UBS's decision to frame these picks across three sectors is worth unpacking. Technology is the obvious leg: chip designers, cloud platforms, and enterprise software vendors all have direct lines to AI revenue. Media and telecommunications are less intuitive additions. Their inclusion suggests UBS analysts see AI as a demand driver for content infrastructure and connectivity, not just compute.

Media companies with large proprietary data assets — content libraries, user-behavior data, licensing catalogs — sit on raw material that AI model training and synthetic media generation increasingly require. That data advantage could translate to licensing revenue or strategic value as training data becomes scarcer and more regulated.

Telecommunications carriers face a different AI catalyst. Network densification required for low-latency AI inferencing at the edge — running models on devices and in local data centers rather than distant cloud regions — demands ongoing investment in fiber and 5G. Telcos that monetize that infrastructure upgrade, whether through wholesale capacity sales or AI-native services, fit naturally into a second-act thesis.

Technology remains the heaviest exposure in any AI infrastructure trade. But the cross-sector construction of the UBS AI stock picks reflects a view that AI's infrastructure demands are pervasive, not siloed — and that the next returns may come from less-crowded parts of the stack.

How to Position Your Portfolio Around the AI Buildout

UBS's publication of a 14-name list is a research call, not a portfolio prescription. Retail investors should approach it as a map of institutional conviction, not a ready-made allocation.

Several structural considerations apply. Concentration risk is real: the sector has attracted significant capital over the past two years, and valuation premiums on perceived AI beneficiaries are elevated relative to historical norms. Diversifying across UBS's three identified pillars — technology, media, telecommunications — offers natural dispersion, since those sectors do not move in lockstep.

Time horizon matters. The "second act" framing implies a multi-year cycle. Shorter-horizon investors face more volatility from quarterly earnings misses and capex commentary shifts. Amazon guides capital expenditure in ranges that move market expectations significantly when revised. Palantir's commercial contract growth, while expanding, can produce lumpy quarterly results depending on deal timing.

Investors tracking the UBS AI stock picks should also monitor regulatory developments. Data-center energy consumption has attracted policy attention in the European Union and, increasingly, in U.S. state legislatures. Changes to permitting, power pricing, or AI-specific regulation could alter buildout timelines and the relative attractiveness of individual names on UBS's list.

What Long-Term Investors Should Take Away From UBS's Call

The September 2026 UBS AI stock picks represent a coherent institutional view: AI infrastructure is a durable capex cycle, it spans more sectors than the consensus trade acknowledges, and the buildout's second phase will reward investors who look beyond obvious first-act names.

Amazon and Palantir serve as anchors — the hyperscaler and the analytics platform — but the list's breadth signals that UBS sees value distributed across the ecosystem. Investors who entered AI through a narrow semiconductor lens in 2023 and 2024 captured meaningful returns. The bank's current note suggests the next tranche may emerge from less-crowded parts of the stack.

Long-term investors should treat the note as a starting point for independent research, not a substitute for it. Verifying UBS's premises — examining Amazon's most recent capex guidance, reviewing Palantir's commercial contract additions, and cross-referencing infrastructure spending forecasts from IDC or Gartner — turns a research note into an informed position.

The AI buildout is neither finished nor guaranteed. But UBS analysts, with cross-sector TMT coverage, are making a concrete case that the second act has begun.

Source: MarketWatch.com - Top Stories