WisdomTree

The 4 long-term themes investors agree on

Published 3 November 2026

Mobeen Tahir
Mobeen Tahir

Director, Macroeconomics and Thematic Research

@MobeenTahirWT

Key Takeaways

We surveyed both retail and professional investors this year, and something interesting stood out from the results. In April, Opinium asked 4,000 retail savers and investors across the UK, Italy, France and Germany, all with at least £5,000 in savings, which themes offer the most compelling long-term opportunity. In July, Censuswide put a similar question to 800 professional investors across Europe.

Both groups put the same four themes at the top, in the same order.

Theme

Professionals

Retail

Artificial intelligence software

17.25%

33%

Artificial intelligence infrastructure

15.50%

29%

Cybersecurity

14.88%

24%

European defence

14.63%

23%

Both retail and professional investors have made smart choices in picking these themes. In this blog, we outline WisdomTree’s thinking and why the ranking makes a lot of sense.

Artificial intelligence software

Apple launched the iPhone in 2007. The App Store followed in 2008. Instagram arrived in 2010. In 2007 nobody could name the companies that would make the most of mobile internet, because many of them didn't exist yet.

AI software looks a bit like the early days of the App Store. We are in the buildout phase. Investors are right to wonder whether OpenAI and Anthropic can earn enough to justify all the capex. But this may be the wrong question to ask. The model makers are one slice of the opportunity, and their numbers are already striking. Anthropic told investors its annualised revenue run rate hit about $65 billion in July, up from roughly $9 billion at the end of 20251.

The bigger prize sits with the companies building on top of these models, and many of them may not exist yet. Some early examples are already live. Palantir is putting AI into decisions at large enterprises and governments. Synopsys is using it to speed up chip design. Recursion is applying it to drug discovery. NICE is building it into customer service. These look like early apps on the App Store, and the Instagram of this cycle may still be to come.

Artificial intelligence infrastructure

Every app needs somewhere to run, and that infrastructure is being built right now.

The big US hyperscalers guided to roughly $720 to $745 billion of combined capex for 20262. Amazon lifted its own figure to around $220 billion during the year, partly because memory costs rose. Alphabet expects a significant increase again in 2027.

That money flows down the supply chain. Nvidia, AMD and Broadcom design the chips. TSMC makes them. ASML, Lam Research and Advantest supply the equipment to make and test them. SK Hynix and Samsung provide the memory, which has become a pricing story of its own.

Then comes power. The International Energy Agency expects data centre electricity demand to roughly double to about 945 terawatt hours by 20303. Every one of those sites needs power, cooling and connections, which brings in companies like Amphenol alongside the chipmakers.

Cybersecurity

Software had a rough start to the year. Investors feared AI agents would eat into traditional software, and the whole group sold off, security names included. People called it the SaaSpocalypse. The bounce back has been strong though and cybersecurity stocks have made strong gains in the second and third quarters of the year.

Markets have worked out that more AI means more to defend. Every new model, agent and integration is another way in. CrowdStrike's 2026 Global Threat Report shows how fast attackers are moving. Attacks by AI enabled adversaries rose 89%. The average time for an eCrime attacker to move across a network after first access fell to 29 minutes, and the fastest on record was 27 seconds. 82% of detections were malware free, which means attackers often get in with valid credentials and built in tools.

Even with rates rising, security spend is hard to switch off. A breach costs far more than the software built to stop it. Companies like CrowdStrike, Okta, Zscaler, Fortinet and Rubrik cover everything from identity to the network to data recovery.

European defence

The headlines explain some of the demand. Defence is about preventing the next conflict and building that strength takes years.

The order books show how long. Book to bill ratio for defence companies compares orders won with sales in a period, and anything above one means the backlog is growing. Saab took SEK 65.8 billion of orders in the third quarter against SEK 20.6 billion of sales, a ratio of about 3.2. Its backlog now stands at SEK 364 billion4. Thales booked €19.1 billion of orders in the first nine months against €16.4 billion of sales and targets a ratio above 1.1 for the year5.

Fincantieri shows how long this can run. Its backlog hit a record €74.2 billion in the first quarter, around 8.1 times 2025 revenue, with delivery visibility out to 20396. A ship doesn't stop earning when it leaves the yard. Maintenance, upgrades and support continue over its life in service, so revenue keeps recurring.

Conclusion

Professionals and retail investors picked the same four themes in the same order. AI software is early in an app era. AI infrastructure is getting hundreds of billions of dollars of investment right now. Cybersecurity demand grows with every new AI system. European defence has order books stretching into the 2030s. Long-term themes ask for patience, and these four have plenty of demand behind them.

1 CNBC, August 2026: Anthropic says annualized revenue climbed to $65 billion in July
2 Datacentres.com, October 2026: Hyperscaler CapEx Trends: 2026 Guidance Passes $700 Billion
3 IEA, October 2026: Energy demand from AI - Energy and AI - IEA
4 Airforce Technology, October 2026: Saab Q3 orders jump 216% on Gripen, GlobalEye and German naval contracts
5 Investing.com, October 2026: Thales Q3 2026 slides: order intake jumps 15%, defence surges By Investing.com
6 Container News, May 2026: Fincantieri raises 2026 guidance as backlog hits record €74.2 billion - Container News

Categories

About the contributor

Mobeen Tahir
Mobeen Tahir

Director, Macroeconomics and Thematic Research

@MobeenTahirWT

Mobeen is a member of WisdomTree’s research team where he focuses on a wide range of asset classes to offer strategic and tactical insights to our clients on global markets and investment products. Before joining WisdomTree in December 2018, Mobeen worked at Willis Towers Watson as an investment consultant advising institutional clients as well as their in-house fund business on asset allocation and portfolio construction with his research focus being equity and multi-asset smart beta. Mobeen has a BSc (Hons) in Accounting and Financial Management from Loughborough University and an MSc in Accounting and Finance from the London School of Economics and Political Science. He is also a CFA Charterholder.