Key points

• The Trust’s strong returns in May were driven by a de-escalation in trade tensions and a recovery in AI-exposed stocks
• Governments around the world are treating AI as essential infrastructure, committing billions to AI development
• Elevated volatility associated with an uncertain macro backdrop is likely to persist, but AI progress and adoption continues unimpeded
  

Market review

Equity markets rallied in May, with the MSCI All Country World Net Total Return Index gaining +5.9%, compared to +5.4% for the S&P 500 and +4.1% for the DJ Euro Stoxx 600 indices (all returns in sterling terms).

Sentiment continued to improve after the Trump administration followed its 90-day pause on reciprocal tariffs with several trade deals and de-escalations. The UK secured the first draft trade deal, which laid the groundwork for future discussions on a range of issues but maintained the baseline 10% tariff (suggesting this will be the case for other nations too). The UK received a reprieve on tariffs levied on vehicle and automotive parts, which will be capped at 10% rather than 25% up to a quota of 100,000 vehicles – approximately what the UK exports to the US – as well as a temporary exemption from additional steel tariffs that were announced and came into force in early June.

The US and China also agreed to lower reciprocal tariffs for a 90-day period to allow time for negotiation, a significant de-escalation. The US lowered tariffs on Chinese goods from 145% to 30%, while China lowered retaliatory tariffs from 125% to 10% and temporarily relaxed export restrictions on rare earth elements which are critical to various high-tech industries. Trade negotiations (or lack of) with the European Union were frustrating Trump, who threatened to increase the tariff on the trade bloc to 50% from 1 June but delayed the implementation until 9 July after a positive call with European Commission President Ursula von der Leyen.

At the end of the month, in the first meaningful judicial ruling on Trump's tariffs, the Court of International Trade (CIT) struck down all tariffs issued under the International Emergency Economic Powers Act (IEEPA), which includes all the Liberation Day (2 April) tariffs and the fentanyl/migration tariffs on Canada, Mexico and China (although not the sectoral tariffs on steel, aluminium, autos and auto parts under Section 232, nor further planned sectoral tariffs). The court ordered the administration to cease implementation of the challenged tariffs within 10 days, although the administration was quickly granted a stay of execution on appeal. The issue will likely end up in the Supreme Court and the Trump administration has other options for imposing similar tariffs, although some have never been used and come with further legal questions.

Another concern for equity markets has been rising public debt, reflected in rising long-term government bond yields. Japanese government bond yields spiked during the month, raising fears of a sovereign debt crisis. Japan's Ministry of Finance signalled a potential reduction in the issuance of super-long-term bonds, alleviating concerns about oversupply for now.

The 30-year US Treasury bond yield also temporarily closed above 5% for the first time since late October 2023. This followed Moody’s downgrade of the US sovereign credit rating and the House of Representatives passing Trump’s “Big, Beautiful Bill”, which raised concerns over the Federal deficit and the trajectory of public debt. Section 899 of the Bill allows the US to impose retaliatory taxes (increasing by 5% each year up to 20%) on foreign entities from countries determined by the Treasury to be unfairly taxing US companies or citizens (e.g. Digital Services Taxes in Europe). This would – if passed – represent another adverse development for non-US companies with high US exposure.

The Federal Reserve held its benchmark interest rate at 4.25-4.50% in May, in line with expectations. This marked the third consecutive meeting without a change. The Federal Open Market Committee (FOMC) reiterated its data-dependent, ‘wait and see’ approach amid evolving economic conditions. It is too early to see the full impact of the administration’s trade and fiscal policy changes, but the data released in May suggested the US economy remains resilient: private payrolls increased by 177,000, significantly surpassing market expectations of 130,000, and inflation remains relatively benign. The Consumer Price Index increased +0.2% month-on-month (m/m) in April, rebounding from the -0.1% decline in March, but below forecasts of +0.3% m/m.

Technology review

The technology sector meaningfully outperformed in May.

Large-cap technology stocks marginally lagged their small and mid-cap peers, the Russell 1000 Technology Index (large cap) and Russell 2000 Technology Index (small cap) returning 10.4% and 11.9% respectively. The Bloomberg Magnificent 7 Price Index, however, performed well (+12.4%) following solid results led by NVIDIA (+23%), Tesla (+22%), Meta Platforms (Meta; +17%) and Microsoft (+15%). Meanwhile, Amazon (+10%), Alphabet (+7%) and Apple (-6%) all trailed due to AI, trade war or regulatory concerns.

The Philadelphia Semiconductor Index (SOX) returned 11.6% during the month, while the NASDAQ Internet (QNET) and iShares Software (IGV) indices returned 7.6% and 7.0% respectively.

Saudi Arabia’s HUMAIN has signed deals worth $23bn with leading technology companies so far, with further investment – accounting for 7% of global AI training and inferencing by 2030.

Sovereign AI initiatives, particularly in the Middle East, have also picked up notably in recent weeks, which coincides with Trump cancelling AI diffusion export controls. Saudi Arabia’s HUMAIN has signed deals worth $23bn with leading technology companies so far, with further investment – accounting for 7% of global AI training and inferencing by 2030. The UAE announced a $15bn Stargate UAE project with OpenAI and there have been further sovereign AI announcements from Indonesia, India and Qatar. As NVIDIA’s CEO Jensen Huang put it: “Countries around the world are recognising AI as essential infrastructure – just like electricity and the internet”.

NVIDIA itself delivered another remarkable quarter during the month, with data centre revenue, including semiconductors (chip) and other components used in AI computing systems, surging +73% y/y, driven by strong demand for the company’s Blackwell chips, which are now in full production. NVIDIA estimates the Chinese AI chip market is a $50bn “missed opportunity”, so US/China negotiations could be a significant source of potential upside.

Advanced Micro Devices (AMD) reported solid March quarter results with revenue up +36% y/y, driven by strength across various segments. Management continues to forecast strong double-digit revenue and earnings per share (EPS¹) growth this year, despite a $1.5bn hit to graphics processing unit (GPU²) sales due to China export controls.

Astera Labs, a provider of AI infrastructure connectivity solutions, delivered another strong quarter with Q1 revenue up +144% y/y, driven by sales of retimers and active electrical cables (AECs) into AI servers. Celestica, a leading electronics manufacturing services firm, continued to rally in May, buoyed by AI-related strength, as well as positive results at the end of April, when the company confirmed no change in hyperscaler demand.

Unfortunately, not all semiconductor firms performed strongly. eMemory Technology, a leading provider of silicon IP (intellectual property), reported weaker-than-expected results due to prolonged decision-making cycles for advanced security IP.

Apple underperformed the technology sector in May. The company delivered slightly better-than-expected top and bottom-line results, driven by a 2% y/y increase in iPhone sales – although this may have been boosted by a tariff-related demand pull-in. June quarter guidance was marginally below analyst expectations, baking in a $900m cost headwind from tariffs. The stock continues to be a lightning rod for the trade war given the company’s high dependence on manufacturing in China. Trump threatened to impose a 25% tariff on Apple unless the company begins manufacturing iPhones in the US, a scenario which would significantly increase costs.

In software, Microsoft rallied following better-than-expected results announced at the end of April. Its Azure business grew above expected levels, with AI’s contribution to growth jumping as use of OpenAI soared. AI usage ex-OpenAI is also accelerating, with management highlighting that Microsoft processed 100 trillion tokens in the quarter, up 5x y/y, including a record 50 trillion in the final month alone, indicative of the extraordinary demand for AI.

Cybersecurity remains a core theme in the Trust due to growing AI threats. CyberArk Software posted solid results and strong operating margins. Management sees no softness in the pipeline but guided conservatively to reflect macroeconomic uncertainty.

Robinhood Markets reported strong results, with revenue up +50% y/y, driven by a +77% y/y rise in transaction revenues. Its Gold subscription product showed impressive growth and many of the company’s new product initiatives are already generating revenue.

Cloudflare, which provides security, performance and reliability services for websites and online applications, reported robust results including a $130m ‘Pool of Funds’ deal, where the company provides a comprehensive, developer-friendly cloud alternative to hyperscalers. Q2 guidance was in line, while full-year guidance was maintained (not increased) with management prudently factoring in greater macroeconomic uncertainty despite no evidence of a spending slowdown.

Outlook

Markets staged a strong recovery in May, the S&P 500 gaining +5.4% and the NASDAQ up +9.1%, delivering their best monthly performances in the past 18 months as macro fears subsided and credit spreads³ tightened significantly. The market rebound reflects a partial de-escalation of trade tensions and hopes that manageable – if highly imperfect – solutions to trade issues can be found before uncertainty impacts ‘hard’ macroeconomic data. Consumers continue to benefit from tight labour markets while strong household balance sheets and sentiment have improved from low levels.

We remain of the view that it is not in policymakers’ interests to provoke a deep global recession and it is (for now) within their capacity to prevent it – a period of ‘recalibration’ rather than a full ‘reset’. Uncertainty will, however, remain elevated and markets volatile given the Trump administration’s ‘escalate to de-escalate’ modus operandi and the fundamental challenges of resolving many of the competing interests of different parties.

Risks from growing deficits and enormous public debt burdens have also come into sharper focus, reflected in the rebuilding of term premiums (those on US 10-year Treasuries are back to 2014 highs, per Goldman Sachs). Equity markets are taking their cue from bond yields and the US dollar as signals of potential distress that could spill over, but for now the speed and scale of moves appear contained. The pace of change perhaps matters as much as the level, with equities historically struggling when bond yields have risen by more than two standard deviations in a month (around 60 basis points (bps⁴) today).

AI investment remains robust while hyperscalers are still capacity-constrained against a backdrop of incredibly strong (and accelerating) AI usage.

As discussed, Q1 earnings were generally solid with positive earnings revisions at the technology sector level. Most importantly, AI investment remains robust while hyperscalers are still capacity-constrained against a backdrop of incredibly strong (and accelerating) AI usage. Alphabet revealed its monthly tokens processed increased c50x in 12 months, primarily because of the huge increase in usage associated with reasoning models as they run ‘chain of thought’ responses to prompts. NVIDIA actually suggested at a recent conference that Deepseek uses less compute to train and run but with a significant increase in token use for inference⁵/queries (increasing the number of tokens per query by 13-20x). Microsoft processed 100 trillion tokens in Q1, of which half were processed in March alone.

Our experience of multiple previous cycles suggests that many of today’s incumbent technology companies will face the ‘innovator’s dilemma’. This may explain Alphabet’s recent move to more widely replace traditional search results with AI overviews; in time, this may prove the correct decision but to us it appears highly defensive. Gartner predicts traditional search volume will drop 25% by 2026, with organic traffic potentially decreasing by more than 50% as consumers embrace AI-powered search. Wells Fargo also estimates that ChatGPT (8% of search use in April) could achieve 30% share, or $100bn, by 2030, with Google Share declining from 90% to 60%.

AI adoption continues to increase. Meta announced one billion monthly active users of Meta AI, up from 500 million in September 2024. A recent survey of Americans discovered that workers found AI tripled their productivity (90-minute tasks become 30-minute tasks) and studies across fields as varied as product development, sales, consulting, coding, legal work and call centres suggest widespread productivity gains are possible, even at this early phase of the technology’s evolution. Translating these gains into firm-wide (and eventually economy-wide) impacts is challenging, but it is encouraging to see internal memos from the management of Shopify, Duolingo* and the Norwegian sovereign wealth fund mandating AI usage as a ‘baseline expectation’ for staff. We are impressed by the increasing maturity and utility of AI tools used within our own team for stock-screening and fundamental research.

It is also important to appreciate the frenetic pace of innovation as the building blocks for agentic AI⁶ are put into place. This month, Microsoft launched an initiative to help build the ‘agentic web’, throwing support behind open protocols such as Anthropic’s Model Context Protocol (MCP), which is intended to provide a common language for AI agents to communicate, akin to how HTTP facilitated the early internet. Google DeepMind announced AlphaEvolve, an AI agent that claims to be able to invent brand new computer algorithms, test them and put them directly to work within Google servers. AlphaEvolve has been running within Google for over a year and has helped it recover an average of 0.7% of Google's worldwide computing resource continuously by identify "stranded resources".

Elsewhere, OpenAI announced the $6.5bn acquisition of Jony Ive’s hardware startup io with plans to ship 100 million ‘AI companions’ following an anticipated launch in late 2026. Alphabet unveiled Veo 3, its latest AI model capable of generating hyper-realistic talking video content. On the company’s recent earnings call, Huang remarked that “AI inference token generation has surged tenfold in just one year, and as AI agents become mainstream, the demand for AI computing will accelerate.” The Trust remains widely exposed to AI enablers to take advantage of this trend.

* not held

1. Measures a company’s value by assessing how much money a company makes for each of its shares; earnings growth is not a measure of future performance

2. An electronic circuit designed to process images and graphics on a computer

3. The difference in yield between two bonds with similar maturities but different credit ratings or risk levels. It measures the additional risk premium an investor receives for investing in a lower-rated bond

4. A basis point is a common unit of measure for interest rates and other percentages in finance. One basis point equals 0.01%

5. The process of drawing an idea or conclusion from evidence and reasoning

6. Agentic AI refers to an advanced AI system that autonomously takes actions, adapts in real-time and solves multi-step problems based on context and objectives