• Expectations for interest rate cuts were pushed out as inflation stayed higher than expected
• Q1 technology results were strong with news flow and corporate investment around AI positive indicators for future AI revenue opportunities
• We believe we are at the start of a new AI-driven technology cycle with strong growth potential ahead
  

Market Review

Sentiment was impacted in April by higher-than-expected inflation numbers and commentary from the Federal Reserve (Fed) that interest rates were likely to stay higher for longer. Global equity markets pulled back in April with global equities (measured by the MSCI All Country World Net Total Return Index) declining 2.3% while the US (S&P 500 Index) and Europe (DJ Euro Stoxx 600 Index) returned -3.2% and -1.0% respectively (all returns are in sterling terms).

The US labour market remains robust, with 303,000 jobs added in March (above forecasts of 200,000), the most in 10 months, with the largest employment gains occurring in healthcare and government sectors. However, the Advance Estimate of US real GDP, a measure of economic growth, grew by less than expected, increasing at an annual rate of 1.6% in 1Q24 below forecasts of 2.5%. This was also down from 3.4% in 4Q23 and the lowest since 2Q22, mainly due to a slowdown in goods consumption.

Fed policymakers have said since the start of the year that interest rate cuts depend on gaining greater confidence that inflation is moving towards the central bank's 2% goal. The US Consumer Price Index (CPI) annual inflation rate accelerated 3.5% year-on-year (y/y) in March, from +3.2% y/y in February (above forecasts of +3.4% y/y). This was the third successive higher than expected monthly CPI number while the Fed’s preferred measure of inflation, core Personal Consumption Expenditure (PCE), was stable month-on-month (m/m) at 2.8% y/y in 1Q24, above forecasts of 2.6%.

Fed Chair Jerome Powell confirmed that recent data means “it is appropriate to allow restrictive policy further time to work” but quashed (in our view unwarranted) speculation of a rate hike or fears of “stagflationary winds”, stating “it is unlikely the next move will be a hike”.

That said, it is important to note how significant the shift in interest rate expectations has been, with initial expectations for six to seven rate cuts in 2024 now revised down to between one and two, starting in September at the earliest. This has seen 10-year Treasury yields rise by c80 basis points and created a significant headwind for longer-duration growth stocks. We believe this largely explains technology weakness this month, rather than any fundamental setback.

While we do not expect a sharp reversal of interest rate expectations, there are signs of softening consumer demand apparent in commentary from Starbucks, Disney and Amazon which, combined with less excess consumer savings and a weaker labour market, should create a more favourable rate backdrop from here.

Technology review

The technology sector modestly underperformed the broader market in April. The Dow Jones Global Technology Net Total Return Index (W1TECN) returned -3.5%. After several hotter-than-expected CPI prints in the US and a move higher in bond yields, technology stocks pulled back after a strong start to the year.

Large-cap technology stocks materially outperformed their small and mid-cap peers; the Russell 1000 Technology Index (large cap) and Russell 2000 Technology Index (small cap) returning -3.4% and -8.3% respectively. After a record year in 2023 and continued outperformance in the first quarter of 2024, the Philadelphia Semiconductor Index (SOX) declined -3.8%, while the NASDAQ Internet Index and Bloomberg Americas Software Index returned -3.6% and -6.9% respectively.

AI-related news flow was overwhelmingly positive, with capital expenditure increases at all the hyperscalers, further evidence of confidence in AI model progress and future revenue opportunities

While the market environment proved challenging during the month, first quarter technology results were robust with AI adoption firmly on (a very fast) track. While expectations may have got ahead of themselves, AI-related news flow was overwhelmingly positive, with capital expenditure (capex) increases at all the hyperscalers (Microsoft, Amazon Web Services, Google and Meta Platforms (Facebook)) further evidence of confidence in AI model progress and future revenue opportunities. While some AI-related names experienced meaningful pullbacks from highs, we expect estimates for this group to move higher during the second half of the year and consider recent weakness an opportunity to accumulate favoured names.

In the semiconductor sector, Advanced Micro Devices (AMD) reported an in-line quarter and guidance given strength in their Data Centre and Client businesses, albeit somewhat offset by weakness in their more cyclical Embedded and Gaming segments. Most notably, the company raised their expectation for sales of their AI GPU. In contrast, close peer Intel (*underweight; u/w) missed market expectations and guided down for the 2Q with underwhelming progress on their foundry and AI business aspirations.

Semiconductor capital equipment maker ASML Holding reported an in-line quarter but orders for their flagship extreme ultraviolet lithography (EUV) tools came in at €3.6bn, below consensus expectations of €5bn. That said, management commentary was bullish around the ramp of the next-generation semiconductor chips and orders from TSMC are forecast to come through in the next few months, underpinning ASML’s confidence in achieving the high end of their 2025 guidance.

TSMC’s results and next quarter guidance exceeded consensus forecasts, with the company highlighting AI/High Performance Computing (HPC) demand remains robust and is expected to grow at a 50% compound annual growth rate (CAGR) for the next five years to reach >20% of sales by 2028. Management reduced expectations for overall semiconductor industry growth (ex-memory) to 10% y/y, highlighting the relatively slow recovery for non-AI silicon.

ASM International had a strong month, driven by continued semiconductor chip orders from all three major foundry customers (including NVIDIA), as well as a large increase in high-bandwidth memory (HBM) orders.

In the internet sector, Meta Platforms (Facebook) reported strong results with 27% y/y revenue growth, driven by online commerce, gaming and media, but next quarter guidance was not enough to meet elevated expectations. The company also raised their capex guidance for the year as they invest heavily in generative AI capabilities and develop Llama 3, their industry-leading large language model (LLM).

Alphabet (Google) also reported strong earnings as core search grew 14.4% y/y and YouTube growth accelerated to 21% y/y, both above consensus estimates. Management focused on their generative AI progress and the associated decrease in query costs for their Search Generative Experience offering, now down 80% since introduction. Conversely, Apple underwent a difficult period of performance in the month as industry data on iPhone sales suggested slow demand for the company’s flagship device.

[Netflix] net subscriber adds in the quarter were 9.33 million, ahead of market expectations, as the crackdown on password sharing began to bear fruit

Netflix reported strong results, posting 16% y/y revenue growth, with margins and earnings per share (EPS) ahead of expectations. Critically, net subscriber adds in the quarter were 9.33 million, ahead of market expectations, as the crackdown on password sharing began to bear fruit. However, the company also noted they would stop reporting subscriber data from 2025 and guidance for 2Q24 was modestly below investors' elevated expectations.

Spotify Technology delivered strong numbers, demonstrating its pricing power, with average revenue per user growth accelerating. This helped drive gross and operating margins ahead of expectations, as management steers the company towards sustainable profitable growth.

Amazon posted an impressive quarter, with AWS growing 17% y/y and overall company operating income of $15.3bn, well above expectations of c$11bn. The company plans to meaningfully increase capex to build out AI infrastructure. While its Q2 guidance was a little light of expectations, Amazon continues to deliver strong retail margins while also seeing re-accelerating growth in their leading cloud business.

In the software sector, Microsoft exceeded expectations on both the top and bottom line, with particular strength in its cloud business, Azure, growing 31%, partly driven by AI demand. Management expects to increase capex materially to meet future AI demand and noted that capex will grow further into 2025. Looking ahead, despite increased spending potentially impacting margins temporarily, Microsoft is seeing strong AI tailwinds with c60% of Fortune 500 companies using its Copilot products.

ServiceNow posted strong margins in the quarter and called out process optimisation as the single biggest generative AI use case in the eEnterprise today. Their Pro+ SKU remains the fastest selling offering in the company’s history. Cadence Design Systems posted an in-line quarter, but management gave softer- than- expected next quarter guidance.

As regards electric vehicle (EV), bellwether Tesla posted weak results as expected by the market, impacted by reduced demand and supply issues. Management reassured investors they expect volume in Q2 to be better than Q1 and to grow in 2024, in line with consensus estimates. Tesla is also accelerating the introduction of a new lower cost model to late 2024/early 2025 and management talked about significant step changes to advance full self-driving. The company has a robotaxi event in August where they are expected to discuss their offering, which may compete against ride hailing incumbents Uber Technologies and Lyft*.

Outlook

AI-exposed areas of the market have been subject to elevated volatility in recent weeks, unsurprising given high investor expectations and the normal challenges and setbacks of a transformative technology theme. Volatility will be an ongoing (and at times uncomfortable) aspect of the AI story, but our own enthusiasm for the opportunity remains undimmed and the Trust remains firmly positioned to benefit from it, in our view.

Macroeconomic and geopolitical developments have also contributed to higher volatility. The move higher in bond yields has weighed on investor sentiment at a time when valuation multiples are high versus history; the US 10-year Treasury yield was up 48bps during the month and two-year yields closed >5%. This has been unhelpful to growth equities in particular.

Our base case remains that rate cuts have been pushed further out as Q1 inflation has been a little higher than the Fed, and market participants, might have hoped. This has put upward pressure on real yields (government bond yields adjusted for inflation), which have moved up from 1.7% coming into the year to c2.2% by the end of April. The sharp rise in real yields has put less pressure on equity returns (which remain positive year-to-date) than might have been expected, given the strong economic growth backdrop and investor excitement regarding the long-term potential for AI. Our base case is the Fed will likely follow other central banks in cutting interest rates this year as inflation trends back towards target, which could provide further support to equity markets, small caps especially.

Volatility also brings opportunity, especially when valuations have been somewhat extended in some subsectors. We are encouraged to see the ongoing return of strategic software M&A with IBM’s* $6.4bn bid for cloud infrastructure orchestration software provider Hashicorp* – valuations in some parts of the market are palatable to the right buyer. In total, there have been six large (>$500m) deals in the software sector announced so far this year.

The adoption of AI products for companies such as Microsoft and ServiceNow are faster than any previous product. ServiceNow found their own ‘deflection rates’ (when an employee or customer query does not need to be passed on to a human agent) have doubled after adopting their own AI products and have continued to improve m/m. We remain very selective within software given uncertainty associated with generative AI which has the potential to reshape the software industry.

Axon launched AI-powered report-writing software ‘Draft One’, which can auto-draft reports based on police body-camera footage and sound and is already saving police officers an hour per day on paperwork in the first trials

A particularly interesting early example of how AI might significantly improve worker productivity came from law enforcement software and hardware provider Axon Enterprise. Axon launched AI-powered report-writing software ‘Draft One’, which can auto-draft reports based on police body-camera footage and sound and is already saving police officers an hour per day on paperwork in the first trials (officers spend up to 40% of their time writing reports). A police force in Colorado has already seen an 82% decline in time spent writing reports while “the quality of reports has improved substantially”. This is just an early glimpse into the kind of AI-driven innovation to come.

The most important AI datapoint of all has been the scale of the positive capex revisions at the largest cloud companies – the hyperscalers. Hyperscaler capex this year is now expected to grow 44% y/y to >$170bn in aggregate, up from expectations of 26% (previously 18%), representing the fastest growth since 2018. AI servers will account for nearly 60% of hyperscalers’ total server spending, according to Gartner.

We treat this data point as particularly significant given the hyperscalers’ visibility into AI product roadmaps and the strong vote of confidence it implies in the ongoing improvement and future potential of an already extraordinary technology. These companies have delivered high returns on their investments over the past few years, and we take seriously the signal that a meaningful step up in capex in AI sends about the potential size and scope of the AI market in the future.

Further, Gartner have revised higher 2024 global IT spending growth expectations from 6.8% last quarter to 8%, surpassing $5trn in the process. Data centre spending is expected to see significant acceleration, from 4% in 2023 to 10%, “in large part due to planning for generative AI”. The AI infrastructure buildout continues at a frenzied pace, however there are capacity constraints apparent in the semiconductor and data centre markets, as well as bottlenecks around sufficient power and labour supply. Against this backdrop, first quarter earnings season was solid but, in some cases, not enough to drive immediate upside after a strong start to the year pushed up investor expectations. We anticipate upside to AI-related demand in the second half of the year.

In our view, generative AI is a rare example of Discontinuous Technology Change. We are at a critical turning point where an entirely new compute architecture/stack is required and, as per earlier technology cycles, previous winners are unlikely to be the future leaders. A change of leadership may follow which will require an active investment approach and a deep understanding of technology change. 

The initial investor reaction to Meta Platforms (Facebook)’s significant increase in AI investment and pivot to lead with AI-based products were taken poorly in the near term, but we note CEO Mark Zuckerberg’s comments from a recent interview regarding scaling laws: “I think it is likely enough that we will keep going. I think it is worth investing the $10bn or $100bn+ in building the infrastructure.” The expected launch of GPT-5 over the summer as well as the launch of Meta Platforms (Facebook)’s 425 billion-parameter Llama 3 model and Amazon’s ‘Olympus’ model will serve as important waypoints to assess the fundamental progress in the underlying AI technology.

The Trust is widely exposed to the AI infrastructure build with 65% of the portfolio accounted for by companies deemed ‘AI enablers’, predominantly in the semiconductor, semiconductor capital equipment and cloud computing subsectors. ‘AI beneficiaries’, which are expected to benefit from the ascent of AI and its sustainable impact on their businesses, make up a further 20%. A small number of ‘AI adopters’ – companies aggressively adopting AI but it is not clear whether this will provide sufficient competitive differentiation in the longer term – make up most of the remainder. We have used recent weakness in AI-related stocks to add to our favoured names. We remain highly constructive on demand, with significant hardware upgrades, model improvements and increased capacity coming online over the next 12 months.

Following a number of thematic false starts in recent years, we understand why some investors might default to ‘bubble’ at times like this. However, we believe we are the beginning of an accelerated GPT buildout, akin perhaps to 1995 – early in a new technology cycle with strong growth ahead. 

*not held