![None](/static//images/frames/Tech_Software_RGB_72dpi.crop.ef301187-8db9-4866-990e-e2b41cc3281e.jpg)
![None](/static//images/frames/Ben_Rogoff_3_transparent.crop.2f7f021e-131d-4732-9074-ff2bb4afdb6e.png)
- Positive outlook with equity market trending upwards and economic data pointing to a ‘soft landing’
- Equity market gains once again led by a narrow range of US large-cap technology companies
- While short-term volatility could reappear, we believe Artificial Intelligence (AI) strength will continue to drive technology sector strength over the medium to long term
Market review
Global equity markets continued to trend higher in June, the MSCI All Country World Net Total Return Index gaining +3.0%. The S&P 500 Index returned +4.3%, although the DJ Euro Stoxx 600 declined -1.7% (all returns in sterling terms). Economic data remained supportive of a US ‘soft landing’ (where inflation moderates without a severe increase in unemployment) although the Federal Reserve (Fed) signalled it was in no hurry to cut interest rates.
The US labour market remains robust, with 272,000 jobs added in May (above forecasts of 185,000), the most in five months, with the largest employment gains occurring in the healthcare and government sectors. Despite the strong labour market, the US Consumer Price Index (CPI) annual inflation rate moderated to +3.3% year-on-year (y/y) in May, from +3.4% y/y in April and below forecasts of +3.4% y/y. The core Personal Consumption Expenditure (PCE) deflator (the Fed’s preferred inflation gauge) also fell to +2.6% y/y in May, the lowest since March 2021.
While noting that there has been “modest further progress” in recent months, the Fed left its funds target range1 steady at 5.25-5.50% for a seventh consecutive meeting, in line with expectations. Policymakers have said since the start of the year that interest rate cuts depend on gaining greater confidence that inflation is moving sustainably towards the central bank’s 2% goal. Despite progress in the month, the Fed’s core PCE inflation forecast for 2024 was raised to 2.8% (from 2.6% in March), while the Federal Open Market Committee (FOMC) – who decide US monetary policy – is now expecting a single rate cut this year (from three expected in March, and 6-7 earlier in the year), followed by four cuts in 2025.
US bond yields helpfully dropped 10 basis points (bps2) to 4.4% during June. However, as we stated last month, while higher for longer interest rates may not be welcomed by long-duration investors3, a robust economy with slowly moderating inflation may prove an optimal scenario for economic growth and technology investment over the medium term.
Technology review
The technology sector strongly outperformed the broader market in June, the Dow Jones Global Technology Net Total Return Index (W1TECN) returning +9.3%, while the Trust returned +9.5% (all figures in sterling terms).
As is often the case during the early stages of a new technology cycle, market breadth remains exceptionally narrow, with returns dominated by companies considered synonymous with the new regime. During the month, the so-called Magnificent Seven4 gained +9.2% with all seven constituents advancing by 5% or more, led by NVIDIA and a rebound in Tesla. As such, large-cap technology stocks continued to significantly outperform their small and mid-cap peers; the Russell 1000 Technology Index (large cap) and Russell 2000 Technology Index (small cap) returning +9.5% and +2.3% respectively. Longer-duration stocks fared worse still, with the Goldman Sachs Unprofitable Tech Index flat over the month.
Software stocks rebounded, the Bloomberg Americas Software Index returning +9.8% in June, lifting first-half returns into positive territory. After the group’s significant underperformance during 2024, some mean reversion was likely with better-than-expected results from Adobe Systems (Adobe) (see below) a rallying cry. Despite less demanding valuations, we remain significantly underweight (‘digital transformation’ is our smallest theme today), with growth slowing and AI-related uncertainty unlikely to dissipate.
In contrast, AI infrastructure spending strength continues unabated and is expected to remain so into mid-2025 as new products ramp and capacity constraints ease.
In contrast, AI infrastructure spending strength continues unabated and is expected to remain so into mid-2025 as new products ramp and capacity constraints ease. AI-related semiconductor stocks outperformed: the Philadelphia Semiconductor Index (SOX)5 added +7.7% in June and +34.7% during 1H24. We were encouraged to see some broadening here, with Broadcom +22.2% and TSMC +16.2% during the month, and both stocks adding to Trust performance partially offset by disappointing performance at Advanced Micro Devices (AMD) where we have continued to reduce exposure. Elsewhere, the NASDAQ Internet Index quietly gained +6% in June (and +15.2% for 1H24), supported by strong returns from Meta Platforms (Facebook), Amazon, Alphabet and Netflix amongst others.
Strong performance was supported by robust off-quarter earnings reports. Within the semiconductor sector, Broadcom delivered results ahead of expectations with revenue +43% y/y as demand for software, networking chips and AI chip design services more than offset cyclical weakness in non-AI semiconductors.
Memory maker Micron Technology also reported earnings during the period for their fiscal third quarter. Revenue grew +82% y/y driven by commodity memory pricing up +20% quarter-on-quarter (q/q) as demand for AI-related high bandwidth memory (HBM) continued to take up capacity and squeeze standard memory supply, driving up prices. Management expects “multiple billions” of HBM revenue next year, with their product offering 30% lower power consumption versus peers. However, margins were pressured by this ramp; we expect HBM to be accretive to corporate margins in the longer term. Management also spoke to a significant step-up in capital expenditure (capex) in 2025, which should benefit some of our semiconductor production equipment (SPE) holdings.
In the software sector, Adobe* reported better-than-expected Q2 results and nudged up their full-year 2024 net new Annual Recurring Revenue (ARR) target by $50m to $1.95bn. Despite ongoing Document Cloud momentum and some early success in their Acrobat AI Assistant product, we remain cautious on the potential risk posed by generative AI to Adobe’s business model. Elsewhere, cybersecurity vendor CrowdStrike Holdings reported strong earnings with Net New Annual Recurring Revenue (NNARR) of $212m above market expectations of $197m, but marginally below some investor expectations. Profit and cashflow metrics were also positive and the company guided NNARR to exceed analyst expectations.
Apple (underweight (u/w)), held its long-awaited Worldwide Developer Conference (WWDC) during the month. We were glad to see Apple finally embracing AI and were impressed by its innovation around small language models (SLMs) that can operate locally on the iPhone, as well as its Private Cloud Compute (PCC) that promises consumers a high level of privacy protection. However, it will be some time before developers can fully utilise these capabilities while Apple Intelligence will not be available in the EU this year due to regulatory concerns. While an AI-driven upgrade cycle appears likely (and nearer term, supply-chain checks suggest improved smartphone demand) we still see this as more a late 2025/early 2026 event. We remain underweight Apple, supplementing our cash equity position with call options6 designed to reduce upside risk should near-term demand prove stronger than forecast.
Outlook
Technology returns have been driven by investor enthusiasm for AI stocks, underpinned by strong fundamental data (including robust revenue and earnings growth) and growing capital spending plans by the hyperscalers (the largest cloud service providers). Mega-cap outperformance primarily reflects superior growth: posterchild NVIDIA is a perfect example where remarkable year-to-date returns have been matched by exceptional revenue/earnings growth, leaving the stock’s price-to-earnings (P/E7) multiple relatively unchanged at c35x 2025. The largest six technology companies are expected to deliver robust second quarter earnings in the coming weeks, with 30% y/y earnings growth, versus 5% expected for the rest of the market.
The Trust has benefitted from our experience and accumulated knowledge from running a dedicated Artificial Intelligence (AI) Fund for the past seven years. In part, this resulted in the application of an ‘AI lens’ to our investment process to help assess every portfolio holding and potential holding in terms of its positioning in an AI-first world. Of the portfolio, 65% is explained by companies we believe are AI enablers, mainly in the semiconductor and cloud computing sectors, as AI infrastructure companies are leveraged to the early buildout phase of a new parallel computing architecture.
There has been a significant increase in aggregate hyperscale AI capex with expectations entering the year for +18% growth, revised up to +26% growth after Q4 earnings, and +44% growth after Q1 earnings (to reach $170bn total spend). Goldman Sachs estimate that there has been $60bn-$80bn of incremental hyperscaler AI capital spending above ‘regular’ cloud capex. While some worry this spend may not deliver an attractive return on investment (RoI), we see these investments reflecting the enormous opportunity the hyperscalers see ahead as every industry – perhaps every activity – has the potential to be reimagined in an AI-first world.
A Bernstein survey found 67% of CIOs believe AI “will be very productivity enhancing” and 79% are exploring applications of AI/LLMs (large language models).
In Microsoft’s case, Azure AI revenue (c$3.4bn run rate) is already higher after just five quarters than Azure’s cloud revenue was six years after launch (and Azure AI is still capacity constrained). A Bernstein survey found 67% of CIOs believe AI “will be very productivity enhancing” and 79% are exploring applications of AI/LLMs (large language models). The proliferation of AI is so early that only 7% have AI as a meaningful part of budgets today, but 66% already agree that AI/LLMs could drive meaningful spending on compute for their organisation in the future.
While we appreciate fears over the sustainability of current AI spend, we believe strong growth will be sustained through the second half of 2024 and into 2025. It is important to note the industry has thus far been supply constrained and we expect significant capacity additions in HBM from major memory makers and advanced packaging technologies (including CoWoS8) at TSMC in the second half of the year. In addition, significant new product launches such as NVIDIA’s Blackwell chip will only begin to ramp in coming quarters, while we expect material improvements in AI model performance with the launch of the next generation of leading-edge models expected before year end. GPT-5.0 is expected to represent a “significant leap forward” both in terms of performance and “human-like” capabilities, according to OpenAI CEO Sam Altman. The AI applications we are all experiencing today (the earliest applications running on current hardware and models) are subject to dramatic improvement. As 2025 progresses, we expect the full potential of AI to become more obvious to all, which will help support continued investment.
It is also important to step back and consider the bigger picture, the race to Artificial General Intelligence (AGI) is now clearly underway. The performance of generative AI models improves in relation to increases in model size, computational resources, and data volume according to observed ‘scaling laws’. There is a strong imperative to increase spending as long as scaling laws hold, chips and models keep improving, and hyperscalers (and potentially other actors like nation states) compete to reach AGI as fast as possible -– expectations now appear anchored around a 2030 rather than 2050 timeline.
In our 25+ years as technology specialists, we have certainly not experienced a technology shift as exciting as we see unfolding today. If history is a guide, almost all industries will be reshaped by AI (and existing profit pools redistributed) while large new markets/opportunities will emerge (impossible to predict let alone size at this early stage). Understanding the impact of AI may be the most important factor investors have to consider this decade and we are hopeful that lessons from previous technology cycles could provide a blueprint for capturing much of the AI growth opportunity and navigating the disruption ahead.
Turning to the near-term outlook, second quarter earnings season will start in the coming weeks and should provide a plethora of supportive AI data points. Our expectation is that results will be solid and guidance upbeat (but, as usual, also conservative). Recent signs of low-end consumer spending weakness will likely continue but are unlikely to detract notably from AI investment strength. After a strong start to the year (driven by a narrow group of stocks), we caution that volatility could resurface whether driven by macro (rates/inflation), geopolitics (elections/policy) and/or sensitivity to perceived negative AI datapoints. However, notwithstanding the possibility of near-term turbulence – especially ahead of earnings season – we anticipate continued AI strength will support the sector through the second half and as such remain constructive on the outlook.
* not held
1. The overnight lending rate among US banks
2. A basis point is one hundredth of one percentage point; it is a common unit of measure for interest rates and other percentages in finance
3. The longer the duration, the more sensitive a stock is to interest rates
4. Microsoft, Amazon, Alphabet, Tesla, Meta Platforms (Facebook), NVIDIA and Apple
5. An index of the 30 largest US semiconductor companies
6. An option allowing the owner the right, but not the obligation, to buy an underlying security at a specific price within a specified time
7. P/E stands for price-to-earnings ratio, which relates a company's share price to its earnings per share
8. Chip on Wafer on Substrate; allows semiconductors to be packaged together more efficiently allowing faster processing, more complex computations and better performance (e.g. in gaming)