- Inflation trending back to target is a boon for equities and the broader global economic outlook
- A mixed first-quarter earnings season was led by AI-exposed companies, with NVIDIA heading the pack
- AI continues to dominate, with semiconductors now the largest technology subsector in the S&P 500 for the first time
Market review
Global equity markets rose in May, the MSCI All Country World Net Total Return Index gaining +2.2% while the S&P 500 Index returned +3.1% and the DJ Euro Stoxx 600 +3.3% (in sterling terms). Major indices broadly recovered April’s losses with the S&P 500 and Stoxx 600 indices making new highs, while Brent Crude prices fell after four consecutive months of gains amid a more muted geopolitical backdrop.
After a strong non-farm payroll number for March, the April labour market report showed the US adding just 175,000 jobs (below forecasts of 240,000), the slowest increase in six months, while the unemployment rate rose to 3.9%. This buoyed equity market sentiment on the expectation of a soft landing (where inflation moderates without a severe increase in unemployment). Unfortunately, a strong jobs print in early May showed how volatile this data can be (although unemployment also ticked up to 4%). In Europe, flash PMI (Purchasing Managers’ Index) data hit a 12-month peak resulting in German 10-year yields moving up +8 basis points (bps) to 2.66%. The expectation of stickier inflation in Europe combined with resilient growth saw the trade-weighted US dollar index fall by 1.5%, the worst-performing currency in the G10.
Elsewhere, there was a notable move in Japanese 10-year government bond yields which rose to 1.06%. This is the first time since 2012 that yields had surpassed 1% in Japan as investors digested hawkish Bank of Japan rhetoric and the expectation of future interest rate hikes. In Taiwan, strength in the labour market and wage growth – no doubt influenced by strength in the domestic technology industry led by growth at bellwether TSMC – led to an increase in full-year GDP growth forecasts to 4% from 3.5%.
Towards the end of the month, Federal Reserve members struck a somewhat more hawkish tone than expected, citing the requirement to still see “many more months” of further positive inflation data before easing monetary policy conditions. Indeed, some members of the FOMC expressed a willingness to tighten policy further should that be required. That said, the global economic outlook remains broadly positive and supportive of equities as inflation continues to trend back to target, economic growth is holding up and labour markets are cooling but not showing signs of distress. However, confidence levels for low-income consumers fell to pandemic levels as the ‘higher for longer’ rate environment continues to disproportionately impact the group.
Technology review
The technology sector outperformed the broader market in May. The Dow Jones Global Technology Net Total Return Index (W1TECN) returned +6%. Once again, there was significant dispersion between technology subsectors as semiconductors dominated returns. The Philadelphia Semiconductor Index (SOX) returned +7.8%, while the NASDAQ Internet Index and Bloomberg Americas Software Index returned +0.5% and +0.9% respectively. Large-cap technology stocks outperformed their small and mid-cap peers; the Russell 1000 Technology Index (large cap) and Russell 2000 Technology Index (small cap) returned +7.1% and +3% respectively.
First-quarter earnings season continued to be mixed, both in terms of results and market reactions, although AI-exposed companies delivered stronger results, led by NVIDIA. Even against elevated expectations, this was another outstanding quarter, with revenue growing +262% year-on-year (y/y) and next-quarter guidance meeting an already high bar. The ramp up of the company’s next-generation AI chips (Blackwell) is happening faster than expected, with management noting "we will see a lot of Blackwell revenue this year", while demand is expected to outstrip supply into next year. Combined with still strong current generation (Hopper) demand, this helped quell concerns about a potential ‘air pocket’ this year, as well as sustained revenue growth into 2025 and 2026.
Advanced Micro Devices (AMD) reported in-line results and guidance, with strength in data centre revenue driven by the ramp up of the company’s MI300 AI GPUs (graphics processing units) and market share gains in server CPUs (central processing units) from Intel, offset by continued weakness in gaming and communications markets. Management noted MI300 is the fastest ramping product in the company’s history and raised their revenue guidance for the product from $3.5bn to >$4bn in 2024. The debate about ultimate market share will continue, however we have confidence in these numbers given the urgency to build out AI infrastructure, as indicated by the positive revisions to capital expenditure forecasts from the hyperscalers (the largest cloud service providers). If anything, AMD and others remain supply-constrained, with TSMC CoWoS and high bandwidth memory (HBM) in tight supply.
Many other companies with AI exposure enjoyed a strong month, particularly those with high correlation to NVIDIA. Disco, a leading semiconductor production equipment vendor, benefited as it produces equipment for the manufacture of HBM used in NVIDIA solutions. Fabrinet is a supplier of complex optical and electro-mechanical components to NVIDIA. The company had a strong quarter driven by its Datacom segment, which grew +150% y/y due to AI demand, more than offsetting softness in other areas such as automotive. In contrast, Samsung* continued to struggle with HBM yields and underperformed the market.
First-quarter earnings season continued to be mixed, both in terms of results and market reactions, although AI-exposed companies delivered stronger results, led by NVIDIA.
Pure Storage, another potential AI winner, rallied before reporting solid results, with management calling out strength in enterprise and international markets. Management noted the quantity and quality of its discussions with hyperscalers “advanced considerably during the quarter” and continue to believe the company will secure a design win this year. Arista Networks also raised their full-year revenue forecast due to strong momentum in the first quarter, driven by strength in the enterprise market and cloud service provider demand. Management reiterated their target for $750m of AI-related revenue in 2025, but confidence has increased, given all four of the big customer trials are now migrating to deployments.
In the internet subsector, Amazon reported solid results, with Amazon Web Services’ (AWS) growth accelerating to +17% y/y in constant currency (cc), benefiting from generative AI (GenAI) workloads. Management indicated they expect capital expenditure to be up meaningfully in 2024 to support further AI infrastructure buildout. However, second-quarter revenue growth guidance was modestly below expectations, raising concerns about consumer sentiment. E-commerce platform Shopify reported gross merchandise value (GMV) growth of +23% y/y, but this did not meet elevated expectations as healthy trends in North America were offset by weaker consumer spending in Europe (just as Starbucks*, Amazon, Disney* and others have seen some low-end consumer softness).
Cloud connectivity platform CloudFlare delivered strong results, however management gave conservative guidance given macroeconomic headwinds and recent changes in sales leadership which are expected to have a positive impact in the long term. The company has launched GenAI features on its Workers app development platform but is not yet monetising them, instead focusing on driving adoption among its two million developers.
Elsewhere in the software market, weak results and negative revisions among some of the largest application companies including Salesforce*, Workday* and Intuit* disappointed, while weak Q1 results and guidance at MongoDB weighed on the consumption/usage-focused software group, including Confluent. Some software companies are facing pressure when renewing seats-based contracts against a tighter hiring environment, especially as enterprises have one eye on the potential productivity gains AI may help generate from existing staff. Additionally, some consumption software companies have yet to enjoy the same tailwinds enjoyed by the hyperscalers following a prolonged period of cloud optimisation. This may reflect nascent AI disruption, manifesting as ‘macro’ or SMB (small and medium-sized business) uncertainty; we see a particularly uncertain future for the software industry which is why we have continued to reduce exposure and recently exited our remaining positions in Workday and Salesforce.
The one exception to our software concerns is cybersecurity. AI is likely to significantly change and broaden the threat landscape and vendors embracing AI to counter this threat are likely to thrive. CyberArk Software delivered a solid quarter, with revenue up +37% y/y and margin and cashflow headwinds from the business model transition abating. The company also announced the acquisition of Venafi, a leading provider of machine identity security, expanding the addressable market by $10bn to $60bn. The share price of cybersecurity peer Crowdstrike Holdings recovered after a pullback in April despite strong results, benefiting from announcements and meetings at the RSA security conference, which reinforced strong fundamentals.
Semiconductors have eclipsed software as the largest technology subsector in the S&P 500 for the first time.
There was a relief rally in Apple (underweight (u/w)) shares after the company posted better than expected results, with revenue down -4% y/y. Revenue from China only declined -8% y/y (improving from -13% y/y last quarter), better than indicated by third-party data leading into the earnings report, although concerns about market share losses in the region remain. Services growth remained healthy at +13% y/y and the company announced an incremental $110bn share buyback, the largest in company history ($90bn last year). While Apple appears to be a laggard in AI, a rumoured partnership strategy (with either Google or Open AI) along with GenAI-focused devices may still be sufficient to drive a product refresh cycle.
Outlook
Higher interest rates (or lack of expected cuts) and the uncertain macroeconomic backdrop have weighed on software spending, initially for smaller customers but increasingly for enterprises. AI spending also appears to have ‘crowded out’ some spending on other IT initiatives, including software projects. AI has even caused some enterprises to adopt a more considered approach to their software investments as the frenetic pace of model improvement and architectural shift to a parallel computing world increase uncertainty (even while enhancing technological capabilities); there is no clear best practice in a domain evolving this fast.
We have been reducing our exposure to software for some time as we have been unsure if the benefits of AI will accrue to the application layer, and more importantly whether listed software companies of a ‘pre-AI’ vintage are best placed to capture them. Sector weakness has left software stocks looking more attractively valued with average multiples (6.3x EV/NTM sales) below both the Covid trough (c7x) and five-year average (10x). In the past, we might have bought this dip, but sales growth expectations continue to decelerate, risk-free rates are higher and recent misses from both Workday* and Salesforce* suggest that AI disruption – to budgets, sales cycles or headcount – may be playing a part.
Trends in AI adoption and GenAI continue to accelerate, with annual spending estimated to reach $1.5trn this year, rising to $2.3trn by 2028, according to Gartner.
Semiconductors have eclipsed software as the largest technology subsector in the S&P 500 for the first time. Software companies have struggled to keep pace with semiconductor companies exposed to AI infrastructure investment, which have benefitted from positive cloud capital expenditure (capex) revisions and increasing investor enthusiasm. The Trust has been overweight companies exposed to the buildout of AI infrastructure (led by NVIDIA) which have proved the antithesis of the stuttering software trade.
Trends in AI adoption and GenAI continue to accelerate, with annual spending estimated to reach $1.5trn this year, rising to $2.3trn by 2028, according to Gartner. According to IDC, spending on GenAI alone is set to double in 2024. A McKinsey Global survey this month found 65% of respondents reported their organisations are regularly using GenAI, double the percentage 10 months ago. Klarna* reported Q1 revenue growth of +29% at the same time as operating expenditure declined -11% y/y as "90% of Klarna employees have integrated AI into their daily workflows” and AI was responsible for nearly 40% of the cost savings in the sales and marketing budget.
That said, it is important to remember that GenAI remains in its early stages and, as such, the implementations today are mostly early or experimental, leveraging older technology and models. Naturally, we expect significant and ongoing silicon and AI model improvements to improve capabilities dramatically, so much so that the focus on the return on investment of hyperscale capex and potential risk to margins misses the fact that the ‘race’ to artificial general intelligence (AGI) appears firmly underway. Just three years ago, the timeline to AGI looked to be 2050 or beyond; today, experts believe it could be achieved by 2030 with profound implications for life as we know it. This explains some of the urgency of AI spending, not just at the hyperscalers, but also by enterprises and nation states alike.
We continue to hold a portfolio of companies which we believe are well-positioned for the buildout of parallel computing architecture and the rebuilding of global data centres which could reach $1trn in incremental spending over the next 4-5 years, according to NVIDIA’s CEO. Even in this scenario, it will not be plain sailing; investors should be prepared for periods of elevated volatility tied to macroeconomics, geopolitics or AI-related news flow. As it relates to AI, this volatility is an expected and healthy dynamic during early accelerated adoption of a disruptive new technology architecture buildout.
* not held
