Key points

• AI continues to dominate as a transformative technology, driving productivity gains and revenue opportunities across sectors
• Large-cap technology stocks, particularly AI-focused companies, outperformed in 2024
• The technology sector remains well-positioned, providing a favourable environment for AI-driven growth and investment opportunities
  

Market review

The MSCI All Country World Net Total Return and S&P 500 indices both declined by -0.8% in December, while the DJ Euro Stoxx 600 declined by -0.9% (all returns in sterling terms).

December weakness belied a strong year for global equity markets. The MSCI All Country World Net Total Return Index gained +19.8%, while the S&P 500 Index returned +27.4%, significantly outperforming the DJ Euro Stoxx 600 Index rise of +4.5% (all returns in sterling terms).

The strength of the US equity market reflected better than expected economic growth while the Federal Reserve pivoted to an interest rate cutting cycle, enabled by moderating inflation. Donald Trump’s decisive victory in the US presidential election alleviated political uncertainty and prompted the biggest ever post-election market rally. The Republican ‘clean sweep’ further buoyed sentiment, increasing expectations for pro-business policies (deregulation, lower taxes) that should hopefully offset potential negative effects from tariffs. As the major macroeconomic and political risks dissipated over the year, equity valuation multiples expanded and credit spreads tightened.

Unfortunately, the post-election rally lost steam in December due to worries that some of the President-elect’s policies, such as tariffs, may prove inflationary. While the Federal Reserve (Fed) cut the federal funds rate at the December meeting as anticipated, this was accompanied by upward projections for the path of future interest rates and inflation, in part due to potential policies from the incoming administration. The 10-year Treasury yield jumped 40 basis points (bps¹) during the month to 4.58%, up sharply from 3.88% at the start of the year.

The 25bps cut to the federal funds rate at the December meeting marked the third consecutive reduction, lowering borrowing costs to the 4.25-4.5% range, in line with expectations. However, Fed Chair Jerome Powell called the rate cut a "closer call" than recent decisions with monetary policy² now "significantly less restrictive" and "significantly closer to neutral".

The Fed also revised its GDP growth and personal consumption expenditure (PCE) inflation forecasts upward for 2024 and 2025 while the committee’s ‘dot plot’³ indicates that policymakers now anticipate just two interest rate cuts in 2025, totalling 50bps, compared to the full percentage point⁴ of reductions projected in the previous quarter.

Technology review

The technology sector significantly outperformed the broader market in December. The Dow Jones Global Technology Net Total Return Index gained +3.6% (all figures in sterling terms). The large-cap Russell 1000 Technology Index gain of 3.8% saw it outperform the small/mid-cap Russell 2000 Technology Index which was flat during the month. Positive AI sentiment saw semiconductor stocks rebound, the Philadelphia Semiconductor Index gaining 2.9%, in contrast with the iShares Software index which declined by 3.9%.

For 2024, the technology sector once again outperformed the broader market with the NASDAQ Composite Index delivering positive returns for the 18^th time in 22 years, in dollar terms The Dow Jones Global Technology Net Total Return Index gained +35.9%, significantly outpacing the MSCI All Country World which returned +19.8%.

This outperformance was driven by US large-cap technology stocks which significantly outpaced US small caps. During 2024, the Russell 1000 Technology Index (large cap) gained 40.9% compared to the Russell 2000 Technology Index (small/mid-cap) which rose +26%. US large-cap outperformance also led the US to significantly outperform international markets, the Dow Jones World ex-US Technology sector rising just +16.6% during the year.

The narrowness of the technology market during 2024 reflected the remarkable performance of the Magnificent Seven (+71%), which once again led markets higher.

The narrowness of the technology market during 2024 reflected the remarkable performance of the Magnificent Seven⁵ (+71%) which once again led markets higher. This select group of companies continued to benefit from positive earnings revisions and excitement about artificial intelligence (AI), accounting for almost 60% of the S&P 500’s return. NVIDIA was again the standout performer, returning +177%, driven by exceptional growth (the forecast for the year to the end of January 2025 sees revenue up +112% y/y and GAAP earnings +144%).

Generative AI (GenAI⁶) remained the dominant market theme, with significant dispersion in performance between perceived AI winners and losers. Key enabling technologies supporting AI infrastructure performed well across the compute, network and power sectors as hyperscalers⁷ raced to build out their AI offerings and train ever-larger models.

Internet stocks outperformed in 2024, gaining +32.5%, ahead of software, up +25.8%. Despite the semiconductor sector being a key focus for investors, the Philadelphia Semiconductor Index underperformed the wider technology market, gaining +22.7% for the year as non-AI/more cyclical end markets (such as PC, smartphone, industrial and automotive) disappointed.

Returning to more recent developments, news flow during December remained heavily biased towards AI infrastructure companies with strong results from Broadcom, Marvell Technology (Marvell), Credo Technology Group Holding (Credo) and Ciena and the tone here remained supportive. Elsewhere, however, stocks such as Micron Technology (Micron) and Samsara underperformed following mixed quarters.

Broadcom posted results largely in line with consensus, with AI semiconductors now accounting for more than 40% of total semiconductor revenues, growing +150% y/y in the quarter. The stock reacted very positively to commentary around the ASIC (application-specific integrated circuits) opportunity where the company sees a serviceable addressable market (SAM) of $60-90bn in FY27. In addition, Broadcom announced two additional hyperscaler wins to develop custom AI chips with line of sight to revenue generation before 2027.

ASIC chipmaker peer Marvell also delivered a strong quarter. Data centre revenue grew +25% and 98% quarter on quarter (q/q) and y/y respectively, led by strong ASIC and optical growth. More importantly, Marvell announced an expanded strategic relationship with Amazon through a five-year agreement for its custom silicon (ASIC) efforts for both AI training and inference. Astera Labs also benefitted from ASIC-related excitement given its platform of data centre connectivity products.

In memory, industry bellwether Micron reported a solid quarter including a doubling of high bandwidth memory (HBM) revenue – memory chips used in AI systems – and raised its total addressable market (TAM) outlook to $100bn by 2030. However, the company guided near-term results well below expectations due to elevated customer inventories in the PC and smartphone sectors.

Oracle reported in December, with GPU (graphics processing unit) consumption in its OCI (Oracle Cloud Infrastructure) business up +336% in the quarter. The company expects Cloud revenue to reach $25bn this year and announced a new agreement with Meta (Facebook) for them to use Oracle’s AI cloud infrastructure and collaborate on the development of AI agents based on Meta’s Llama models. However, due to incremental foreign exchange headwinds of c200bps the company will need to see acceleration in F4Q to achieve its FY25 targets.

In the software sector, Samsara grew revenue by +36%, but shares were weak as net new annual recurring revenue fell below buyside expectations, while revenue guidance of 30% was also slightly below Street expectations.

Elsewhere, Tesla was strong as Elon Musk’s expected prominent position in Donald Trump’s administration could support the company’s growth ambitions. Apple and Google also posted strong returns as mega-caps outperformed, while Shopify and DoorDash experienced some year-end profit-taking.

Technology outlook

The market reversal into the end of the calendar year was a useful reminder that there remains a wide range of potential outcomes under a Trump administration. His unpredictable rhetoric on a range of issues in early 2025 has exacerbated uncertainty although we are hopeful that these represent starting points for a broader negotiating strategy. In addition, elevated volatility is normal at the outset of a new technology cycle of this magnitude, when the innovation curve is at its steepest and both the pace and scale of the opportunity ahead are hard to define.

Valuations in the broader market and within the technology sector are towards the high end of historical ranges as the transformative potential of AI has begun to be reflected in fuller multiples. The S&P 500 IT Sector Index trades at 35.5x price/earnings, or 1.44x the market multiple. A growth premium is not unusual, and we expect multiples for ‘AI winners’ (enablers and beneficiaries) across all sectors to remain elevated, allowing strong revenue/earnings growth to drive absolute returns. In addition, if the US economic outlook remains robust, we see the potential for upside to estimates as 2025 progresses and AI revenue and efficiency tailwinds begin to contribute.

Indeed, the most significant area to monitor for the technology sector in 2025 will be the evolution of leading-edge AI models, which will continue to shape both the AI supply chain and the prospects for AI beneficiaries. The course and velocity of AI model development will also directly inform hyperscale capital expenditure: Goldman Sachs currently expects 24% growth in aggregate cloud capital expenditure to >$260bn in 2025 (Microsoft; Alphabet; Amazon; Meta) following a 70% estimated increase in 2024.

The hyperscalers have strong competitive incentives to continue investing aggressively and investors seeking a straightforward, visible ROI (return on investment) on these generational investments may be disappointed by a skew towards efficiency and cost benefits rather than incremental revenue streams, the details of which corporates may be less willing to share. Academic studies have found a 23% uplift in labour productivity, while anecdotal commentary from companies implies c30%, according to Goldman Sachs. This is extremely encouraging at such an early stage while the underlying technology continues to evolve at a rapid pace and best practice is many years from even being established.

We expect 2025 to be seen as the year when the benefits of GenAI become more widely understood. Today, much of the innovation and experimentation is being done behind closed doors (internal research projects and learning by experimentation) and often for competitive reasons not openly discussed. However, we believe the productivity gains and/or revenue opportunities for those who embrace AI will become more obvious as the year progresses. This should support a wider range of stocks (AI beneficiaries) across a broad array of sectors which in turn will support continued investment and growth in spending on AI infrastructure.

While the timeline of NVIDIA’s flagship Blackwell processor build (specifically the GB200) has frustratingly slipped due to typical design, manufacturing and supply chain issues, we believe the scale of the opportunity continues to grow. New product ramps are rare at this scale, and few are so keenly watched with the additional scrutiny adding to early cycle volatility.

Our base case is that the rapid progress in the AI story should be more than sufficient to support strong returns absent a significant macro shock or political crisis.

The pace of AI innovation remains ferocious with two significant developments in the past few weeks. DeepSeek v3 was launched by a Chinese company which claims similar performance to OpenAI’s GPT-4o and Claude 3.5 Sonnet but achieved with significantly lower training costs. This raises obvious questions about the technological advantage the largest model builders are likely to enjoy. However, OpenAI also announced its o3 model which showed better than human performance on the ARC-AGI benchmark⁸ and achieved a score that would have put it within the top 200 of the Codeforces competitive coding leaderboard. The creator of the ARC-AGI benchmark described o3’s results as “a surprising and important step-function increase in AI capabilities, showing novel task adaptation ability never seen before in the GPT-family models. All intuition about AI capabilities will need to get updated for o3”.

The fact that o3 was able to achieve such significant breakthroughs just three months after the launch of OpenAI’s first reasoning model (o1) suggests test-time compute scaling on the inference side is capable of meaningful AI performance improvements faster than when relying solely on LLMs and their months-long training runs. A new vector for scaling model performance could have big implications for timelines to AGI/tasks in scope and the pace of capital deployment.

In addition to these new scaling vectors, xAI (founded by Elon Musk in 2023) and Meta are expected to deliver the next generation of frontier models trained on >100,000 GPU coherent clusters during the first few months of 2025, the next major milestone for the anticipated continuation of pre-training scaling laws for LLMs (which state that performance predictably improves as model size, computational resources and training data increase). NVIDIA’s Blackwell architecture will also be rolled out more widely and the company is expected to announce details of the second-generation B300 at its GPU Technology Conference in March before the new Rubin platform expected in 2026.

Returning to the economic outlook and political influences, concerns around changes to US trade, fiscal and immigration policy are understandable – and the recent move higher in US bond yields highlights the market’s sensitivity, as well as providing an unhelpful headwind for longer-duration (particularly unprofitable) stocks. However, this sensitivity should also help contain the worst-case scenarios; despite its more hawkish tone at the December meeting, the Fed (and other central banks) are still likely to respond quickly and decisively to any meaningful softening in the labour market.

Our base case is that the rapid progress in the AI story should be more than sufficient to support strong returns absent a significant macro shock or political crisis. The Trust is still positioned heavily in favour of AI winners and the new US administration should help ensure the US remains one of the stronger global economies and at the forefront of GenAI development. While regulation (particularly of Big Tech and GenAI leaders) will rightly remain a feature, a more pro-business environment may help overcome regulatory barriers to AI experimentation and adoption given the huge potential productivity opportunities.

Investors should not underestimate the significance of enterprise IT budgets and AI projects. Technology research specialist Gartner expects IT spending to grow by nearly double digits in 2025 (+9.5% y/y), the highest annual growth rate since 2011. This is anticipated to follow 7.2% growth this year, marking the strongest back-to-back increase in expected annual IT spending so far this century. The RBC Capital Markets’ December CIO survey found 39% of respondents are already in production with AI projects, with another 43% expected to be in production in the next six months.

Looking ahead in 2025, we are hopeful that the technology market should broaden as a wider range of companies both enable and benefit from AI advances. However, the first weeks of January are often a challenging period for the previous year’s winners, with preannouncement season ahead and additional uncertainty associated with a new presidency. This year, equity markets also have to contend with sharply higher risk-free rates. While we have tactically taken profits in some winners in anticipation of further volatility, we remain extremely bullish on the outlook for AI-related assets and envisage moving back to a fully invested position in the coming weeks.

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

2. Monetary policy is an action that a country's central bank or government can take to influence how much money is in the economy and how much it costs to borrow

3. A chart that records each Fed official's projection for the central bank's key short-term interest rate

4. The difference between percentages (i.e. a 1 percentage point fall takes 10% to 9%; a 1% fall takes 10% to 9.9% - a 10% fall takes 10% to 9%)

5. Apple, Microsoft, Amazon, Alphabet (Google), Tesla, Meta (Facebook) and NVIDIA

6. GenAI: a type of AI that can create new content and ideas including conversations, stories, images, videos and music

7. The largest cloud service providers (AWS; Microsoft Azure; Google Cloud; Meta Platforms; Apple; TikTok)

8. Abstraction and Reasoning Corpus for Artificial General Intelligence, measure of general intelligence and skill-acquisition efficiency in AI systems