Key points

• Equity markets suffered a setback as elevated policy uncertainty has weighed on sentiment and the growth outlook in the near term
• We remain convinced that the AI story represents the most exciting opportunity in the market today, albeit more complex and potentially volatile post-DeepSeek
• This year, we expect to see AI make significant further gains that begin to change how industries operate and enable the creation of large new markets
  

Market review

Global equity markets were bifurcated in February, the MSCI All Country World Net Total Return delivering -1.7%, while the S&P 500 and DJ Euro Stoxx 600 indices returned -2.4% and +2.1% (all returns in sterling terms).

Uncertainty was the flavour of the month. The Federal Reserve (Fed) held interest rates steady in January, while the minutes of the Federal Open Market Committee (FOMC) revealed significant uncertainty about the path of rates ahead. Meanwhile, market participants had to account for changing US government policy as the Trump administration has been assertive in advancing its goals – negotiating the end of the Russia/Ukraine war on the geopolitical front, implementing government spending cuts with the Department of Government Efficiency (DOGE) on the fiscal front and imposing tariffs on Mexico, Canada, China and Europe on the trade front.

Data released in February suggested the US economy has been resilient, with Q4 GDP (Gross Domestic Product) estimated to have grown at an annualised +2.3%, driven primarily by personal consumption. US growth appears to have slowed in recent weeks, driven by lower consumption, waning consumer confidence and falling sentiment among small businesses on tariff concerns. February’s consumer spending (which is around two-thirds of US economic activity) shrank, its first decline since March 2023 and the biggest decrease in nearly four years.

Consumer surveys in February also indicated a softer outlook, with the Conference Board’s Consumer Confidence Index falling the most since August 2021 (when the Covid Delta variant hit), while the University of Michigan’s Surveys of Consumers found 40% of respondents spontaneously mentioned tariffs as a concern, up from less than 2% before the election. Fed futures, reflecting views of the Fed’s future monetary policy, are now pricing three rate cuts this year, up from just one in the middle of the month as the market anticipates the Fed will be more concerned about the potential impact of softer growth on the labour market than the risk of a second wave of inflation.

Despite the current turmoil, we believe that the AI story represents the most exciting opportunity in the market today.The implementation of tariffs on Canadian and Mexican imports by the Trump administration, effective 4 March, coupled with ongoing tariffs on Chinese goods, has intensified concerns of a potential escalation into a full-blown trade war, with adverse effects on economic growth.

Employment numbers rose by 151,000 in February, up from a downwardly revised 125,000 in January. The unemployment rate ticked up to 4.1% from 4%, driven by a swelling labour force and a spike in part-time workers unable to secure full-time roles. Federal government employment dropped by 10,000, reflecting the early impact of the DOGE spending cuts.

Understandably, market volatility remains elevated, with the VIX – a volatility index – rising notably, underscoring investors’ apprehensions about the economic fallout from escalating trade disputes and unpredictable, aggressive policies emanating from the White House. Treasury Secretary Scott Bessent recently made clear the US economy may slow during this “detox period” as the economy transitions towards private investment (from public spending) and indicated the administration is not bothered by the stock market weakness. However, we hope we are close to peak uncertainty and the market may recover if some degree of policy clarity, or even simply less uncertainty, returns.

Technology review

The technology sector underperformed the broader market in February; the Dow Jones World Technology Net Total Return Index returned -4.2% while the Trust returned -6.8% (both figures in sterling terms).

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) returned -4.6% and -13.1% respectively. The Philadelphia Semiconductor Index (SOX) returned -6.0% while the NASDAQ Internet Index and IGV Software ETF returned -4.1% and -6.3% respectively.

After the release of DeepSeek’s R-1 model at the end of January, the market spent February digesting the potential implications for AI infrastructure and related capital expenditure (capex). China significantly outperformed, in part because it is no longer perceived as a loser from AI with DeepSeek efficiency gains and advances in its own semiconductor design and manufacturing, which means even the tightening of US restrictions may not derail its rapid progress. We also saw the bulk of Q4 earnings reports where we continued to see strength in the AI theme and further upward revisions to capex by the largest cloud service providers.

In the semiconductor sector, NVIDIA posted another ‘beat and raise’, growing +78% year on year (y/y) and delivering c$11bn of revenue from its new Blackwell platform in the quarter to the end of January, above market expectations of high single-digit billions. Blackwell processes 25x more ‘tokens’ at a 20x lower cost to the previous Hopper generation. The company guided to $43bn of revenue in its fiscal Q1 (to the end of April 2025), above consensus estimates of c$42.2bn.

In the near term, some NVIDIA uncertainty remains due to the transition from Hopper to Blackwell chips currently underway. We believe risks are largely overstated and expect strong demand (and increased supply and improving yields) to drive a robust 2H25 accompanied by gross margin improvement, as guided. That said, the implementation of tighter US export licence rules (agreed under the Biden administration in January) comes into effect in May and creates uncertainty in the interim (there is an issue of export licences perhaps being delayed by DOGE-related changes).

In February, NVIDIA peer AMD (Advanced Micro Devices) also reported seeing particular strength in PC sales, +23% quarter on quarter, while posting a slightly disappointing server number sequentially. The company sold >$5bn of its MI300 GPUs (graphics processing units) in 2024 but focus has now shifted to its next-generation MI355 product which has been pulled forward to the middle of the year. Meanwhile, AMD continues to see strong traction on its Turin CPU (central processing unit) gaining share from Intel.

Taiwanese smartphone System-on-Chip (SoC) maker Mediatek delivered a very strong Q4, driven by a doubling of its flagship SoC revenues to $2bn in 2024. The company guided above expectations as continued smartphone stimulus in China accelerated sales of its Dimensity 9400 chipset, while its ASIC (application-specific integrated circuit) business is on track to contribute >$1bn in 2026 at accretive operating margins. Arm Holdings reported another quarter of growth in both licensing and royalties at +14% and +24% y/y respectively. V9 adoption continues to show strength and the company is gaining share in data centres with AWS Graviton, Microsoft Cobalt, Google Axion and NVIDIA Grace.

In semiconductor capital equipment, Nova Measuring Instruments delivered strong results, with revenues and guidance ahead of expectations. ASM International printed reasonable numbers. Although orders missed, commentary around GAA and TSMC’s A16 node was very strong, with the company seeing a double-digit percentage increase in ALD (atomic layer deposition) intensity with each node.

Amazon saw solid growth in its AWS (Amazon Web Services) cloud business at +19% y/y which was constrained by supply (not expected to ease until later in the year). Operating margins at 11.3% were above estimates of c10%. However, guidance was disappointing on revenues and profitability due to FX (foreign exchange) headwinds and leap year impacts. The company spent $26.3bn in capex in Q4 and at this run rate expects to spend >$100bn in 2025, above expectations and a positive read for AI infrastructure. Meanwhile, Shopify printed a strong quarter with revenues up +21% y/y driven by merchant count, increasing penetration of payments and improving monetisation overall. The Q1 guide for revenue to grow in the mid-20s% was in line with analyst expectations but below some investors’ hopes.

Alphabet (Google) also posted solid results with search revenue in line with expectations and YouTube ahead. GCP (Google Cloud Platform) revenue growth missed expectations, however, which was surprising to the market. Similar to Amazon, the capex guide of $75bn was well above investor expectations. While the stock valuation is low, especially relative to growth, Alphabet still has to navigate the Department of Justice (DoJ) remedy and the ‘Ad tech versus DoJ’ trials, which present a potential overhang on its share price.

Meanwhile in China, Alibaba Group Holding (Alibaba) performed strongly during the month. Overall revenues grew 8% y/y and AI-related product revenue achieved triple-digit growth for the sixth consecutive quarter. More importantly, the company also noted it intends to spend c$52bn on cloud computing and AI infrastructure over the next three years, more than it spent on AI and cloud computing over the past decade. The company also recently benefitted from the 5 March release of its latest QwQ-32B (QWEN), a new 32 billion parameter reasoning model which matches DeepSeek’s performance with even smaller compute requirements.

Spotify Technology continued to execute well as it entered 2025, with accelerating monthly active user (MAU) growth, product enhancements in audiobooks and video podcasts, plus the ability to continue to drive both operating and gross margins higher. Food delivery company DoorDash saw a solid Q4 with orders growing 19% y/y. Restaurants continued to grow by strong double digits throughout the year with new verticals and rapid international growth.

Reddit grew +71% y/y, driven by both better than expected strength in its core advertising business and continued scaling of data licensing revenues. The guidance was strong at 48-52% y/y albeit below some heightened investor expectations and due to hits from slowing DAU (daily active user) growth.

Finally in internet/advertising, Applovin saw strong advertising revenues and software strength driving better than expected adjusted EBITDA¹. The guide for Q1 of 31% growth at the mid-point was also above expectations.

In cybersecurity, CyberArk Software delivered a clean quarter with annual recurring revenue (ARR) of +51% y/y and organic +30%. Q1 guidance for revenue at 37% y/y was slightly better than expected and accompanied by healthy operating margins and FCF. Cloudflare delivered 27% y/y growth despite some FX headwinds and FCF margin compression due to a ramp in its infrastructure build. 2025 guidance came in above expectations at 25% y/y and key performance indicators (KPIs) were supportive of growth having bottomed, with net revenue retention expanding, and paying customer growth accelerating to 25% y/y.

AI’s adoption is clearly expanding beyond the technology sector, with significant applications emerging in healthcare, finance and manufacturing, potentially broadening investment opportunities.

Turning to infrastructure software, Snowflake saw product revenue growth of 28% which was nicely above expectations. Remaining performance obligation (RPO) was also strong at 33% and customer adds accelerated. The Q1 guide was slightly soft due to tough quarterly comparisons and leap year impacts, but FY26 (to 31 January) guidance was largely in line.

Elastic also delivered a strong beat on both the top and bottom line while raising its FY25 (to 30 April) guidance once again. The company is seeing strong interest from custom building out of generative AI (GenAI) applications and the sales team continues to execute well. Gross margins were a little light but operating margin was very strong. Twilio delivered Q4 growth of c11%, above the guide and FY25 targets were reiterated. The Q1 revenue guide was slightly below expectations and management noted it expects a minimal AI impact on FY25 sales, albeit early proof points with Open AI and internal use cases are promising.

The networking and storage sector has struggled recently. Strong results were unfortunately overshadowed by uncertainty related to NVIDIA’s next-generation product roadmap (with more details expected at its GTC AI Conference). Coherent saw revenue and earnings per share beats and a strong gross margin as the new CEO implements his efficiency drive. Arista Networks posted a small beat and raised full-year guidance to 17% from 15-17%, which is atypical for the company at this point in the year suggesting good visibility. Revenue from key customer Meta declined to 15% from 21% due to its spending patterns, but the company is adamant that Meta revenue will grow this year and that this has been factored into guidance.

Pure Storage beat expectations on product and operating margins but its gross margin came in below. FY26 (to 1 February) guidance was broadly in line but Q1 guidance missed on operating margins.

Axon Enterprise delivered solid 4Q24 results with revenue of $575m, above consensus of $566m. Taser 10 continued to scale and customers continued their adoption of virtual reality training solutions. 2025 guidance was strong and total future contract bookings were solid at $10.1bn. The company closed its first 10 AI Era Plan deals in 4Q24 after launching in October. In FinTech, Robinhood Markets posted a very strong set of numbers, reaching 60% EBITDA margins in Q4 on the back of growth from all three transaction segments, while maintaining operating expense discipline. Revenue growth was driven by new product launches and strong retail crypto trading, which was the fastest growing segment.

Finally, DataDog posted a reasonable quarter but guidance was below expectations. In particular, both revenue and operating margins were light at 21% versus 25% in 2024. AI-native annual recurring revenue totalled 6% which was unchanged from the previous quarter despite volumes growing. We subsequently sold our small remaining position in the Trust.

Outlook

Following Trump’s election victory, the US equity market upgraded its economic growth expectations as optimism around deregulation, lower taxes and a revival of animal spirits offered further fuel to the ‘US exceptionalism’ narrative and cyclicals outperformed defensives. Equity markets initially took signs of political upheaval in their stride, but the emerging reality of the new administration’s policy agenda and modus operandi have been more challenging. The US waging a trade war with its top three trading partners representing over 40% of its imports, DOGE-related federal headcount and funding cuts, transgressing diplomatic norms and alienating historical allies have pushed economic policy uncertainty to its highest level since 2008. US policy – while presenting a headwind to markets – has also taken a further pro-AI pivot. Vice President JD Vance declared that America opposes the idea that government should direct AI’s development and excessive regulation of AI would “kill a transformative industry just as it’s taking off”.

Market action is in contrast to fundamental AI news flow.Furthermore, growth worries have emerged as policy/tariff uncertainty has weighed on consumer and small business confidence (and perhaps enterprise investment) leading to softer macro data (consumer spending; confidence; services PMI; ISM manufacturing data). The Fed will likely look through any short-lived tariff impact because its Federal Open Market Committee believes policy is already restrictive (i.e. the neutral rate is lower than the current Fed funds rate), but it will be monitoring future inflation expectations closely.

Equity markets have, understandably, responded negatively and the S&P 500 has given back all its post-election gains. Within the technology sector, a momentum unwind in small and mid-cap, long duration and AI infrastructure stocks following the DeepSeek announcement combined to create a perfect storm over the past several weeks.

Despite a flurry of AI announcements and an increase in the growth expectations of capex spending by the larger cloud service providers in 2025 to +32% from +20% at the start of the year, the GS TMT Growth Long Basket² dropped -20% from its highs. This has coincided with a significant TMT (technology, media and telecoms) de-grossing³ by hedge funds (similar to July last year on the yen carry trade⁴ scare and on par with the largest we have seen in the past five years) as well as single-stock selling by retail investors.

Market action is in contrast to fundamental AI news flow. AI scaling laws appear intact across three compute-intensive scaling vectors: pre-training (multi-modality, reasoning data), post-training (human reinforcement learning/AI feedback), test time/inference scaling (chain of thought). The latest generation of frontier AI models from xAI, Anthropic and Google are all setting new performance records against benchmarks at ever cheaper costs per token. ChatGPT’s Deep Research product launched for paying subscribers and is – in our view – astonishing. OpenAI Co-Founder Ilya Sutskever is raising $1bn+ for his superintelligence startup at a reported $30bn+ valuation, up 500% from the $5bn prior round in September, while xAI is in talks to raise $10bn at a $75bn valuation.

Despite the current turmoil, we believe that the AI story represents the most exciting opportunity in the market today. That said, we also acknowledge it has become a little more complex, and likely more volatile, due to the advent of new scaling vectors including innovations highlighted by but not exclusive to DeepSeek. This year, we expect to see AI make significant further gains that begin to change how industries operate and enable the creation of large new markets, many of which are unimaginable today. AI can address a much larger percentage of GDP than technology previously as it is integrated into every application: agentic AI (enterprise technology and knowledge work); physical AI (autonomous vehicles); industrial AI (robotics/manufacturing). The race to artificial general intelligence (AGI) is also underway, with the potential to unleash productivity gains and a potential for deflationary growth acceleration.

AI’s adoption is clearly expanding beyond the technology sector, with significant applications emerging in healthcare, finance and manufacturing, potentially broadening investment opportunities. On the regulatory front, governments are grappling with how to balance AI innovation with risk mitigation. Despite these challenges, the fundamental outlook for AI remains robust and the recent market pullback may present attractive entry points for investors with a long-term perspective. As we have regularly reminded investors, volatility is a persistent feature of new technology cycles: there were seven >15% corrections between 1995-98 (i.e. before the 1999 ‘melt up’), while the NASDAQ gained 354% over the same period. Our base case is that we are currently experiencing something similar and remain constructive on, in our view, the most transformative opportunity in the market.

This constructive view has been reinforced by recent company meetings. In early March, five of the Polar Capital Technology investment team attended several US conferences where they met with the management teams of many portfolio holdings (including NVIDIA and TSMC) across a range of technology subsectors, as well as attending presentations by Sam Altman (OpenAI), Dario Amodei (Anthropic) and Elon Musk (Tesla). Despite some uncertainty at the margin around the macro and tariffs, we felt these meetings were reassuring while confidence exhibited by both Altman and Amodei made post-DeepSeek’s stock action in AI training stocks look detached from fundamentals today.

Finally, several indicators we follow suggest we may have already experienced the worst of this market selloff. Markets appear oversold with relative strength indicators on both the NASDAQ and the S&P 500 indices falling below 30. Likewise, the magnitude of the current selloff (NASDAQ 100 -13%, S&P 500 -10%) is now commensurate with the median maximum drawdowns over the past 15 years (-12% and -10% respectively). Investor sentiment is in fear territory; the AAII Sentiment [bull/bear] Index recently registered its 13^th most bearish reading since 1960, and the lowest since 2022, while the CNN Fear & Greed index was recently at 15 (out of a maximum of 100).

We feel that with the combination of reassuring meetings and an oversold (and cheaper, to our mind) technology market, together with a constructive annual strategy paper recently completed by the team, has seen us move back to a more fully invested position in the Trust.

1. Earnings before interest, taxes, depreciation, and amortization; used to assess a company's profitability and financial performance

2. A portfolio created by Goldman Sachs focusing on long positions in companies within the TMT sectors

3. When long positions are sold down to buy back shorts

4. A strategy where one borrows at a low interest rate and reinvests in a currency/financial product with a higher interest rate