December Investment Views
Strategy
What is Artificial Intelligence?
- Defining AI is far from simple
- Generative AI shifts focus from calculation to creativity
- Artificial General Intelligence is the ultimate goal
We have recently reached the three-year anniversary of the launch of OpenAI’s chatbot ChatGPT. Some clients are very familiar with AI from personal use or adoption in their own workplaces, but we are conscious that for some clients the new technology remains a bit of mystery.
In truth, many investors have more questions than answers on how the AI landscape will evolve. There is a general consensus that AI is a significant technological breakthrough but it is too early to tell whether it is as consequential as electricity, which companies will profit from it, which will be disrupted, and what it really means for the jobs market.
Some historical context is helpful. Alan Turing famously cracked the Enigma code of scrambled messages during the Second World War. Turing’s work combined machines with human reasoning to rapidly test code combinations. The machines themselves did not think or learn. Instead, cracking the code involved the very quick testing of millions of combinations. There is a key distinction between mathematics and machines following rules, compared to machines actually being able to think and learn.
“Artificial Intelligence” was first coined in the 1950s, but without sufficient computing power at a reasonable cost, progress was slow. Rapid advances in semiconductors in the 1970s and 1980s then enabled a technological resurgence for software and AI in the 1990s. The next milestone was IBM’s Deep Blue computer, famous for defeating world chess champion Garry Kasparov. But again, this wasn’t really intelligence. The machine could calculate millions of chess positions per second and “think” 10 moves ahead, but it wasn’t really thinking. It was maths and high-powered engineering based on rules written by humans.
ChatGPT was such a pivotal moment for AI because it ushered in an era of Generative AI. The introduction of the transformer in 2017 was a key enabler of generative AI as this was a breakthrough in the way AI models are designed. This massively improved the scalability of models to billions of data points. Rather than solve one logical problem, such as chess, Generative AI can handle many different tasks, recognise patterns, or hold human-like conversations. For example, one can ask AI to “write a poem about the 2025 Formula 1 season in the style of Shakespeare”. This is not a maths challenge like chess, or testing millions of combinations like breaking the Enigma code, it requires a form of thinking.
The range of what Generative AI can be used for is vast. From writing computer code, generating text, images, video, songs, and creating music. It can summarise long documents and analyse huge amounts of data. It can also turn text descriptions into visual and audio creations, read seismic data for mining, read medical images such as X-rays, and improve autonomous driving technology.
Semiconductor firm Nvidia has been the biggest beneficiary of AI. They offer the cutting-edge chips for training AI models. The company has grown to be the largest in the S&P 500, at a huge 7.7% weighting in the index. AI is also impacting the stock market in a number of other ways. Data centres are needed to house the semiconductors and these data centres require a lot of energy. Then there are the AI models of which ChatGPT, owned by OpenAI, is arguably the most well-known with 800 million weekly active users. However, there are many different models, including Google Gemini, Claude (owned by private company Anthropic), Grok (owned by private company xAI), and Deepseek (a privately owned Chinese company).
Applications are the final piece of the AI jigsaw. These can be apps on a smart phone or tools used by companies with potential use cases in every industry. Companies can use AI to help them dynamically price their goods or services, for example an airline maximising revenue while also minimising empty seats on flights. Facebook, Instagram, and YouTube have used AI to understand more about users and pattern recognition, which then helps target advertising more efficiently.
AI’s ability to write computer code is already disrupting the job market for software developers. AI has the ability to mimic human-like mental abilities, but whether it is really “intelligence” is still debated. Technology companies are investing heavily to both lead in AI but also in the race for Artificial General Intelligence (AGI). This is the ability for machines to understand, learn, and then apply knowledge to solve intellectual tasks in the same way as humans. This remains elusive, and there is considerable debate around whether AGI is two years away, 20 years away, or if it is even possible.
For markets, AI has been a key driver of returns for the last three years, but we have now reached a more challenging phase of the AI cycle; a lot of money has been spent, yet profits remain uncertain. The market is now closely discriminating between winners and losers from the AI revolution. We have responded to this shift by focusing more on company specific factors compared to big-picture economics, which struggles to capture everything that this happening under the surface.
Fixed Income
Elevated Uncertainty Heading into 2026
- Data thin, policy uncertainty dominates
- Global macro cycles continue to diverge
- US credit and Treasuries remain subdued
November opened with continued economic disruption, but the 43-day government shutdown ended on the 12th November. Key economic releases were lost, confidence in near-term data quality was undermined, and markets were left trying to infer economic momentum with few reliable signals. At the same time, the Federal Reserve moved closer to a major operational shift as they begin to expand, rather than contract their balance sheet.
Against that backdrop, expectations for a December rate cut lurched sharply back and forth through the month, reflecting shifting interpretations of incomplete data and renewed focus on internal Fed dynamics. This includes ongoing speculation around the next Chair and their influence ahead of the December meeting. The result was a month in which conviction was scarce, narratives moved quickly, and the bond market was forced to price policy uncertainty rather than clean macro trends.
In the US, the macro picture remained more resilient than many had expected. The Atlanta Fed’s GDPNow estimate for Q3 maintained a strong reading of 3.9% in real terms, reinforcing the view that the economy is buoyant and broadly consistent with trend growth of around 2% for 2025. Surprisingly, US Treasuries outperformed all major developed bond markets, with the 10-year yield drifting lower, ending the month at 4.02%. The shape of the yield curve remained largely unchanged. Inflation expectations moved lower despite still-sticky realised inflation, a signal that markets remain comfortable with the medium-term disinflation story, even as the shutdown left large holes in the data. The data that did emerge pointed to a modest reacceleration in activity, awkwardly arriving just as the Fed restarted easier monetary policy. The labour market continues to complicate the picture: hiring has slowed materially but layoffs remain limited, leaving policymakers facing an economy that is cooling at the margin but without the weakness in the jobs market to justify a faster or more confident easing path.
Across Europe, the UK and Japan, November highlighted how divergent late-cycle dynamics have become. With Japan’s domestic data strengthening, inflation still elevated and persistent pressure on the yen, markets have effectively forced the Bank of Japan’s hand. The upcoming meeting is now widely expected to deliver a rate increase, the first since January, reflecting both economic momentum and the need to reassert policy credibility as JGB yields and the currency signal rising discomfort with prolonged accommodation.
In the UK, the policy challenge is very different. The latest budget reinforced a stagflationary (inflation and weak growth) outlook, with higher taxes and tighter fiscal settings weighing on already weak growth and rising recession risk. That backdrop is broadly supportive for gilts, particularly at the front end, but leaves the long end vulnerable to ongoing supply and credibility concerns.
By contrast, the euro area appears to have threaded the needle: inflation has slowed materially without a collapse in activity, allowing the ECB to signal that its easing phase is effectively complete. The forward path now hinges less on monetary policy and more on fiscal choices. The scale and composition of German spending will play a decisive role in shaping Europe’s growth and rates outlook into 2026.
Corporate credit markets stayed resilient, with the spread (premium that investment-grade companies pay to borrow) remaining low. High yield bonds continue to offer attractive income, supported by benign default rates and healthy access to funding. Looking ahead, while average yield premiums over US Treasuries of 2.69% on high yield bonds and 0.8% on investment grade bonds remain far from cheap, interest rate cuts and higher levels of government spending in China, Europe, and the US are all supportive factors for corporate bonds.
Geopolitics remained a secondary but persistent source of risk rather than a primary driver. The war in Ukraine stayed entrenched, with no change in the stalemate despite efforts to sign a ceasefire. Tensions around Venezuela edged higher, as the US increased its military presence in the region and sharpened its focus on the country’s political trajectory. For fixed income investors, Venezuela matters mainly through the energy channel and confidence: any credible path to a more market-friendly regime with higher and more stable production would likely lower risk premia over time, but the route is uncertain. More generally, November showed that even without a single dominant shock, the interaction of policy uncertainty, fiscal sustainability and geopolitics is enough to keep term premiums rising and long-end volatility firmly in play.
Equities
Another Year of Double-Digit Returns
- A broad-based equity market rally
- Software again under pressure
- AI ambitions increasingly funded by debt
The MSCI World Index posted eight consecutive positive months, with only two months of declines this year. As we close out the year and head into the festive season, equity investors are celebrating a third straight year of double-digit returns, well above the long-term expectations for equities. Unlike 2023 and 2024, when market gains were driven by a handful of sectors, 2025 saw a broader rally.
Technology was only the fourth-best performing sector so far this year, while Industrials and Utilities, key players in AI infrastructure, took the lead. This year's returns have predominantly been driven by earnings growth, while valuation multiples have largely remained stable, albeit at elevated levels. A combination of S&P 500 earnings growth projected at 13% for next year and anticipated interest rate cuts provide continued support for the market.
In November, the MSCI World Index rose by 0.3%. Healthcare emerged as the leading sector with a strong 8.2% gain, rebounding after underperforming over the past three years. This surge was primarily driven by large market capitalisation pharmaceutical companies reporting robust earnings, increased merger and acquisition activity, and a series of positive developments. In contrast, technology lagged, becoming the worst-performing sector with a 4.7% decline, following two months of impressive results.
Software stocks dropped by 7.4% amid prevailing speculation that AI could disrupt the industry’s future. Oracle's gains were erased as market enthusiasm waned, with shares returning to June levels and sitting 45% below their September peak. Investor anxiety grew over Oracle’s uncertain growth prospects and rising debt. Additionally, disappointing revenue and unexpectedly high capital expenditures in Oracle’s latest earnings report prompted further selling. The semiconductor sector also fell by 5.8%, with Broadcom outperforming Nvidia as investors hedged their bets on which company would lead the semiconductor market, Broadcom gained momentum with Hyperscalers through custom chip offerings.
The third quarter earnings season brought reassurance, with unusually few downgrades beforehand and an impressive 6% beat on earnings expectations. Earnings grew by 13% year-on-year, underscoring a growing economy. The season began with major banks reporting robust capital market activity. Financials have become the second-best performing sector so far this year, benefiting from increased equity capital markets activity, partly spurred by deregulation. Large market capitalisation technology companies experienced a mixed season: Nvidia exceeded forecasts, but its shares dropped on the day; Amazon surged 10% after strong cloud computing results, while Alphabet rose 3% due to better-than-expected earnings and increased capital spending. Meta was the main loser, falling 11%.
Alphabet, Meta, and Oracle issued debt to support their growing AI ambitions, yet only Alphabet has been rewarded, thanks in part to the release of Gemini 3, which shifted its status from AI laggard to leader. Investors continue to support Alphabet despite its significant investment requirements. In contrast, Meta's comparable level of investment raises concerns among some stakeholders, given its market capitalisation is approximately half that of Alphabet. Similarly, Oracle faces scrutiny due to its increasing debt levels and the uncertainty surrounding future returns.