2025 market wrap-up: Volatility, AI dominance and geopolitics

As 2025 draws to a close, we can look back on a year that tested market resilience through extraordinary swings in both sentiment and fundamentals.

Despite headwinds including tariff shocks and central bank policy shifts, global stock markets managed to reach new record highs, powered by what appears to be an insatiable appetite for artificial intelligence investments.

Nvidia's wild ride

January 2026 began with one of the most dramatic market moves in history. Nvidia, the undisputed AI revolution trailblazer, experienced a staggering single-day market capitalisation loss of approximately $589 billion on January 27. This was by some margin the largest decline ever recorded for any company.

The catalyst for the fall was DeepSeek, a Chinese AI startup that claimed to have developed a high-performance large language model for a fraction of the cost typically spent by the US tech giants.

DeepSeek's announcement that it trained its R1 model for under $6 million (compared to the hundreds of millions spent by companies like OpenAI) sent shockwaves through the investment community.

The startup claimed it achieved this feat using fewer, less powerful Nvidia H800 GPUs; or in other words, chips specifically designed to comply with U. export restrictions to China. The immediate implication was that if AI models could be trained more efficiently with less expensive hardware, the massive capital expenditures that hyperscalers like Microsoft, Amazon and Google were planning might be excessive.

The tech-heavy Nasdaq plummeted 3% that Monday, with semiconductor stocks dropping across the board. Industry analysts scrambled to assess the implications, with some questioning whether the AI infrastructure boom had been built on unsustainable assumptions about computational requirements.

However, the panic proved premature. Market analysts quickly pointed out several critical factors that DeepSeek's claims overlooked. The reported $6 million figure didn't include hardware costs, pre-training expenses or the full infrastructure required to run the model at scale. Furthermore, several industry experts suggested DeepSeek likely had access to far more powerful chips than disclosed, including as many as 50,000 Nvidia H100 GPUs, but couldn't reveal this due to U. export controls.

Wall Street analysts from JPMorgan and Citigroup then argued that DeepSeek's efficiency breakthrough, rather than threatening Nvidia's business, could actually accelerate AI adoption by making it more accessible and cost-effective.

The argument they made, and continue to make, is Jevons paradox, where technological efficiencies that lower costs typically lead to increased overall consumption rather than decreased spending. If AI development became more efficient, demand for computing infrastructure would likely rise as more companies could afford to deploy AI solutions.

By late summer, Nvidia had not only recovered but soared to new heights. In October 2025, the chipmaker achieved another historic milestone, becoming the first publicly traded company to surpass a $5 trillion market capitalisation. The achievement came on the back of continued strong demand for its Blackwell GPU architecture and optimism about sustained AI infrastructure spending.

Despite falling back to $4.2 trillion today, Q3 results saw revenue hit a record $57 billion, up 62% year-over-year.

Founder and CEO Jensen Huang noted that ‘Blackwell sales are off the charts, and cloud GPUs are sold out. Compute demand keeps accelerating and compounding across training and inference — each growing exponentially. We’ve entered the virtuous cycle of AI. The AI ecosystem is scaling fast — with more new foundation model makers, more AI startups, across more industries, and in more countries. AI is going everywhere, doing everything, all at once.’

Tariff turmoil

If January's DeepSeek scare tested investor nerves, April's ‘Liberation Day’ tariff announcement shattered them completely.

On 2 April, President Trump unveiled a sweeping new trade policy featuring a universal 10% tariff on all US imports, plus significantly higher ‘reciprocal tariffs’ on dozens of trading partners. China faced a cumulative 104% tariff, the European Union 20% and Japan 24%, pushing the effective overall US tariff rate to approximately 25%, the highest in over a century.

The market reaction was immediate. Over the following two days, the S&P 500 plunged by more than 10%, marking one of its worst two-day performances since World War II, trailing only events like the pandemic crash, the Global Financial Crisis and Black Monday in 1987. More than $5 trillion in market value evaporated, equivalent to the entire economy of France. Both the S&P 500 and European indices teetered on the brink of bear market territory.

The tariff shock represented what one Morningstar analyst called ‘the most radical economic proposal of my lifetime.’ Unlike earlier trade tensions that had been largely priced into markets, the scope and immediacy of these measures caught investors off guard. Companies had almost no time to adjust supply chains or purchasing strategies before the tariffs took effect on 5 April.

Federal Reserve Chair Jerome Powell delivered perhaps the starkest warning possible on 16 April, acknowledging the ‘challenging’ impacts of tariff-related uncertainty while warning that the policies were ‘likely to generate at least a temporary rise in inflation’ with effects that ‘could also be more persistent.’

The tariff impact rippled through corporate America. During earnings season, companies across almost all market segments warned of higher input costs, with Nvidia among those highlighting increased expenses.

However, markets proved remarkably resilient. After the initial shock, President Trump and his economic team began selectively pausing tariffs for countries willing to negotiate. By summer, many of the United States' largest trading partners had come to the bargaining table, easing some of the most extreme provisions. The S&P 500 gradually climbed what investors termed the ‘Wall of Worry,’ eventually reaching new all-time highs despite the lingering uncertainty.

The recovery was aided by the fact that many tariff impacts were delayed or less severe than initially feared. While inflation did initially accelerate, the increases were more moderate than worst-case scenarios.

Central Banks pivot

Against this backdrop of market volatility and trade tensions, 2026 saw a significant shift in global monetary policy. After years of aggressive rate hikes to combat post-pandemic inflation, central banks around the world began cutting interest rates, though with varying degrees of conviction and at different paces.

The Federal Reserve led the way with three rate cuts in 2025, bringing the federal funds rate down from its peak to a target range of 3.50%-3.75% by December. However, the path forward for 2026 became considerably more uncertain.

In its December 2025 meeting, the Fed signalled that only one additional 25-basis-point cut was likely in 2026, a cautious stance reflecting concerns about persistent inflation pressures from tariffs and a resilient but slowing labor market.

Chairman Jerome Powell, whose term remains set to expire in May 2026, emphasized that the Fed had reached what he described as ‘the high end of a neutral range,’ meaning policy was no longer significantly restrictive but wasn't particularly accommodative either.

The unemployment rate had risen to 4.6% by November 2025, the highest in more than four years, while job creation had slowed to just 17,000 new positions per month in the latter half of the year. Yet inflation remained somewhat elevated, with tariff-induced price pressures complicating the picture.

Across the Atlantic, the European Central Bank took a more aggressive easing stance. After beginning its rate-cutting cycle in June 2024, the ECB delivered eight consecutive cuts, lowering the deposit rate from 4% to 2% by late 2025.

With eurozone inflation contained at around 2.1% and growth remaining modest but positive, ECB President Christine Lagarde declared the bank was ‘in a good place,’ suggesting rates may now steady barring any significant economic deterioration.

The Bank of Japan has moved in the opposite direction, continuing its slow normalisation of ultra-loose monetary policy. The BoJ hiked its policy rate to 0.75% in December, representing a historic shift for a central bank that had maintained rates near zero for decades. Concerns linger over the end of the Yen carry trade, which may persist into next year.

Meanwhile, China's People's Bank of China remained cautious, implementing only selective, gradual easing measures as it navigated trade tensions with the US, weak loan growth and concerns about financial stability. This measured approach reflected Beijing's delicate balancing act between supporting economic growth and avoiding excessive stimulus that could fuel capital outflows.

Central banks also continued to buy Gold at a record pace, with the precious metal rising to a record $4,400/oz towards the end of December. The metal was also perhaps driven by falling confidence in USD as a store of value, amid retail buying.

Silver has also reached record highs in 2025 due to a convergence of factors. Surging industrial demand from solar panels, electric vehicles, and AI infrastructure is colliding with stagnant mine production and supply deficits. This structural imbalance is amplified by safe-haven investment flows amid geopolitical tensions, falling real yields, and speculative momentum driving prices perhaps beyond what fundamentals alone would support.

AI frenzy

Despite the volatility from tariffs, geopolitical tensions and monetary policy uncertainty, 2026 proved to be another banner year for global stock markets, with artificial intelligence stocks serving as the primary engine of growth. The S&P 500 advanced approximately 15% year-to-date, repeatedly breaking new record highs, while technology stocks dramatically outperformed the broader market.

However, the USD (US Dollar) fell in value through the year, so the real rise may be smaller than first appears.

The concentration of market gains in AI-related companies reached unprecedented levels. Many constituents of the so-called ‘Magnificent Seven’ tech giants rose to valuations that would have seemed unthinkable just a few years earlier.

Capital expenditures on AI infrastructure reached staggering heights. Big tech companies collectively spent an estimated $405 billion on AI-related investments in 2025, a 31% increase in spending projections during the year alone.

This massive spending was driven by the race to build the infrastructure necessary for increasingly sophisticated AI models and applications. Hyperscalers like Amazon, Microsoft, Google and Meta announced that they would set new capital expenditure records in 2026, with most of that money directed toward constructing and equipping massive data centers.

These facilities required not just advanced chips from Nvidia, AMD, and others, but also enormous amounts of energy, cooling systems, networking equipment and physical real estate.

Nvidia remains at the epicenter of this AI boom. Despite the January DeepSeek scare, demand for the company's chips never wavered. The introduction of the Blackwell architecture in late 2024 had set new performance benchmarks, and by 2026, customers were clamouring for the even more advanced Blackwell Ultra chips. Nvidia's gross margins remained in the remarkable mid-70% range, demonstrating its pricing power in a market where demand far exceeded supply.

However, competition is intensifying. Alphabet's Tensor Processing Units (TPUs) emerged as a meaningful alternative to Nvidia's GPUs. Analysts have suggested that Alphabet could generate revenue equivalent to 10% of Nvidia's total sales from TPUs alone, potentially representing $31 billion in annual revenue.

The AI frenzy wasn't limited to chip makers and cloud providers. Companies across sectors integrated AI into their operations and products. Meta leveraged AI to deliver more relevant content and advertising, allowing it to charge premium rates to advertisers. In pharmaceuticals, AI-powered drug discovery moved from theoretical promise to practical application, with several AI-discovered molecules entering clinical trials. Financial services firms deployed AI for fraud detection, risk assessment, and algorithmic trading.

Yet concerns about an AI bubble grew louder as valuations stretched, while the concentration of market gains also raised systemic concerns. The largest companies (Nvidia, Microsoft, Apple, Amazon and Alphabet) have at times made up more than a quarter of major market indices by weight.

Critics also point to the circular nature of AI investments, where companies like OpenAI receive funding from Microsoft, Oracle, and Nvidia, then spent that money buying services and hardware from those same investors.

The key question for 2026 and beyond remains: when will AI investments start generating returns that justified the enormous capital expenditures? While early adopters reported tangible value creation, monetization is arguably still lagging capex.

Oil volatility

Global oil markets experienced dramatic swings in 2026, beginning with a sharp geopolitical shock in June before collapsing to multi-year lows by year-end. This volatility encapsulates how quickly market narratives can shift from supply disruption fears to oversupply concerns in commodities in general.

In early June, escalating tensions between Israel and Iran erupted into direct military confrontation. On 13 June, Israel launched preemptive strikes against Iran's nuclear facilities at Natanz and Isfahan, along with ballistic missile sites and senior IRGC commanders. Iran retaliated with approximately 200 ballistic missiles targeting Israeli territory, damaging the Haifa oil refinery.

Oil prices unsurprisingly surged, with Brent Crude jumping more than 10% from around $68-69 per barrel to approximately $78 per barrel. The spike reflected fears of potential disruption to the Strait of Hormuz, through which roughly 20% of global oil supply flows. Energy analysts modeled worst-case scenarios pushing prices above $100 or even $130 per barrel if the conflict escalated further.

However, the geopolitical premium proved short-lived. Within days, markets recognised that actual supply disruptions remained limited, with Iran's production continuing largely uninterrupted. Israel had deliberately avoided major oil infrastructure to prevent broader regional war, while the Trump administration reportedly restrained Israel from more aggressive actions. By late summer, prices stabilised around $70-75 per barrel.

The temporary calm gave way to a dramatic collapse driven by fundamental oversupply. Throughout the second half of 2026, oil entered a sustained downtrend as production surged while demand growth disappointed. OPEC+ unwound voluntary production cuts, adding approximately 2.9 million

barrels per day as Saudi Arabia prioritised market share over price support. US shale producers reached record output, while Brazil, Canada, Guyana and Argentina added significant volumes.

This supply surge collided with tepid demand growth. Weak economic conditions, improving vehicle efficiency, electric vehicle adoption and structural transportation shifts all weighed on consumption. China, previously a reliable source of demand through strategic stockpiling, pulled back sharply by late 2026, with imports and refinery runs declining to six-month lows.

Investment banks are issuing increasingly bearish forecasts, with JPMorgan projecting Brent at $58 per barrel for 2026 and Goldman Sachs even lower at $56. Most alarmingly, analysts warn that without coordinated OPEC+ production cuts, prices could fall into the $30s by 2027, reminiscent of the 2020 pandemic crash.

The collapse has mixed implications. For consumers and oil-importing nations, lower energy costs are providing relief and helping to contain inflation, potentially giving central banks more flexibility on monetary policy. For oil producers and exporting countries, however, sustained low prices threatened fiscal sustainability, with even major state producers like Saudi Aramco requiring higher revenues to balance government budgets.

The critical question for oil is now whether OPEC+ might implement significant production cuts to stabilise prices or continue prioritising market share. The answer will shape not only energy markets but broader economic conditions well into 2027, as the oil glut represents both a disinflationary tailwind for consumers and a fiscal challenge for producer nations.

2025 Market wrap-up summed up

• Nvidia lost $589 billion in a single day after DeepSeek's efficiency claims in January but rebounded to reach a $5 trillion market cap by October on sustained AI infrastructure demand
• Trump's April tariff announcement triggered a $5 trillion two-day market wipeout, but stocks recovered to new highs as tariffs were selectively paused and negotiated
• The Federal Reserve cut rates three times in 2025 but signalled caution for 2026, while the ECB delivered eight consecutive cuts and the Bank of Japan hiked to 0.75%
• Big tech companies spent an estimated $405 billion on AI infrastructure in 2025, raising concerns about market concentration and whether returns will justify the massive expenditures
• Oil prices spiked above $78 in June after Israel-Iran confrontation but collapsed to multi-year lows by year-end as OPEC+ increased production amid weakening demand