In our latest Markets in a Minute, Guy Foster, Chief Strategist, discusses the recent announcement on U.S. tariffs, earnings results from some of the ‘Magnificent Seven’, and a new disruptor in artificial intelligence (AI).
The return of tariffs
In his first term, it took Donald Trump a year to impose the tariffs he had threated during his campaign. Delays were compounded by the efforts of his own administration to thwart him.
This time around, it took him less than a fortnight.
Last weekend there was the announcement that 25% tariffs would be imposed on all U.S. imports from Canada and Mexico, with the exception of oil imports from Canada, which suffer a preferential rate of just 10%. The scale of these tariffs is much more severe than during the first Trump presidency when tariffs were limited to solar panels, washing machines, steel and aluminium.
Broadening the scope of countries affected obviously increases impact, but it also compounds it by reducing the workarounds where goods can be funnelled through different countries.
The announcement of these tariffs could have been seen as crystallising one of the major uncertainties which the market had been facing since it became likely that President Trump would return to the White House. But I avoided writing that it brings certainty. The tariff threat could very easily be dialled up or down. Trump was already warning that the European Union would be confronted, while hinting that the UK could possibly avoid it. Meanwhile, he’d already threatened Colombia with rising tariffs before dialling them back in less than an hour.
The tariffs against Canada and Mexico were delayed a month following calls with their respective leaders and apparent deals over enhancing border security. Companies had been working to get stock across the borders ahead of President Trump taking power. They’ll redouble those efforts now.
What happens when that month elapses? Here are a few potential scenarios:
• Canada and Mexico face a material risk of recession when, or if, these tariffs are imposed, while the U.S. may also suffer weaker growth.
• The average weighted tariff rate on U.S. imports will rise threefold on the levels seen in 2018.
• The impact on growth depends upon whether the money taken from U.S. consumers and businesses via tariffs finds its way back to them or just reduces the deficit.
• Inflation in the U.S. will rise, with some expecting it to reach 3%. Although the impact of tariffs should be a one-off adjustment which drops out of the Consumer Price Index (CPI) rate after a year, unless it impacts inflationary expectations.
A big week for earnings
All of this came after a pretty hectic week of market movements. Last week was a big week for earnings, with Apple, Tesla, Meta and Microsoft from the ‘Magnificent Seven’ all reporting, but the volatility was driven by events taking place on the other side of the world.
In terms of those earnings, Apple’s were pretty good. Although its performance in China was weak due to increased competition, the company remains positive on broader emerging market sales. It also emphasised that iPhone sales have been stronger in regions that offer Apple Intelligence, the firm’s AI system, which has attracted some scepticism from the investment community.
Tesla’s results were behind estimates and vehicle sales guidance for 2025 was lowered. However, the shares shrugged this off due to strong performance in battery
sales, plans for self-driving cars, and, particularly, a focus on Optimus humanoid robots, which the company is developing.
A new player in the AI space
These companies were reporting after last weekend’s revelation that Chinese hedge fund High-Flyer’s AI lab, DeepSeek, has created an AI model that performs on
par with industry leaders.
What’s notable about this company is its ability to achieve impressive results with limited resources. While other companies have spent hundreds of millions of dollars on developing similar technologies, this firm has managed to do so at a cost of just a few million dollars. Its model also requires significantly less computing power to operate, making it a more efficient and cost-effective solution.
The efficiency comes from building a ‘mixture-of-experts’ approach. This means that instead of trying to produce one enormous model that can do the widest range of tasks, DeepSeek’s model is essentially a series of smaller specialist models that are optimised for different tasks. A gatekeeper breaks down a task into components and sends them to appropriate ‘experts’ – comparable to the division of labour into specialised roles.
One of the key factors contributing to the company’s success is its approach to innovation. By making its AI algorithms and models open source, the firm
is allowing developers and researchers to access and build upon its technology. This route, which is also followed by Meta, enables models to build a community of users. Being able to see a model’s source doesn’t make it copyable because development is still expensive.
The release of DeepSeek saw a sharp drop in many technology-related shares. A lot of the value lost was claimed back in the following days.
The most obvious company to focus on would be Nvidia. Nvidia sells the latest cutting-edge chips required to build bigger and more powerful models. As DeepSeek was able to produce its model with older weaker chips, investors questioned Nvidia’s prospective demand.
Other companies like Google and Meta were also perceived to be at risk because the DeepSeek model would undercut processing costs in their own models.
In a market that’s ridden high on the hopes for AI, is the edifice about to come crashing down?
What does this mean for AI?
We’ve found the analogy between the AI boom and the gold rush to be an instructive one. In the 19th century, the promise of gold encouraged people to migrate and work to try and uncover gold. Few succeeded, but fortunes were made selling them the equipment (picks and shovels) to do so. If we view the promise of AI to be the profits that can be generated by an AI model, then that business case has indeed been dented. However, if you see the promise of AI to be the gains made from the use of AI, then the falling cost is a good thing.
The principle of the Jevons paradox applies here. Let me explain…
The Watt steam engine burnt coal more efficiently, creating fears that coal demand would fall. Instead, steam travel increased. The same has been noted in
countless technological developments ever since. Therefore, more competition in the AI world should accelerate the pace of adoption. The picks and shovels of AI are the semiconductors, the data centres and maybe even the power sources.
So, at the end of this tumultuous week for the technology market, the biggest questions surround the route to monetisation for the most cutting-edge AI models, while the route to greater adoption of AI has become a little smoother.
What’s ahead?
This week, there will be more earnings from ‘Magnificent Seven’ members Alphabet and Amazon.
Purchasing manager indices will suggest the state of different economies one month into the year. Provisional estimates revealed last week suggest the manufacturing sector may be recovering, although attempts to frontload purchases before U.S. tariffs are imposed may be distorting things.
On Thursday, the Bank of England is widely expected to cut interest rates from 4.75% to 4.5%, following in the footsteps of the European Central Bank, which cut
rates last Thursday.
The week will then finish with the U.S. employment report, where 150,000 new jobs are expected to have been produced during January.
But these are just the scheduled and anticipated announcements. Recent weeks have been dominated by announcements from businesses and policymakers which had not been anticipated. With the tariffs issue now moving towards the top of the agenda for businesses and maybe consumers, the President’s language will be scrutinised for any hint of how this saga is going to develop.
