Hello and welcome to World Business Report from the BBC World Service.
I'm Roger Hearing and on this edition, interest rates on hold in the US.
The Fed decides to leave them where they are in its first decision this year and come to heavy criticism from President Trump.
Also, as US tech is still reeling from the new cut price Chinese AI app, we'll hear from the boss of the world's largest semiconductor manufacturing machine maker and from the CEO of the company making the largest AI chip and how the Ghanaian economy deals with the price of a nut.
You've got cashew nuts here.
People like cashew.
People like cashew.
That's all coming up.
But first, it was one of the most heavily telegraphed central bank decisions of recent times.
The US Federal Reserve has held interest rates where they are.
The first Fed decision of the Trump era came without much surprise after three rate cuts last year.
And this time, they're no longer saying inflation reduction has made progress, only that the pace of price increases remains elevated.
Here's the Fed chair, Jerome Powell, explaining the decision.
The economy is strong overall and has made significant progress toward our goals over the past two years.
Labour market conditions have cooled from their formerly overheated state and remain solid.
Inflation has moved much closer to our 2 % longer run goal, though it remains somewhat elevated.
In support of our goals, today the Federal Open Market Committee decided to leave our policy interest rate unchanged and to continue to reduce our securities holdings.
Jerome Powell there.
Let's talk to Susan Schmidt, who's Portfolio Manager at Exchange Capital Resources.
This is Susan. Thanks for being with us.
What are the markets reading into this decision?
Well, Cameron Powell's comments were very much in line with what investors expected.
So there is a lot of uncertainty in the policies, tariffs, and therefore price changes that might happen in the cost of goods for the U .S.
consumer. And they expected the Fed to hold pat with current rates until we get more clarity on that.
Remember that inflation has been plaguing many companies, many economies across the globe, it hasn't gone away, and Chairman Powell is indicating that.
It's still something they're keeping watch on.
And of course, he's coming under a lot of flack from President Trump, who says that the Fed's policies on things like it's reading into this decision.
Well, Chairman Powell's comments were very much in line with what investors expected.
So there is a lot of uncertainty in the policies, tariffs, and therefore price changes that might happen in the cost of goods for the US consumer.
And they expected the Fed to hold pat with current rates until we get more clarity on that.
Remember that inflation has been plaguing many companies, many economies across the globe.
It hasn't gone away.
And Chairman Powell is indicating that.
It's still something they're keeping watch on.
And of course, he's coming under a lot of flack from President Trump, who says that the Fed's policies on things like diversity were part of a problem and that they wouldn't have messed up with inflation if that hadn't happened.
I mean, he's feeling under a bit of fire at the moment, I imagine.
He absolutely is. And this is a little bit of the Fed trying to say we're independent and we act independently.
Remember that President Trump in Davos made some very strong statements that he wanted to lower interest rates right away.
Well, that's not really in the presidential capacity.
And so Chairman Powell, I think, is taking a little bit of a stand here.
But following the course that investors were expecting him to take regardless.
And at the same time, I mean, just looking further forward, if we see pressures coming in because of tariffs and other things, I mean, is it possible they could go back to interest rate rises?
The Fed always wants to keep all their options open and to keep the market guessing.
So certainly that's possible.
However, I think investors are expecting the Fed to hold interest rates where they are should they see inflation back to prior levels or start to creep up and not raise interest rates at this time.
Suzanne, stay with us if you will.
We'll talk about some more issues on Wall Street in a little while.
But first, let's turn our attention to what happened on the share markets this week because they were seeing quite a lot of upheaval sparked by the arrival of China's new AI app DeepSeek, with tech stocks taking a big hit, in fact, on fears the US was losing the AI race.
On Wednesday, things had calmed, but NVIDIA and Microsoft were still down.
So it was an interesting moment for ASML, which is the world's largest semiconductor manufacturing machine maker, to release its latest earnings.
Now, they revealed that 2024 was a record -breaking year, and its outlook for 2025 remains equally promising.
My colleague Sam Fennig spoke to the ASML president and chief executive officer, Christophe Fouquet.
So how have the last 12 months been for his business?
We have seen a major shift of AI, which created, of course, a huge opportunity for the future and some success in providing the right technology for those applications.
And those are the customers that are doing very well.
Some of the other players are playing a bit catch up.
And for them, of course, I would say we are looking maybe at a delayed success on AI.
So who are those that are having to catch up?
I think if you go back to some of the press last year, I think some of the names have been mentioned many, many times, so I don't need to do that.
It's also true for the winners.
I think they are well known.
This week has been a busy week for AI, and we've seen that a Chinese artificial intelligence startup, which seems to be able to show that AI systems can be made more efficiently.
Does their approach question the need for powerful AI hardware, like the stuff that you're making?
No, it does not. You know, AI creates a huge opportunity in this market.
So I think that the fact that a new company brings a novel approach to try to beat whoever maybe seems to be the leader today is very, very normal.
And we believe also that for AI to be really accessible to a very large public, the cost of AI has to go down.
So I think innovation that will drive cars down, either for the model or later on with the hardware, later on with the process, I think all of that is very necessary.
Can we talk about export controls now?
These are rules which prevent you selling some of your kit to China.
There are restrictions from the US government as well as the Dutch government.
How difficult do they make your business in navigating businesses in China and the US?
The difficult part for us usually on that is the lack of stability.
You don't know exactly what you can, cannot do tomorrow.
I think that's been the biggest challenge.
Once the rule is set, you know, we are a company that acts within the law and we are going to adjust to that.
I think stability is the key word here.
We hope, of course, that at some point of time we reach that point.
Do you agree that there should be restrictions on what you can sell and to whom?
Well, I think it's, you know, it's not, I would say, a question for ASML when it comes to matter of national security.
We are good citizens in this case, and we have to trust our political leader to make the right call for their citizens.
I think as a citizen, we also understand that in some cases, this is a tool that may be used.
Some companies have tried to circumnavigate those rules by sort of making tweaks to their technology so it doesn't come within a particular...
It's the key word here.
We hope, of course, that at some point of time we reach that point.
Do you agree that there should be restrictions on what you can sell and to whom?
Well, I think it's, you know, it's not, I would say, a question for ASML when it comes to a matter of national security.
We are good citizens in this case, and we have to trust our political leader to make the right call for their citizens.
I think as a citizen, we also understand that in some cases, this is a tool that may be used.
Some companies have tried to circumnavigate those rules by sort of making tweaks to their technology so it doesn't come within a particular regulation.
Is that something that you've looked to do?
No, I think the biggest decision when it came to export control was to prevent the export of EUV to China.
And, well, we have not shipped any tool there.
So I think it's not something we'll look at.
I think you should have a discussion before if you believe that the decision is not the best.
I don't think that trying to go around decision is right in any way.
Your net sales are up.
They are flying. You said it's been a very good year, 2024, but the share price over the last year has been down 20 % and then down again 12 % this week, which perhaps suggests that investors are concerned about the future growth of the company.
Do you think they're right to be concerned?
When we look at the opportunity for SML, we always look at the long term.
We had in November what we call a capital market day where we share that view with our investor, but also with a very large public.
And we are focusing on the long term because we know by experience that this industry is very cyclical.
We also know that when you have a new technology like AI, you know, it's not going to just be a walk in the park, right?
You're going to have challenges.
You're going to have changes.
So this is in the long term that you win or you lose.
I think a lot of what you describe is more related to reaction on short term news than anything else.
Christoph Fouquet there.
Well, what was making the short -term news, of course, the past week, as I said, was the furore over the advent of Deep Seek, the artificial intelligence app brought in by a Chinese company.
And the row about that has now reached the U .S.
Congress. Howard Lutnick, President Donald Trump's nominee to lead the Commerce Department at his U .S.
Senate confirmation hearing for his nomination, accused China of stealing OpenAI's IP.
How could it be more clear than this week when DeepSeek, a China...
I mean, let's just talk, first of all, about the upheavals we've seen this week, because it is interesting.
Why do you think suddenly this AI app from China has upset everyone?
Well, first, thank you for having me.
I think DeepSeek showed several things that changed the playing field.
They showed that a small group of engineers, relatively, about 200 engineers, using less compute than was being used by Google or OpenAI could do very important work and could move the industry forward.
I think we have underestimated what a small number of exceptionally focused and talented engineers can do.
But what does that say, if I may interrupt you, about your company?
Because you're a company that makes a very, very big chip, as I just said, and a lot of money goes in, is $4 billion is the value of your company.
And yet something that's much, much smaller, it seems, can set everything on fire.
Does that suggest that perhaps you are moving in the wrong direction?
Well, I don't think smaller in this case is little.
They used an enormous amount of compute.
They used hundreds of millions of dollars' worth of compute to do their work.
And the way we use the model, called inference, how we use AI, continues to need an enormous amount of compute.
So quite the contrary.
DeepSeq's announcement is great for Cerebris.
It extends the AI universe.
It makes AI available to more people.
And we are, they're using this AI on our hardware.
And so I think there has been an ongoing mistake made particularly in the public markets over the last 50 -odd years when people think that innovations that reduce the cost and increase the performance of compute make the market smaller.
And in every single instance over half a century, they've made the market bigger.
So you think that DeepSeek will actually help a company like yours?
We know it will. We are awash in demand for DeepSeek.
Correct. And as a company like yours, do you feel that what's happening in China, that they can produce something like this and they haven't had great access to a lot of microchips, the latest ones for a little while because of sanctions of various kinds and tariffs and barriers.
Does that prove that they actually are innovating faster than companies like yours in the US at the moment?
Well, we don't compete with them.
We compete with NVIDIA, and we compete with other chip makers.
They are software makers, and I think what it showed, and I think it's a reminder, and we believe very much in this, is that this is a market in which underdogs can have tremendous impact.
We compete against a company worth several trillion dollars.
They're competing against companies worth hundreds of billions of dollars.
And in both cases, small teams have been able to innovate, do things that people bet against, said couldn't be done.
I think that's exciting, and I think that's very positive for the AI ecosystem.
Are you able to export into China?
We are not. We have chosen long before the formal rules came into place not to export to China.
Why? We thought it was the right thing to do.
Because of the way that they operate, because of the political atmosphere in the U .S.?
What was the reason?
Because we build the fastest AI processors in the world, and we thought that they could be used for the wrong purposes.
And we chose, therefore, to build a business plan in line with our ethics.
And now you see Chinese innovation, which we've talked about, coming from there.
That doesn't change your mind at all?
They chose to make innovations that were built on the innovations of U .S.
companies, and they chose to open source them.
So much of the DeepSeek models are standing on the shoulders of work done at Meta, work done at Google, work done at other startups.
DeepSeek, in turn, open -sourced their model.
They made it freely available.
They used an extremely permissive license, the MIT license.
And I think there are customers all over the world that want to use it.
They may not want to use the DeepSeek app in which the data goes to China, But you can run the model on us today.
It'll be the fastest in the world.
All the work is done on computers we built that are housed in the US.
We store none of your data and never will.
And so I think it is part of the open source ecosystem that there is a community that's building, that's moving the industry forward.
I think that's one of the very interesting.
And I think there are customers all over the world that want to use it.
They may not want to use the DeepSeek app in which the data goes to China, but you can run the model on us today.
It'll be the fastest in the world.
All the work is done on computers we built that are housed in the US.
We store none of your data and never will.
And so I think it is part of the open source ecosystem that there is a community that's building, that's moving the industry forward.
I think that's one of the very interesting things that happened here.
They're competing against a closed source ecosystem.
They're competing against Google's Gemini and OpenAI's GPT models.
And DeepSeek put their model out for free.
And it's been a very disruptive affair as well.
Well, Andrew, thank you so much for talking to me.
Andrew Feldman there, founder and CEO of Cerebras Systems.
You're with World Business Report from the BBC World Service.
Now, Facebook's parent company, Meta, has beaten Wall Street estimates, posting a record revenue of $48 .39 billion for the fourth quarter, and that compared with analysts' average estimate of $47 .03 billion.
The reason seems to be, at least in part, It's artificial intelligence -powered marketing tools which continue to attract people spending for advertising across the social media platforms.
Let's talk now to Venki Hario -Ryan, who is an entrepreneur, investor, and partner at the venture capital firm Rocket Ship.
Now, Venki, you were an early investor in Facebook, I think back in 2005.
Is that right? Yep, that's correct.
So what do you make of this?
I mean, is Meta really, we've heard of the ups and downs of tech this week, but Meta really seemed to be out there and doing pretty well.
Yeah, actually, it's sort of very interesting, the parallel between DeepSeq and Meta.
If you recollect, a few years back, Meta went through a really difficult phase because Apple cut off access to some key data for ad targeting for Meta.
and given that Meta had to innovate on AI to really survive and thrive and that challenge of not having access to information that they had really I think has propelled them to be sort of to be maybe the largest user and deployer of AI at scale today in the world and so that is what you're seeing at some
level in these earning results.
And do investors like you worry about what I was just talking about there?
As proprietary systems is a game changer and I think DeepSeek has shown that.
The second big sort of thing that DeepSeek has shown is that you can actually inference at a very, very low cost.
That again means you can apply AI more broadly.
So overall I believe that it's actually very, very positive for AI going forward.
The big question is all the sunk investments in AI so far, which of those are impacted and which are not?
I think that is going to take some sorting through.
Venki, I mean, you'll remember as I do the moment where Facebook meta seemed to be betting their house on VR, virtual reality.
That kind of went away a bit.
Now they're doing the same with AI.
Is this a company that really knows where it's going?
Yeah, I think part of being in tech, part of being in Silicon Valley is you have to have a thick skin and be willing to be wrong because the world changes very quickly, and if you are stuck in a direction, you're usually going to be going in the wrong direction.
So from that point of view, if you look at Amazon, you look at Meta, I'm very appreciative of how quickly they're able to change, how quickly they're able to recognize where the world is going and make the right investments.
And yes, you are going to make a few mistakes along the way, but so long as you get one right out of, like, three or four, I think you're doing very well.
Venki, thank you so much for your insights there.
Venki Harinarayan there of Rocketship.
Now, let me bring Susan Schmidt back in, Portfolio Manager at Exchange Capital Resources.
Susan, I want to bring up another couple of tech results, because we've seen a lot of tech earnings now.
Tesla, their gross profit margin missed its estimates, 16 .3 % in the fourth quarter and their shares in after hours trading took a bit of a hit, didn't they?
A bit, but down just a few percent and moving around quite a bit.
It'll remain to be seen what happens tomorrow.
Tesla had an interesting report.
They sold 1 .8 million cars in 2024.
That sounds impressive, but that's the first drop in over 10 years where there was a lower number than the prior year.
Investors are looking at that and worried that despite all the sales, financing, et cetera, initiatives that they tried to bring more customers in with, they still didn't manage to beat last year's sales numbers.
And let's also talk about Microsoft.
It's been everyone in Silicon Valley, I think, reporting today.
Slow and expected growth in their Azure cloud business is what they report, but they did beat estimates for overall quarterly revenue.
Again, people looking fondly at Microsoft or not?
Overall, yes. I think the cloud business was a disappointment to investors.
as you're seeing that in this lackluster stock price performance years where that was a lower number than the prior year.
Investors are looking at that and worried that despite all the sales, financing, et cetera, initiatives that they tried to bring more customers in with, they still didn't manage to beat last year's sales numbers.
And let's also talk about Microsoft.
There's been everyone in Silicon Valley, I think, reporting today.
Slow and expected growth in their Azure cloud business is what they report, but they did beat estimates for overall quarterly revenue.
Again, people looking fondly at Microsoft or not?
Overall, yes. I think the cloud business was a disappointment to investors.
You're seeing that in this lackluster stock price performance.
But where Microsoft really delivered was wonderful growth in the AI side of their business.
And so investors will weigh that out given the DeepSeek issues going forward.
Indeed, which brings us full circle pretty much, doesn't it?
Thanks so much for being with us, Susan.
Susan Schmidt there.
are to exchange capital resources.
Now, as countries in Africa attempt to pull themselves out of something of an economic slump since the pandemic, many are going to be looking to their start -ups and small businesses to generate better growth.
But it's a challenge.
Many largely agricultural African economies still depend heavily on imported food for their staple diet.
My colleague Ed Butler has been looking into this in Ghana and taking the example of one heavily farmed local product, the cashew nut.
so we stopped the side of the road now this is just one of the busiest thoroughfares of central accords a lady here with a whole bags of nuts can i see you've got cashew nuts here people like cashew people like cashew i'm buying my nuts and they're costing me about 75 u .s cents for to be honest a pretty
tiny 30 gram or one ounce bag.
Now these nuts have been shelled, roasted and packaged, but I've done the mass and this tiny bag I'm crunching on now, it comes at a pretty astonishing mark up.
40 times the farm gate price.
Yep, that's 40 times more than you pay for these raw nuts from a Ghanaian farmer.
So how does that happen?
My theory of why Africa's economic transformation has been slow, We focus a lot on the supply side.
How do we get the soil to yield more cashews?
But a lot of the beauty is on the demand side.
Bright Siemens is a social commentator, innovator and entrepreneur himself in Accra.
And he's given a lot of thought to this cashew question.
What he calls a basic market failure that's symptomatic of comes at a pretty astonishing markup.
40 times the farm gate price.
yep that's 40 times more than you pay for these raw nuts from a Ghanaian farmer so how does that happen?
My theory of why Africa's economic transformation has been slow is we focus a lot on the supply side oh how do we get the soil to yield more cashews but a lot of the the beauty it's on the demand side.
Bright Siemens is a social commentator innovator and entrepreneur himself in Accra, and he's given a lot of thought to this cashew question, what he calls a basic market failure that's symptomatic of many others in Ghana.
So basically, Ghanaians haven't learned to love cashews yet.
Very good. You need to turn Ghanaians into cashew lovers.
And that's a different set of capacities.
The best entrepreneurs are not performing.
It's a massive brain drain.
Many of them have left.
And you don't have an entrepreneurial class that can create new markets.
But it's not all bad news.
Some Ghanaian entrepreneurs are also cracking open the local market.
Hello. Hi. Good to have you.
Yes, it is. This is Mildred Acoutia, founder and CEO of Acouaba, a cashew processing firm based on the outskirts of the capital.
It's an automated roasting machine.
We've had to acquire this to increase our efficiency and scale us up and all that.
In a residential house that she's adapted into a factory, she and a small workforce roast some 25 tonnes of cashews a year.
The objective of my business is actually to give a facelift to Ghanaian processed foods.
But, she says, her problems come from massively overpriced credit from the banks, 30 % interest or more sometimes, and at a market that is geared towards supplying foreign multinationals.
Currently, we can't even meet demand.
We can't get enough canals to roast because we have a lot of foreigners from India and Vietnam coming in with funding with lower interest rates and they are able to carry a lot from Ghana away.
And that competition from better -funded and more powerful exporters is what's driving many Ghanaian processors out of business, according to Darren Hachamoglu.
He's a Nobel Prize -winning economist and author of Why Nations Fail.
these entrepreneurs really
