Connect with us

Business

The US factory that lays bare the contradiction in Trump’s policy

Published

on


Faisal Islam profile image
BBC A treated image of a TSMC factory, set against a red backgroundBBC

Among the cactuses in the desert of Arizona, just outside Phoenix, an extraordinary collection of buildings is emerging that will shape the future of the global economy and the world.

The hum of further construction is creating not just a factory for the world’s most advanced semiconductors. Eventually, it will mass produce the most advanced chips in the world. This work is being done in the US for the first time, with the Taiwanese company behind it pledging to spend billions more here in a move aimed at heading off the threat of tariffs on imported chips.

It is, in my view, the most important factory in the world, and it’s being built by a company you may not have heard of: TSMC, Taiwan Semiconductor Manufacturing Company. It makes 90% of the world’s advanced semiconductors. Until now they were all made on the island of Taiwan, which is 100 miles east of the Chinese mainland. The Apple chip in your iPhone, the Nvidia chips powering your ChatGPT queries, the chips in your laptop or computer network, all are made by TSMC.

Its Arizona facility “Fab 21” is closely guarded. Blank paper or personal devices are not allowed in case designs are leaked. It houses some of the most important intellectual property in the world, and the process to make these chips is one of the most complicated and intensive in global manufacturing.

A factory building, next to solar panels, in the middle of the desert.

TSMC’s Arizona factory is closely guarded

They’re hugely protective of the secrets that lie within. Important customers, such as Apple and Nvidia, trust this company to safeguard their designs for future products.

But after months of asking, TSMC let the BBC in to look at the partial transfer of what some argue is the most critical, expensive, complex and important manufacturing in the world.

The poster child for Trump’s policy

President Trump certainly seems to think so. He often mentions the factory in passing. “TSMC is the biggest there is,” he has said. “We gradually lost the chip business, and now it’s almost exclusively in Taiwan. They stole it from us.” This is one of the US president’s regular refrains.

TSMC’s recent decision to expand its investments in the US by a further $100bn (£75bn) is something Trump attributes to his threats of tariffs on Taiwan and on the global semiconductor business.

The expansion of the Arizona facility, which was announced in March is, he believes, the poster child for his economic policies – in particular the encouragement of foreign companies to relocate factories to the US to avoid hefty tariffs.

Getty Donald Trump standing in front of a podiumGetty

Trump heralded TSMC’s decision to invest in the US as proof of success for his tariff policy

China is also watching developments carefully. Taiwan’s chip-making prowess has been part of what its government has called its “Silicon Shield”, against a much-feared invasion. While the original strategy was to make Taiwan indispensable in this area of critical technology, the pandemic supply chain difficulties changed the calculus because relying on a single country seemed like a greater risk.

China claims the self-ruled Taiwan as its territory but Taiwan sees itself as distinct from the Chinese mainland.

So, many currents of the world economy, frontier technology and geopolitics flow through this one site and within it lies the essential contradiction of Trump’s economic and diplomatic policy.

He sees this plant as the exemplar of America First, and the preservation of economic and military superiority over China. Yet the manufacture of these modern miniaturised miracles at the frontier of physics and chemistry inherently relies on a combination of the very best technologies from around the world.

The cleanest environment on Earth

Greg Jackson, one of the facilities managers, takes me around in a golf buggy. The factories are almost a carbon copy of the TSMC spaces in Taiwan, where he trained. “I would say these facilities are probably some of the most advanced and complicated in the world,” he says.

“It’s quite the dichotomy. You’ve got really, really small chips with really small structures, and it takes this massive facility with all the infrastructure to be able to make them… Just the sheer complexity, the amount of systems that it takes, is staggering.”

Inside the “Gowning Building”, workers dress in protective clothing before crossing a bridge that is supposed to create the cleanest environment on Earth, in order to protect the production of these extraordinary microscopic transistors that create the microchips underpinning everything.

Konstantinos Ninios, an engineer, shows me some of the very first productions from TSMC Arizona: a silicon wafer with what is known as “4 nanometre chips”.

Watch: TSMC engineer Konstantinos Ninios shows Faisal Islam how transistors are made

“This is the most advanced wafer in the US right now,” he explains. “[It] contains about 10 to 14 trillion transistors… The whole process is 3,000 to 4,000 steps.”

If you could somehow shrink your body to the same scale and get inside the wafer, he says that the many different layers would look like very tall streets and skyscrapers.

Manufacturing manipulation of atoms

TSMC was founded at the behest of the Taiwanese government in 1987, when chip executive Morris Chang was directed to start the business. The model was to become a dedicated foundry for microchips, manufacturing other companies’ designs. It became wildly successful.

Driving the advancement of the technology is the miniaturisation of the smallest feature on chips. Their size is measured these days in billionths of a metre or nanometres. This progress has enabled mobile phones to become smartphones, and is now setting the pace for the mass deployment of artificial intelligence.

It requires incredible complexity and expense through the use of “extreme ultraviolet (UV) light”. This is used to etch the intricate building blocks of our modern existence in a process called “lithography”.

The world’s dependence on TSMC is built on highly specialised bus-sized machines, which are in turn sourced almost entirely from a Dutch company called ASML, including in Arizona.

These machines shoot UV light tens of thousands of times through drops of molten tin, which creates a plasma, and is then refracted through a series of specialised mirrors.

The almost entirely automated process for each wafer of silicon is repeated thousands of times in layers over months, before the $1m LP-sized wafer of 4nm silicon chips is formed.

“Just imagine a particle or a dust particle falling into this,” Mr Ninios says to me incredulously. “The transistors are not going to work. So all of this is cleaner than hospital operating rooms.”

Caution in Taiwan

Taiwan does not have special access to the raw materials – but it has the know-how to stay years ahead of other companies in the intricate process of producing these atomic building blocks of modern life.

Some in the Taiwanese government are cautious about spreading the frontier of this technology off the island. Trump wasted little time in claiming the firm’s decision to bring its highest level of technology to the US was due to his economic policies.

He said this would not have happened without the stick of his planned tariffs on Taiwan and semiconductors. Those I speak to at TSMC are diplomatic about that claim.

Much of this was already planned and subsidised under former US President Biden’s Chips Act.

A woman, wearing a black jacket, sits inside a room, smiling.

The supply chain for semiconductors is global as no single country can do everything at the moment, says Rose Castanares, President of TSMC Arizona

On the walkway into the building are photographs showing Biden’s visit in 2022, with the building site draped in the Stars and Stripes and a banner saying “a future Made in America”.

“The semiconductor supply chain is global,” says TSMC Arizona President Rose Castanares. “There’s really no single country at this moment that can do everything from chemicals to wafer manufacturing to packaging, and so it’s very difficult to unwind that whole thing very quickly.”

‘Non-Red’ supply chains to counter China

As for the semiconductor supply chain, tariffs will not help. The supply chain stretches all over the world. Whether it’s the silicon wafers from Japan, the machines required from the Netherlands, or mirrors from Germany, all sorts of materials from all around the world are required. Now, they could face import charges.

That said, TSMC’s boss was quick off the mark in confirming the expansion of the US site at an event with Trump at the White House. In recent weeks, America’s tech elite – from Apple’s Tim Cook, to Nvidia’s Jensen Huang – have been queueing up to tell the world that TSMC Arizona will now produce many of the chips in their US products.

The global chip industry is very sensitive to the economic cycle, but its cutting-edge technology enjoys very healthy margins, that could cushion some of these planned tariffs.

A black and red sign sits in the middle of the desert. On the sign it says 'TSMC Arizona, Corporation, Fab 21'.

The company was founded in Taiwan in 1987. In March this year, they announced an expansion of their Arizona facility

There are many geopolitical subtexts here. The factory sits at the heart of US strategy to gain technological, AI and economic supremacy over China.

Both the Biden and Trump administrations have developed policies to try to limit Chinese access to the frontier semiconductor technology – from a ban on exports to China of ASML’s machines, to new legislation to ban the use of Huawei AI chips in US software or technology anywhere in the world.

Taiwan’s President Lai this week urged democracies such as Japan and the US to develop “non-Red” supply chains to counter China.

Not everyone is convinced that this strategy is working, however. Chinese technologists have been effective at working around the bans to develop competitive indigenous technology. And Bill Gates this week said that these policies “have forced the Chinese in terms of chip manufacturing and everything to go full speed ahead”.

Trump wants TSMC Arizona to become a foundation stone for his American golden age. But the company’s story to date is perhaps the ultimate expression of the success of modern globalisation.

So for now, it’s a battle for global tech and economic supremacy, in which Taiwan’s factory technology, some of which is now being moved to the Arizona desert, is the critical asset.

BBC InDepth is the home on the website and app for the best analysis, with fresh perspectives that challenge assumptions and deep reporting on the biggest issues of the day. And we showcase thought-provoking content from across BBC Sounds and iPlayer too. You can send us your feedback on the InDepth section by clicking on the button below.



Source link

Continue Reading
Click to comment

Leave a Reply

Your email address will not be published. Required fields are marked *

Business

OBR says pension triple lock to cost three times initial estimate

Published

on


Kevin Peachey

Cost of living correspondent

Getty Images Older man and woman sit at a kitchen table with paperwork and a laptop in front of themGetty Images

The cost of the state pension triple lock is forecast to be three times’ higher by the end of the decade than its original estimate, according to the government’s official forecaster.

The triple lock, which came into force in 2011, means that the state pension rises each year in line with either inflation, wage increases or 2.5% – whichever is highest.

The Office for Budget Responsibility (OBR) said the annual cost is estimated to reach £15.5bn by 2030.

Overall, the OBR said the UK’s public finances were in a “relatively vulnerable position” owing to pressure from recent government U-turns on planned spending cuts.

The recent reversal of proposed the welfare bill, on top of restoring winter fuel payments for most claimants, have contributed to a continued rise in government debt, according to the report.

It said: “Efforts to put the UK’s public finances on a more sustainable footing have met with only limited and temporary success in recent years in the aftermath of the shocks, debt has also continued to rise and borrowing remained elevated because governments have reversed plans to consolidate the public finances.

“Planned tax rises have been reversed, and, more significantly, planned spending reductions have been abandoned.”

Spending on the state pension has steadily risen, the OBR said, because the triple lock and a growing number of people above the state pension age was contributing to costs.

It added: “Due to inflation and earnings volatility over its first two decades in operation, the triple lock has cost around three times more than initial expectations.”

Pensioner protection

The UK’s state pension is the second-largest item in the government budget after health.

In 2011, the Conservative-Liberal Democrat coalition brought in the triple lock to ensure the value of the state pension was not overtaken by the increase in the cost of living or the incomes of working people.

Since then, the non-earnings-linked element of the lock has been triggered “in eight of the 13 years to date”, the OBR pointed out.

That was because inflation “has turned out to be significantly more volatile” than expected.

In April 2025, the earnings link meant the state pension increased by 4.1%, making it worth:

  • £230.25 a week for the full, new flat-rate state pension (for those who reached state pension age after April 2016) – a rise of £472 a year
  • £176.45 a week for the full, old basic state pension (for those who reached state pension age before April 2016) – a rise of £363 a year

Chancellor Rachel Reeves has said the Labour government will keep the triple lock until the end of the current Parliament.

However, before and since that manifesto promise, there has been intense debate over the cost of the triple lock and whether it is justified.

Last week, the influential Institute for Fiscal Studies, an independent economic think-tank, suggested the triple lock be scrapped as part of a wider overhaul of pensions.

It argued that it should rise in line with prices, but the cost should be linked to a target level of economy-wide average earnings.

Pensioner groups say many older people face high living costs and need the protection of the triple lock to avoid them falling further into financial difficulty, especially because the amount actually paid was far from the most generous state pension in Europe.



Source link

Continue Reading

Business

Pluriva Invests €250K in AI Virtual Assistant for Romanian Firms

Published

on









Pluriva Invests €250K in AI Virtual Assistant for Romanian Firms – The Romania Journal





























Source link

Continue Reading

Business

AI in healthcare: What business leaders need to know

Published

on


Artificial intelligence may seem like a new, untested technology, but the reality is that AI is already integrated into our everyday lives. For instance, Siri, Amazon Alexa and Google Assistant use natural language processing and natural language understanding to analyze and respond to voice commands. Emails and text messages use NLP for predictive text and auto correct.

The rapid development of AI brings with it enormous concerns, especially regarding its applications in healthcare. However, AI is already transforming patient care in positive ways, for example, by making it easier for clinicians to diagnose and treat illness sooner, potentially reducing the need for costly specialized treatment or hospitalization.

Read more:  Sick of answering the same benefits questions from employees? Let AI do the work

Chronic condition management and early detection

While clinical judgment by an actual human is still critical to ensuring patients receive the best possible care, AI can support clinicians and their decision-making by providing a more complete view of patient health.

For instance, radiologists are now using AI to more accurately analyze X-rays, MRIs, CT scans and mammograms. AI’s sensitivity to distinguish slight changes from image to image can help detect chronic diseases earlier and more accurately. In one study, researchers found an AI system could predict diagnoses of Berger’s kidney disease more accurately than trained nephrologists. In an attempt to slow the progression of kidney disease among veterans, such as Berger’s disease, the Veterans Administration partnered with DeepMind, an AI research lab, to identify risk predictors for patient deterioration and alert clinicians early. DeepMind developed an AI model based on electronic health records from the Veterans Administration that identified 90% of all acute kidney injuries that required subsequent dialysis, with a lead time of up to 48 hours. 

Earlier intervention in the case of Berger’s disease and other kidney conditions significantly impacts the economic burden of the disease, potentially saving plan sponsors between $276.80-$480.79 per member per month. 

Read more:  AI can help benefit leaders with the compensation process

Automating administrative tasks

One of AI’s greatest assets is its ability to quickly assess large volumes of data to optimize clinical and administrative time. Medical practices are utilizing AI-enabled technology to improve administrative efficiency and patient care. Automated documentation tools can reduce the time physicians spend on patient charting by 72%, which means physicians can spend more time treating and diagnosing plan members. AI can also integrate with electronic health records to pull relevant data, identify missing information and complete and submit prior authorization forms on behalf of providers.

Administrative expenses account for 15% to 25% of total healthcare expenditures. Reducing administrative overhead and claims errors, along with early diagnosis and treatment of chronic disease, can improve member outcomes and produce impressive cost savings for plan sponsors. AI has the potential to save $265 billion in overall healthcare costs by eliminating administrative overhead and documentation errors.

AI’s ability to process vast quantities of data also benefits health plan administrators. Plan sponsors can implement AI tools that provide members with personalized treatment and support, identify health plans during enrollment that best fit specific member needs and determine additional benefits for members and their families. 

Read more:  Leaders share their most popular summer benefits

Overcoming barriers to adoption

Despite its potential to reduce healthcare costs, improve patient outcomes and improve member experience, AI adoption is still slow. The initial investment required to implement AI can be high, and it includes the cost of the technology, staff training, system integration and maintenance of AI models, not to mention potential liability concerns. 

When considering utilizing AI for the purposes of improving efficiency and outcomes, organizations in the healthcare industry are: 

  • Analyzing how AI solutions can support their population, and which modalities are likely to be (or have proven to be) successful
  • Consulting with internal stakeholders from the beginning to identify potential challenges to adoption
  • Evaluating potential cost savings and member outcomes
  • Considering the quality and source of data used to train AI models
  • Ensuring AI tools meet HIPAA requirements

AI in healthcare is no longer an idea of the future. It is here and already making significant improvements in patient outcomes. However, AI is dependent on data quality and clearly defined learning parameters to eliminate potential bias and make accurate predictions. Organizations must also weigh other risks associated with AI, such as informed consent issues that may arise if patients do not fully understand how their information is being used.



Source link

Continue Reading

Trending