The White House claims that Trump’s “America First” policies are bringing in trillions of dollars in new investments into the US.
They call this the “Trump Effect”.
Numbers of up to $21 trillion are being thrown around. But behind these numbers sit very specific projects, diplomatic trades, and accounting tricks.
Some are real factories and data centres that will change local economies. Others are trade targets and arms deals dressed up as investments.
To understand whether there is a real “Trump effect” and what these AI investments mean, it helps to take the list apart and then put it back together in a different way.
A twenty one trillion dollar story
The US President Donald Trump tells a compelling story. Since he returned to office, his “America First” approach has attracted $21 trillion of investment pledges into the United States by the end of the year.
If all of it were real capital spending, even spread across several years, it would be the largest surge in modern US history.
The White House backs the narrative with its “Trump Effect” web page. That page lists about $9.6 trillion in “total US and foreign investments,” covering manufacturing, energy, pharmaceuticals, and above all, technology and AI.
A recent analysis by Bloomberg Economics went through that list project by project.
Their analysis found that only about $7 trillion looked like genuine investment commitments. Roughly $2.6 trillion on the page relates to things such as long-term purchase agreements, vague plans to expand trade or “economic exchange,” and other items that do not involve new plants, equipment, or infrastructure inside the United States.
Source: Bloomberg
And even that $7 trillion figure is still very large. If delivered on the timelines described by companies and officials, it would work out to roughly $1.5 trillion a year. That is about 5% of US GDP.
That would put the “Trump Effect” among the biggest injections of capital in modern US history.
Source: Bloomberg
The AI megaprojects that really matter
Inside the AI bucket sit a handful of projects that carry most of the weight, involving land, steel, turbines, and thousands of workers.
The most talked about is Project Stargate, a joint venture between OpenAI, SoftBank, Oracle, and an investment vehicle called MGX.
It was launched at the White House early in 2025. The stated plan is to invest up to $500 billion into AI infrastructure in the United States by 2029.
Stargate has an initial capital pool of around $100 billion and intends to use significant debt. It has anchor sites in Texas and other states and ties in deals with banks and equipment suppliers.
It has also already faced delays and questions over funding. That is not unusual for very large projects, but it does underline that the upper end of the promised spending is not automatic.
Around Stargate sit the big corporate programs that the White House highlights.
Nvidia talks about producing up to $500 billion worth of AI servers and supercomputers in the United States in the next few years.
Apple has lifted its total US commitment to $600 billion. Meta talks about a similar figure. Amazon and Google are building and expanding their own data center campuses.
TSMC, Micro,n and other chip firms are building or enlarging fabs and packaging plants.
Source: Understanding AI
In the short term, these AI investments show up in construction and equipment orders.
They also strain local power grids. Several regions already report data centres waiting for extra grid capacity before they can fully switch on.
In capital markets, these projects concentrate risk.
A large share of the entire Trump era investment story rests on a few firms and one technology cycle, sparking worries about an “AI bubble”.
If AI spending slows, the overall boom would look very different.
Genesis and the state bet on AI science
One risk of any tech boom is that the gains stay inside a small circle of firms. The Trump administration points to the Genesis Mission as an answer.
Genesis is an executive order and program that asks the federal government, led by the Department of Energy, to turn decades of publicly funded science into a platform that can use advanced AI.
The mission is to link national lab supercomputers, federal datasets, and new AI models into what the White House calls an American Science and Security Platform.
In practice, that means cataloguing and opening up compute capacity, training domain-specific models for areas like materials, biology, and energy, and then letting researchers and agencies run faster experiments and simulations.
It is an attempt to treat AI infrastructure as something more than a private tool for a few consumer internet and cloud giants.
The dollars attached to Genesis are modest compared with Stargate or the headline corporate pledges, as it leans on existing budgets and facilities.
Its importance lies in how it might affect the productivity of US science if it is funded and managed well.
It will also test how comfortable Americans are with private cloud providers sitting at the centre of sensitive research and security workloads.
If Genesis works, it increases the odds that the AI investment wave leaves a lasting mark in fields like drug discovery, clean energy, and defence technology instead of stopping at language models and recommendation systems.
Is there really a Trump effect?
Once the numbers and projects are unpacked, the phrase “Trump effect” takes on a more precise meaning.
While it does not describe a $21 trillion wall of new capital, it describes the way this administration has chosen to ride and shape an existing explosion in global investment.
Ultimately, the AI boom did not start in January 2025.
The chip and green industry programs that supported it were designed and passed earlier in the decade.
The United States was already an attractive destination for investors, with deep capital markets and relative political stability compared with many peers.
But Trump did add a particular style. Tariffs, export controls, and access to US technology and markets are used as bargaining chips.
Foreign governments and companies are encouraged to arrive in Washington with very large public pledges.
A mix of tax and regulatory decisions then encourages firms to place AI infrastructure, some manufacturing, and some energy projects on US soil rather than elsewhere.
The administration also counts earlier plans and existing spending toward its own tally.
There is indeed a real effect there. Some factories, data centres, and grid upgrades are in places and at scales that reflect these negotiations.
The sovereign funds and state-owned firms that want US chip access or tariff relief have paid a visible price in the form of investment promises.
The $7 trillion figure is closer to reality, however.
It is also heavily concentrated in data centres, chips, and the energy systems needed to power them. It includes a smaller but meaningful set of manufacturing, pharmaceutical, and infrastructure expansions, but excludes the bulk of the sovereign promises that rely on trade growth rather than capital spending.
If the power lines get built, the data halls fill up, the Genesis platform makes the labs more productive, and enough of these projects survive the next downturn, the impact on US capacity and productivity will be real.
The post America’s trillion dollar promise: Is the “Trump Effect” real? appeared first on Invezz
