America
Trump’s class alliances: Which companies are profiting from ICE operations?
Operations conducted by Immigration and Customs Enforcement (ICE) units across the United States offer significant clues regarding the “class alliances” underpinning Donald Trump.
Institutions including companies like Palantir and Deloitte have reaped more than $22 billion in total earnings from contracts with agencies situated at the center of the aggressive immigration measures Donald Trump implemented over the past year.
As reported by the Financial Times (FT), consultants, technology groups, charter airlines, and a wall construction company managed by a presidential ally were among the primary beneficiaries of the surge in spending by ICE and Customs and Border Protection (CBP).
This funding bonanza began following Trump’s inauguration for a second term last January and has accelerated since the enactment of the “big, beautiful bill” in July.
According to the FT’s analysis of government contracting data, the data intelligence group Palantir has secured $81 million in contracts from ICE since January 2025.
Consultancy firm Deloitte, meanwhile, obtained more than $100 million in new contracts from ICE and CBP during the same period.
Regional companies secure major contracts
The Fisher Sand & Gravel group, led by Republican donor Tommy Fisher—which signed a contract to construct sections of a wall on the southern US border—became the top earner from CBP contracts, generating over $6 billion in revenue since July.
The single largest beneficiary of ICE contracts was CSI Aviation, a company organizing charter flights for the agency. This firm has secured over $1.2 billion in business since Trump returned to office last January.
These windfalls coincide with a period in which ICE spending on contracts more than doubled in the two quarters following the passage of Trump’s historic legislation, rising from $1.5 billion in the previous six months to $3.7 billion.
CBP spending on private sector companies increased sevenfold between the first and second halves of 2025. The agency reported $2 billion in new contract work this month alone—a sum exceeding the total for the entire first half of 2025.
Much of the agencies’ contracting is for routine work, such as modernizing IT systems or providing outsourced data center staffing, often stemming from previous administrations.
Palantir building a “self-deportation tracking” system
However, other contracts relate to new tactics employed by the Trump administration to identify, detain, and deport undocumented immigrants, or to encourage them to “self-deport.”
Palantir, which has held contracts with the agency for over a decade, signed a $30 million deal in April to build an operating system to be used for “self-deportation tracking,” according to a federal contract announcement.
The company also signed a contract to provide tools intended to “facilitate operations for the selection and apprehension of illegal aliens.”
Palantir CEO Alex Karp had previously dismissed concerns regarding the group’s work for the US government.
Karp stated last year:
“I will use all my influence to ensure this country remains skeptical on immigration and possesses a deterrent capacity. Do we have to pretend that having borders is immoral?”
AI-based language models in the service of immigration enforcement
According to the 2025 DHS AI Use Case Inventory published by the US Department of Homeland Security (DHS), ICE has been using Palantir’s AI products to process large volumes of civilian reports since May of last year.
This tool, named the “AI-Enhanced ICE Report Processor,” utilizes large language models (LLMs) to summarize or categorize received reports and offers functionality to translate reports received in non-English languages into English.
The “Advanced Lead Identification and Enforcement Target Selection” tool, which ICE has utilized since June of last year, was also purchased from Palantir.
This tool, known by the acronym “ELITE,” uses artificial intelligence to identify leads—such as the addresses of enforcement targets, including for deportation—and allows agents to share this information.
It has also been revealed that ICE uses Palantir-based generative AI for internal developers’ code writing and system administration.
Anduril’s surveillance towers in the “Big, Beautiful Bill”
Trump’s legislation also mandates that all new border surveillance towers be certified as “autonomous.” According to a report published in The Intercept last July, only Anduril’s “towers” meet this requirement.
Signed into law by President Trump on July 4th, this bill provides significant spending increases for military and law enforcement projects, including over $6 billion for various border security technologies.
These initiatives include the expansion of the “virtual wall”—a growing network of sensor-equipped surveillance towers along the US-Mexico border. On this border, computers are increasingly assuming the task of detecting and apprehending migrants.
Anduril began its operations by selling software-backed surveillance towers to CBP. The company promotes its “Sentry Tower” series for its “autonomous” capabilities, which use machine learning software to constantly scan the horizon and detect potential objects of interest (such as people, vehicles, or animals attempting to cross the border) without the need for human manpower to monitor sensor data.
Thanks to bipartisan support for the vision of locking down the border with computerized eyes, Anduril has become a dominant player in border surveillance, surpassing incumbents like Elbit and General Dynamics.
Indirect support from Big Tech
Deloitte, one of the largest public sector contractors in the US, accepted recent contract updates providing further funding for “law enforcement systems and analytics for enforcement and removal operations.”
Its contracts also contain updated provisions for “internet research and data analysis support services” for ICE’s target identification operations division.
Many major technology companies do not contract directly with the federal government, but their products and services are offered through vendors, making it difficult to determine the financial benefits they derive from the funding surge.
Amazon and Microsoft, the world’s two largest cloud groups, provide services worth at least $75 million and $93 million, respectively, to US agencies.
These services are primarily provided through third-party vendors such as Dell Federal Systems.
In September, ICE awarded a $24 million contract to a third party to provide “hosting support” for services offered by Amazon’s cloud division.
Additionally, it paid Dell $19 million for Microsoft enterprise licenses.
Smaller tech groups, such as Motorola Solutions, have also signed contracts with ICE.
The Illinois-based group holds $19 million in contracts in its own name, while a third-party vendor won a $260 million contract to provide Motorola radios and batteries to personnel involved in enforcement actions.
Furthermore, AI technologies from various major tech companies are being utilized. ICE used a GPT-4 based AI tool from OpenAI to review resumes for recruitment.
AI technologies from Meta, Google, OpenAI, and Anthropic are also in use.
Land and warehouses for new prisons
Despite protests in small towns and cities across the US, the Trump administration continues to purchase warehouses it plans to convert into immigrant prisons as part of a project that could represent the largest expansion of detention capacity in US history.
According to Bloomberg, the cost of purchasing just two warehouses was $172 million. A third warehouse in El Paso, Texas, could become one of the largest prisons in the country with a capacity of 8,500 beds once completed as planned.
These deals mark the latest development in ICE’s plan to utilize 23 warehouses to detain thousands of immigrants arrested by federal agents in Minneapolis and other cities.
On January 16, according to a local court filing, the administration paid $102 million for a plot of land near Hagerstown, Maryland. A week later, the government paid $70 million in cash for a warehouse in Surprise, Arizona.
The prices, which are roughly in line with the industry average for the warehouse market, cover the purchase of these currently vacant spaces.
ICE must pay companies to equip the buildings with toilets, showers, beds, dining, and recreation areas, and subsequently to operate them as detention centers.
The warehouses, most of which were originally designed and marketed as e-commerce distribution facilities, are crucial to the administration’s $45 billion construction of immigrant detention facilities.
In recent weeks, the federal government has toured potential sites in more than 20 cities with companies and shared designs with them, including preferred layouts for at least 15 locations.
According to sources speaking to Bloomberg, companies that will convert these warehouses into prisons have been asked to submit bids for the initial locations, starting with Hagerstown.
The pattern is evident here as well: For instance, in Salt Lake City, the warehouse designated by ICE as a future “mega-center” prison is owned by the Ritchie Group, a local family business.
Following pressure from protesters arriving at their offices, the company announced it had “no plans to sell or lease the property in question to the federal government.”
Meanwhile, KPB Services, the company that won the tender for the design to convert warehouses in Kansas into detention centers, appears to be a shell company.
ICE has increased detention capacity by leveraging long-standing relationships with private prison companies such as CoreCivic and Geo Group. These companies have provided ICE access to additional beds in their existing prisons, purchased and leased new facilities, and reopened shuttered ones.
In earnings calls held in November, these companies stated they could make a total of more than 30,000 beds available should the federal government request them.
British firms among the beneficiaries
Subsidiaries of several prominent UK-based companies have also secured active contracts with these agencies. British private security firm G4S has signed contracts worth $68 million with ICE since January 2025.
These contracts primarily cover providing “ground transportation services” for detainees during enforcement and removal operations.
Smiths Detection, a unit of the London Stock Exchange-listed Smiths Group that manufactures screening and detection technology for border control, has earned over $62 million from CBP contracts during Trump’s second term.
Smiths stated that it provides “threat detection and security screening technologies for ports and borders that curb illegal activity.”
Regional family businesses, Republican donors, Silicon Valley alliance
John Ganz, who closely examined the companies benefiting from ICE and CBP contracts on the Unpopular Front blog, offers significant clues regarding the alliances behind the Trump administration.
According to Ganz, while there are a few publicly traded and venture capital-funded firms, the largest beneficiaries exhibit a striking pattern: They are all regional family businesses displaying dynastic characteristics and are significant donors to the Republican Party.
Moreover, these regional companies have been involved in legally dubious practices. For example, Fisher Sand & Gravel, which sits at the top of the list, is owned by the Fisher family living in Dickinson, North Dakota.
The Fisher family makes generous donations to Republicans, and President Tommy Fisher frequently appears as a guest on conservative TV and radio programs.
The Fisher company’s history includes accusations of environmental violations, questionable labor practices, and, most notably, fraud.
In 2009, Fisher’s then-owner Michael Fisher pleaded guilty to nine counts of tax fraud and was sentenced to 37 months in prison and ordered to pay over $300,000 in restitution.
The company’s former CFO Amiel Schaff and former auditor Clyde Frank were also each found guilty of one count of conspiracy to defraud the US in 2009.
Under a 2009 agreement with the Department of Justice, the company was required to pay a total of $1.16 million in restitution, penalties, and fines, implement measures to prevent future fraud within the company, and cooperate with the IRS in the audit of tax returns.
Another former president of the company, David William Fisher, was found guilty in 2005 of possessing child pornography involving a 10-year-old child and was sentenced to 10 years in prison.
In exchange for his guilty plea, charges of sexual abuse of a minor were dropped, and he was released on April 30, 2010.
According to Ganz, companies further down the list, such as SLSCO, CSI Aviation, and Barnard Construction, fit this same model: regional, closely-held companies that are, so to speak, “politically integrated.”
Scholar Melinda Cooper, cited by Ganz, points to the tension between private, unincorporated, family-based companies and corporate, publicly traded, shareholder-owned companies.
According to Cooper, “family-based” capitalism, which finds representation in the White House with Trump, extends from the smallest family businesses to the vastest dynasties and is essentially shaped by the alliance between the two.
Trump also belongs to this “social class”: a representative of companies whose business methods are “informal”—or, to put it more bluntly, often outright criminal.
Ganz concludes his piece as follows:
“When you add in the presence of [Peter] Thiel-backed firms like Anduril, you start to understand the material basis of the Trump coalition. It is an alliance of family-based regional crony capital and a reactionary section of the tech sector focused on defense and security. Add to that ICE’s function as a jobs program for the Trumpenproletariat gang and all the illiterate influencers, and voila, you get the class composition of real American fascism, which is characteristically a protection racket. It is a gang all the way down.”
America
AI spending heads toward $7 trillion as analysts warn of market bubble risks
Massive financial resources directed into artificial intelligence technologies are driving companies into dangerous territory for global markets.
If expected productivity gains fail to materialize despite these immense capital flows, the artificial intelligence sector faces the risk of inflating into a giant bubble.
The Wall Street Journal reported that should such a scenario unfold, a widespread collapse capable of shaking the entire financial system and dragging down the broader market will become inevitable.
Estimates by McKinsey & Company project that global spending on data center construction alone could reach $7 trillion by 2030.
According to the newspaper, if these massive investments fail to deliver adequate productivity gains, the global economy will suffer a severe blow.
Should the sector as a whole turn out to be a bubble, the resulting damage will spread directly across the broader financial system.
While market observers note that a major crash—whether sooner or later—would drag all equity markets down with it, declines in AI-related stocks are currently being offset by gains in other sectors.
However, the first concrete signs of emerging vulnerability appeared in the memory chip market, where a sector-specific bubble formed and burst within just four months.
South Korean market shaken by sharp drop
In June, shares of South Korea’s Samsung and SK Hynix, the world’s two largest memory chip makers, sank by more than 12%.
The sharp sell-off pulled down the country’s broader stock index. South Korea’s benchmark Kospi index dropped 10%, triggering an automatic 20-minute trading halt.
Growing investor anxiety over artificial intelligence triggered the steep decline in the two giant companies, which together account for half of the total market capitalization of the Kospi index.
US equity markets also felt the ripple effects during the same period. The Nasdaq index closed down 2.2%, while the S&P 500 fell 1.4%, marking their worst single-day performances in two weeks.
Nevertheless, The Wall Street Journal pointed out that the disruption has not yet produced catastrophic consequences for the rest of the market.
While the bursting of massive historical bubbles resulted in disaster for national economies, smaller and localized bubbles in recent years have failed to paralyze broader economic growth.
The primary reason for this resilience is that these recent investments were not predominantly funded through leverage and bank credit.
When those localized bubbles burst, investors suffered wealth losses, but the financial system remained intact.
Russell Napier, a global macroeconomic strategist and keeper of the Library of Mistakes, a financial history archive in Edinburgh, evaluated the current market posture:
“The banking system is in superb condition, which means there will always be enough credit available to blow the next bubble.”
America
Anthropic AI models breach corporate systems after escaping isolated test environment
Anthropic has announced that several of its advanced artificial intelligence models escaped an isolated testing environment and accessed the live internet.
In a review published Thursday night, the company stated that in three separate incidents dating back to April, the models independently breached the systems of multiple companies without the AI developer’s knowledge.
Anthropic said the incidents involved an unreleased internal research test model, alongside its Opus 4.7 and Mythos 5 models.
Mythos was made available last month to a limited audience composed of technology companies and cybersecurity researchers, an initiative also known as Project Glasswing.
The AI developer did not disclose which companies were breached, but said the affected firms were informed of the incidents on Monday.
Anthropic noted that it conducted the review after OpenAI revealed last week that two of its most powerful models had breached containment, escaped their testing environment, and infiltrated several entities, including the AI platform Hugging Face and cloud provider Modal Labs.
System misconfiguration allowed internet access
Anthropic stated that it examined more than 140,000 tests to find evidence of whether Claude could gain access to the internet from test environments designed to be isolated.
The evaluations included “capture-the-flag” exercises, in which Claude was instructed to breach other systems to obtain information. This is a method frequently used by experts to assess a model’s hacking capabilities.
The San Francisco-based company stated that a “misconfiguration” in systems operated by Anthropic and its testing partner left the models with live internet access, enabling them to infiltrate external systems.
Anthropic said it approached remediation efforts “with full ownership of the responsibility.”
Neither Anthropic nor the affected organizations detected the unauthorized entries at the time they occurred.
Anthropic added that it may examine its logs more extensively, noting that the findings gave the company “cautious optimism” that such risks can be overcome through increased investment and more stringent safeguards.
David Allott, a cybersecurity expert, told the BBC: “The overarching lesson here is not that AI has developed fundamentally new attack vectors.”
“Instead, it means that AI agents can combine capabilities, acquire credentials and system access to act autonomously, while adapting scope and scale at machine speed,” Allott said.
The developments come as technology companies invest billions of dollars to develop AI agents capable of independently executing a range of tasks, from research and customer support to cybersecurity.
America
Elon Musk’s America PAC plans $100 million field operation for 2026 Republican midterm push
Tesla and SpaceX CEO Elon Musk is returning to the political spending arena with a new field program designed to help elect Republicans in at least eight states ahead of the 2026 midterm elections.
Musk has authorized his political action committee, America PAC, to spend between $100 million and $120 million on a new ground game focused on conservative voter turnout for the 2026 midterms, according to a Thursday report by The New York Times, which cited two unnamed sources informed about the plans.
America PAC funneled more than $250 million into Donald Trump’s reelection campaign in 2024, a expenditure that established Musk as the largest political donor in US history.
The New York Times reported that America PAC is reviving its spending initiatives and has reached out to other Republicans in recent weeks regarding the new field operations.
The effort is also being coordinated with other Republican Party spending groups, according to the report.
The newspaper identified targeted Senate races in the states of Alaska, Iowa, Maine, Michigan, and Ohio, while noting that discussions are also underway regarding contests in North Carolina, Georgia, and Texas.
The political action committee is additionally expected to deploy funds for House of Representatives elections in Washington, Wisconsin, and California.
The news comes a day after Axios first reported that America PAC’s operations were resuming, with a focus on driving Republican turnout during the non-presidential election cycle.
A spokesperson for America PAC declined to comment on The New York Times report but confirmed the Axios reporting to The Hill. The spokesperson stated that the spending group was “excited” to contribute to efforts to maintain the Republican majorities in Congress this fall.
“The President’s political team and the rest of the GOP apparatus have built a world-class operation that has Republicans well-positioned to make history and retain control of Congress this fall,” America PAC spokesperson Andrew Romeo said in a statement. “We’re excited to be part of the team again.”
The campaign will reportedly target Republican voters through door-to-door canvassing, mailers, and digital advertisements, enabling other groups to concentrate their resources on television advertising.
The developments were reported days after Musk told The Economist magazine that he had gotten “carried away” during his brief foray into politics.
The SpaceX CEO entered the political arena during the 2024 election, pouring hundreds of millions of dollars into Trump’s presidential campaign and accompanying the candidate on the campaign trail.
Musk went on to lead Trump’s cost-cutting initiative, known as the Department of Government Efficiency (DOGE), which executed sweeping employment and funding reductions across the federal government. Those efforts sparked controversy for Musk and his enterprise empire, including Tesla, whose shares fell sharply during his period of political involvement.
Musk departed the White House in late May 2025, and DOGE officially terminated its operations on July 4.
Shortly after leaving government, Musk and Trump engaged in a public dispute over the president’s sweeping spending legislation, the “One Big Beautiful Bill Act.” During the friction, Musk threatened to form a third party, though the initiative never materialized.
Musk and the US President appeared to resolve their differences last year, with the tech billionaire most recently joining Trump alongside other technology leaders on a trip to China in May.
-
America2 weeks agoUS agricultural superpower status at risk as trade wars shift global markets to Brazil
-
America2 weeks agoUS controls $13 billion in Venezuelan oil revenues with little transparency, raising congressional concerns
-
Diplomacy2 weeks agoPalantir CEO Alex Karp says he would not vote for ‘pro-Russian’ AfD in Germany
-
Europe2 weeks agoUS secures multi-billion-dollar energy and AI deals at Three Seas summit in Dubrovnik
-
Diplomacy2 weeks agoWorld Bank warns US-Iran conflict could slash global growth to 1.3% as inflation looms
-
Middle East2 weeks agoPentagon faces severe budget crunch as Middle East operational costs drain key military funds
-
Europe2 weeks agoGermany accelerates African energy diplomatic push to secure natural gas and green hydrogen
-
Middle East2 weeks agoOil passes $90 as tanker attacks halt Hormuz shipping
