Not long ago, the United States seemed too optimistic in its ambition to be the global frontier in artificial intelligence. Grand policy announcements and billions in funding were meant to reclaim America’s technological prowess and undo decades of offshoring. Moreover, back in 2022, the White House launched the CHIPS Act, which was intended to lure microchip manufacturing back to the United States.
While Washington touted its initiatives, China aggressively secured supply chains, scaled infrastructure and integrated artificial intelligence into every facet of its economy. Meanwhile, the US remained stuck in bureaucratic inertia, mistaking press releases for progress.
Adding to this chaos, The Department of Government Efficiency (DOGE) office, under Elon Musk’s leadership and President Trump’s mandate, is executing one of the most aggressive purges of the U.S. federal workforce in modern history — with little transparency or accountability. Independent data indicates more than 216,000 jobs in March were slashed, many from AI research and policy.
The very teams meant to shape the American technological future are now casualties of a reckless downsizing spree. The CHIPS Act was meant to restore America’s leadership in artificial intelligence. $280 billion were invested in semiconductor research and manufacturing.
Of this, $52 billion came from federal funding and tax breaks. Major players like Intel and TSMC received significant support, while billions were also promised to Samsung and SK hynix.
But these layoffs are already crushing the country’s AI talent pipeline, hollowing out expertise at a time when the stakes in artificial intelligence have never been higher.
Big Promises, Bigger Pitfalls
Big tech AI firms like Apple, Nvidia, Intel and AMD rely heavily on GPUs and AI accelerators because they are designed for the computationally intensive tasks of training and running AI models. Although the US leads in chip design and software for AI, the actual manufacturing of these high-end semiconductors still takes place overseas. The CHIPS Act was supposed to change that by reshoring production, and by 2032, the US is projected to control nearly 30 percent of the world’s fabrication capacity for cutting-edge chips below 10 nanometers. But this focus on the most advanced semiconductors misses the bigger picture. China is quietly taking over legacy chip production, and the world is barely paying attention.
What Are Legacy Chips?
Legacy chips are mature-node semiconductors may not be cutting edge, but they are the backbone of countless everyday technologies, powering everything from cars and medical devices to industrial machinery and defense systems.These chips are indispensable.
By 2032, China is expected to control nearly 40 percent of the global supply of these legacy chips, compared to just 10 percent for the US. This gives China enormous leverage to disrupt supply chains and even cripple America’s economy with targeted export bans. While Washington obsesses over AI chip restrictions, it is sleepwalking into a far more fundamental vulnerability.
Adding to the problem, President Donald Trump has vowed to roll back the CHIPS Act entirely, dismissing it as a waste of money. Instead, his administration is doubling down on export bans designed to frustrate China’s AI ambitions. But history shows that these coercive measures don’t work — if anything, they accelerate Chinese innovation.
Why the US Needs a New Playbook for AI
DeepSeek’s AI models, which rival OpenAI’s technology, were reportedly developed at a fraction of the cost and relied on chips smuggled through intermediary nations like Singapore. Huawei’s ability to produce 7nm chips despite U.S. restrictions further highlights the futility of Washington’s current approach.
The reality is that the US cannot win the semiconductor war through isolationism, yet current policy seems to assume otherwise. The semiconductor supply chain is one of the most interdependent systems ever built. No country — not even the United States — can unilaterally dominate chip production.
Washington loves to talk about “reshoring” and “tech sovereignty,” but let’s be clear: the United States cannot build advanced chips without its allies. Although U.S. companies like NVIDIA, AMD, and Qualcomm lead in chip design, they lack the manufacturing tools to turn those designs into reality. That’s where ASML, a Dutch company with a monopoly on extreme ultraviolet (EUV) lithography machines, plays a crucial role. Without ASML, there are no advanced chips.
The US’s strength has never been self-sufficiency — it has been through its global network of partners and allies. China, on the other hand, is taking a go-it-alone approach, pouring billions into domestic chipmaking to replace foreign technology.
If Washington does not rethink its semiconductor strategy, it may win the AI race but lose control over the foundational technology that keeps its economy and military running. Chips and the data they power are the oil of the future. The US cannot afford to get this wrong.
DOGE’s Access to Databases Sparks Privacy Concerns
Besides that, under the guise of “efficiency,” DOGE has gained access to government databases holding medical records, financial histories, and tax filings, and this has raised serious concerns about privacy and oversight. The team also has access to at least seven major federal databases, including those belonging to the IRS and the Social Security Administration.
At least a dozen lawsuits have been filed, as unions and privacy advocates scramble to block DOGE’s access to IRS, Social Security and Medicare records. The government has long operated under strict firewalls to prevent misuse of personal data — firewalls that DOGE now appears to be dismantling.
Privacy advocates are worried that these rich datasets could be used to train AI for Musk’s xAI or similar private companies. The risk is more than potential security breaches. Training advanced machine learning systems on personal records may lead to unfair profiling, biased policy decisions and discriminatory practices in services like insurance or credit.
The administration insists DOGE’s access is “read-only,” but that assurance does little to address the broader issue: This level of data centralization hands extraordinary power to a small group with minimal oversight. The real question is not whether DOGE can access sensitive data, but why it wants to, and what happens when it does?
What’s Ahead for the US and AI?
Washington’s current approach shows a troubling lack of readiness for the next wave of machine learning. When a hastily developed app like DOGE AI supplants seasoned federal employees, it yields questionable outcomes. The government loses institutional knowledge when cutting professionals and leaning on undercooked automation tools. And it occurs right when algorithmic bias, cybersecurity vulnerabilities, and ethical data usage demand heightened scrutiny.
A constructive path supposes embracing more decentralized approaches that reduce single points of failure at a governmental or corporate level. From my perspective, distributing AI research, data governance and accountability across diverse stakeholders can help avert abuses of power. In practice, this means embedding transparency, fairness, and oversight into every layer of AI development and deployment.
We create natural guardrails against overreach through ensuring no single gatekeeper holds all the data or decision-making authority. Privacy-preserving frameworks and decentralized ledgers offer further assurance that personal information will not be misused or exploited for corporate gain.
