TikTok has made a decision to roll out a new age-detection system across Europe, reflecting a recalibration by the platform as governments move from warning shots to hard regulation on children’s access to social media, Reuters reports.

The move follows Australia’s enactment of a landmark law restricting children under 16 from using social media, a policy shift that has already forced platforms to shut out large numbers of young users and sharpened pressure on regulators elsewhere to act.

The ByteDance-owned company said the new system, which will be deployed in the coming weeks, is designed to better identify accounts belonging to children under 13, TikTok’s global minimum age. Unlike traditional age gates that rely on self-declared birthdays, the technology analyzes a mix of profile information, posted videos, and behavioral signals to assess whether an account is likely underage.

Accounts flagged by the system will not be automatically removed but sent to specialist moderators for review, a step TikTok says is meant to reduce errors while complying with Europe’s strict data-protection rules.

The rollout builds on a year-long pilot in Britain that resulted in the removal of thousands of additional underage accounts. The UK trial offered TikTok both evidence that the approach could work and a template that could be adapted for the rest of Europe, where regulators have long complained that platforms are failing to enforce their own age limits.

Australia’s under-16 social media law, the most far-reaching of its kind globally, has altered the regulatory conversation. By placing the burden squarely on platforms to prevent children from accessing social networks, the law has effectively cut off millions of minors from mainstream social media services and demonstrated that governments are willing to accept disruption in the name of child safety. European policymakers have taken note. The European Parliament is pushing for clearer age limits, while countries such as Denmark are openly discussing bans for children under 15.

Against that backdrop, TikTok’s Europe-focused system looks less like a voluntary upgrade and more like a pre-emptive response to the risk of tougher mandates. European authorities are increasingly skeptical of approaches they view as either ineffective or overly invasive, and age verification sits at the center of that tension. Demanding identity documents from all users raises privacy concerns under the General Data Protection Regulation, yet self-reporting has proved easy to circumvent.

TikTok says its solution is designed to thread that needle. The company argues there is no globally agreed method for confirming age while preserving privacy, and that inference-based systems offer a middle ground. For users who appeal a ban, TikTok will still rely on more explicit checks, including facial age estimation from verification provider Yoti, credit card verification, and government-issued identification. Meta also uses Yoti to verify ages on Facebook, pointing to a broader industry shift toward layered verification rather than a single gatekeeping tool.

The regulatory pressure is not abstract. On Friday, a judge in Delaware was scheduled to hear TikTok’s bid to dismiss a lawsuit filed by the parents of five British children who died while allegedly attempting online challenges. The lawsuit alleges that TikTok’s algorithms promoted dangerous content to children, including the so-called “blackout challenge.”

TikTok has expressed sympathy for the families and said it strictly prohibits content that encourages harmful behavior, but the case underscores the legal exposure platforms face when underage users slip through enforcement gaps.

TikTok has emphasized that the new technology was built specifically for Europe and developed in consultation with Ireland’s Data Protection Commission, its lead privacy regulator in the EU. European users will be notified as the system goes live, the company said, a nod to transparency requirements under EU law.

However, it is not clear whether the approach satisfies regulators. Inference-based systems risk false positives that could lock out legitimate users, as well as false negatives that allow underage accounts to persist. There are also unresolved questions about how transparent platforms will be about the signals they use and how consistently human moderators apply standards at scale.

What is clear is that the regulatory climate has shifted. Australia’s under-16 ban has shown that governments are willing to impose blunt restrictions if they believe platforms are not acting fast enough. TikTok’s European rollout suggests the company is betting that smarter detection, rather than outright bans, can convince regulators that tougher measures are unnecessary. The coming months will test whether that bet holds as Europe weighs its next steps on child safety online.

The Trump administration is moving to redraw the balance of power costs in the U.S. electricity market, pressing the nation’s largest grid operator to make technology companies shoulder a greater share of the burden created by the AI boom.

At the center of the push is PJM Interconnection. This vast regional grid supplies electricity to more than 65 million people across 13 states and Washington, D.C. PJM also happens to sit at the heart of the global data center industry, with northern Virginia alone hosting the largest concentration of facilities anywhere in the world. Those data centers, built to train and run artificial intelligence models, are now reshaping the economics — and the politics — of power.

According to a White House official cited by CNBC, President Donald Trump wants PJM to hold an emergency auction that would force large technology companies to bid directly on contracts for new power plants. The administration is seeking roughly $15 billion in new baseload generation, capacity designed to run around the clock, not just when the sun shines or the wind blows. At the same time, Trump wants limits placed on what existing power plants can charge in PJM’s capacity market, a move aimed at protecting households already grappling with rising utility bills.

Energy Secretary Chris Wright, Interior Secretary Doug Burgum, and governors from the mid-Atlantic region are expected to announce a joint agreement on Friday, urging PJM to adopt these measures. The White House framed the effort as a corrective to years of underinvestment and policy drift.

“Under President Trump’s leadership, the administration is leading an unprecedented bipartisan effort urging PJM to fix the energy subtraction failures of the past, prevent price increases, and reduce the risk of blackouts,” spokeswoman Taylor Rogers said.

But behind the rhetoric sits a stark reality. Electricity prices across the PJM region have surged in recent years, and watchdogs say data centers are a major driver. Monitoring Analytics, which oversees PJM’s markets, estimates that about $23 billion in capacity costs can be traced directly to the explosion of data center demand. Those costs are ultimately passed through to consumers, a dynamic the watchdog has described as a “massive wealth transfer” from households to large power users.

In its most recent capacity auction, PJM came up six gigawatts short of its reliability requirement for 2027. That shortfall is equivalent to the output of six large nuclear plants. Without a new generation, the margin for error shrinks dramatically.

“Instead of a blackout happening every one in 10 years, we’re looking at something more often,” said Abe Silverman, a researcher at Johns Hopkins University and former general counsel to New Jersey’s public utility board.

The political stakes are rising alongside the technical risks. Trump campaigned on a promise to lower energy costs, yet utility bills have continued to climb in many parts of the country. High power prices were a major undercurrent in the landslide victories of Democrats Mikie Sherrill and Abigail Spanberger in the recent governors’ races in New Jersey and Virginia, two states deeply tied to the PJM grid.

By pushing PJM to make tech companies pay directly for new generation, the administration is signaling a shift in how it views the AI economy’s infrastructure needs. For years, data centers have been welcomed as engines of investment and tax revenue, often receiving favorable treatment from states and local governments. Now, as their electricity consumption rivals that of entire cities, the question has become who pays to keep the lights on.

Capping capacity prices would also mark a significant intervention in PJM’s market design, one that power producers are likely to resist. Generators argue that high prices are necessary to attract investment, especially as older coal and gas plants retire and new projects face regulatory and supply-chain hurdles. The administration, however, appears willing to test those arguments as it seeks to blunt the impact on consumers.

The outcome will shape not just PJM but the broader U.S. power system, as grids nationwide confront the same challenge: an AI-driven surge in demand colliding with aging infrastructure and years of underbuilding. Whether emergency auctions and price caps can deliver new power quickly enough — without triggering unintended consequences — remains an open question.

What is clear though, is that electricity, once a quiet background issue, has become a frontline political and economic battleground. And in Trump’s view, the era of ordinary ratepayers quietly subsidizing Big Tech’s power hunger may be coming to an end.

Dutch semiconductor equipment company ASML shares climb to fresh record highs following TSMC’s blockbuster earnings, underlining a quiet but powerful truth in the global semiconductor race that no matter how fierce the competition among chip designers and manufacturers becomes, the real bottleneck — and leverage point — sits with the companies that control the tools.

Shares in the Dutch semiconductor equipment maker have risen about 7% since TSMC reported earnings on Thursday, pushing ASML’s market value above the $500 billion mark and making it only the third European company ever to cross that threshold. The stock is now up roughly 25% in 2026, extending a rally that has increasingly been driven less by hype around artificial intelligence and more by hard spending commitments from the world’s biggest chipmakers.

At the center of ASML’s appeal is not growth alone, but scarcity. The company remains the world’s only supplier of extreme ultraviolet lithography machines, the tools required to manufacture the most advanced semiconductors used in AI accelerators, data centers, and high-performance computing. No viable alternative exists at scale, and replicating ASML’s technology would take years, if not decades, even for well-funded rivals.

That monopoly is becoming more consequential as the AI race shifts from experimentation to infrastructure. TSMC’s capital expenditure guidance, which significantly exceeded expectations, sent a clear signal that demand for advanced chips is not peaking. Instead, it is spreading across industries and regions, forcing manufacturers to lock in long-term investment plans. Bank of America said this dynamic underpins near-term upside for ASML, as EUV and other advanced tools become essential for improving yields and energy efficiency at cutting-edge process nodes.

Morgan Stanley took the argument further, saying ASML shares could rise as much as 70% in a bullish scenario. The bank expects higher spending by foundries and memory producers into 2027, combined with better-than-feared demand from China, to drive earnings upgrades. Its base case price target sits at 1,400 euros, but its bull case envisions shares approaching 2,000 euros if tech valuations continue to expand and margins surprise to the upside.

The memory segment is a critical part of that story. AI chips rely heavily on fast, high-density memory, and demand for Dynamic Random Access Memory has surged alongside Nvidia and AMD’s most advanced processors. Counterpoint Research expects memory prices to rise by another 40% to 50% in the first quarter of 2026, adding urgency to capacity expansion plans. JPMorgan said Samsung, the only major DRAM producer with available clean room capacity, is likely to increase orders for equipment most aggressively, with TSMC also expected to place strong orders.

What makes ASML’s position even stronger is that this wave of investment is increasingly being shaped by geopolitics. Governments are no longer passive observers of the semiconductor cycle. The United States, Europe, and key Asian economies are actively steering where chips are made, how supply chains are structured, and which technologies are prioritized. TSMC received another boost this week after Washington said tariffs on Taiwan would be limited to 15%, following commitments by Taiwanese chip and technology firms to invest at least $250 billion in U.S. production capacity.

That policy push aligns neatly with ASML’s business model. Whether new fabs are built in Arizona, Texas, Germany, or Japan, they all require the same high-end lithography systems. For ASML, the location of production matters less than the scale and urgency of it. In effect, industrial policy has become an indirect demand engine for its machines.

Recent signals from the broader AI ecosystem reinforce the momentum. Foxconn, a key Nvidia partner and the world’s largest contract electronics manufacturer, reported a 22% surge in revenue in the final quarter of 2025, reflecting strong demand for AI servers used in data centers. Those servers ultimately house chips produced using ASML’s tools, tying the company’s fortunes to every layer of the AI buildout.

Yet the rally also comes with longer-term questions. ASML’s dependence on a small number of very large customers, particularly TSMC, means its outlook remains closely tied to the investment discipline of a handful of players. Export controls, especially those affecting sales to China, remain a risk, even as analysts note that demand has held up better than expected. At the same time, valuations across the AI supply chain are rising rapidly, raising the stakes for upcoming earnings reports.

ASML is due to report fourth-quarter results on Jan. 28, with investors expected to focus less on near-term revenue and more on order intake, backlog strength, and management’s view of customer spending beyond 2026. Confirmation that capital expenditure plans remain intact could reinforce the view that the current rally is not speculative, but structural.

What ASML’s ascent ultimately highlights is a shift in where strategic power sits in the AI economy. While attention often gravitates toward chip designers, cloud providers, and software breakthroughs, the constraints are increasingly physical. The ability to manufacture at the most advanced nodes has become the limiting factor, and ASML sits squarely at that choke point.

In a global race defined by speed, scale, and sovereignty, ASML is not just supplying machines. It is shaping the pace at which the AI future can be built.