HardwareFreaks

Tag: Export controls

  • Supermicro denies that its offices were raided by Taiwanese authorities in Nvidia GPU smuggling case — company says that it coordinated with the police and provided access to investigated employees’ workstations and gadgets

    Featured image Supermicro denies that its offices were raided by Taiwanese authorities in Nvidia GPU smuggling case company says that it coordinated

    The high-stakes world of artificial intelligence is increasingly defined by complex trade regulations, and recently, an investigation into the movement of advanced AI hardware brought these global tensions directly into focus. Taiwanese authorities have launched a significant probe into alleged smuggling of Nvidia AI chips into China, targeting employees within the technology sector.

    The situation escalated when police reportedly conducted a raid on the Taipei office of Supermicro, a major player in the server industry, as part of this broader examination. While the event generated immediate attention, the company quickly stepped up its defense, asserting full cooperation with investigators. Supermicro Chief Revenue Officer Matt Thauberger confirmed that the firm has been working alongside authorities since May of this year, voluntarily providing access to workstations and electronic devices of the suspected employees and placing them on administrative leave.

    Supermicro maintained that the police intervention was not a formal “raid” but rather a cooperative measure, emphasizing their commitment to zero tolerance for policy violations. Despite their cooperation, the investigation underscores the immense pressure placed on technology companies operating in sensitive international supply chains.

    This inquiry is part of Taiwan’s determined effort to enforce regulations concerning Nvidia AI chip exports. While Taiwan does not have export control laws mirroring those of the United States, it utilizes a loose interpretation of existing regulations to manage the flow of these critical components across borders.

    The alleged smuggling scheme involved sophisticated methods designed to obfuscate the movement of hardware. Reports suggest that some individuals linked to the company utilized devices, such as hairdryers, to soften serial numbers on real servers and move them between legitimate hardware and thousands of dummy units. These manipulated items were then allegedly shipped through entities based in Thailand before landing in Chinese tech giants’ warehouses.

    Beyond the specific Supermicro case, the investigation has led to wider arrests in the United States. Several other individuals connected to the company, including co-founder Yih-Shyan “Wally” Liaw, have been detained on charges related to conspiring to violate export controls. This broader context highlights how deeply intertwined technology supply chains are with international legal frameworks.

    The implications of these actions ripple throughout the industry. The situation has prompted major players to reassess their compliance protocols. Nvidia’s CEO, Jensen Huang, publicly urged Supermicro to tighten its export compliance controls, signaling that the integrity of the supply chain is now paramount for maintaining competitive advantage.

    This scrutiny has had immediate financial repercussions. News of the investigation and subsequent legal actions led to an 8% slide in Supermicro’s stock price during U.S. trading, underscoring how swiftly geopolitical investigations can translate into market volatility for major technology firms.

    Buy on Amazon

    Credit: Tom’s Hardware

  • Singapore cops seize $42 million mansion, freeze $772k bank account of suspected Nvidia AI GPU smugglers — individuals alleged to have illegally exported data center servers to China charged with fraud, money laundering

    Featured image Singapore cops seize 42 million mansion freeze 772k bank account of suspected Nvidia AI GPU smugglers individuals alleged to have ille

    The intersection of cutting-edge artificial intelligence and global commerce has recently shone a harsh light on the complexities of international trade and regulatory enforcement. A sophisticated scheme involving the smuggling of high-tech components, particularly AI chips, has led Singaporean authorities to pursue major money laundering charges, revealing a complex network operating in the shadow of global technology.

    The investigation originated in the United States, spurred by inquiries into the operations of the DeepSeek frontier model released in late 2024. Officials were examining whether this groundbreaking AI company utilized third-party firms based in Singapore to circumvent restrictions on exporting Nvidia GPUs to China.

    The probe uncovered a fascinating logistical reality: although Singapore holds a significant portion of Nvidia’s revenue, the actual flow of these crucial chips was drastically limited. The investigation revealed that only about 1% of Nvidia’s GPU sales were actually delivered to China through this specific route, highlighting how easily international trade controls can be bypassed.

    This regulatory gap provided fertile ground for illicit activity. The smuggling operation involved using Singapore as a critical transshipment point, allowing highly restricted technology to move across borders undetected. This scheme was masterminded by individuals who were involved in purchasing these servers and forwarding them illegally.

    The investigation ultimately led to significant arrests. In the first quarter of 2025, police actions resulted in the apprehension of Woon Guo Jie Aaron, along with fellow Singaporean Wei Zhaolun Alan and Chinese citizen Li Ming, for their roles in this illicit trade.

    Beyond the smuggling charges, the scope of the criminal activity extended deeply into financial crimes. Authorities uncovered that the suspects had accumulated over USD 926,000 (SGD 1.2 million) across their bank accounts as proceeds from these illegal activities. Furthermore, assets acquired through these laundered funds were seized, including a mansion valued at $42 million (SGD 55 million).

    The financial trail also included the freezing of USD 772,000 (SGD 1 million) held in a bank account, further illustrating the scale of the illicit wealth involved.

    Because Singaporean law does not directly cover the U.S.’s export controls, prosecutors chose to charge the individuals with fraud and money laundering. This dual approach reflects the need to address both the criminal flow of funds and the illegal handling of regulated technology.

    Wei Zhaolun Alan is facing separate money laundering charges concerning USD 4.5 million (SGD 5.8 million) stored in his personal accounts, which were determined to be proceeds of criminal conduct.

    The legal consequences for these actions are serious. If found guilty, the suspects could face imprisonment of up to twenty years each, along with substantial fines potentially reaching USD 385,000 (SGD 500,000).

    Singapore police issued a strong statement emphasizing their commitment to maintaining order and integrity. They affirmed a zero-tolerance stance against such violations, stating they will act resolutely against any individuals or businesses that contravene the law, thereby safeguarding Singapore’s standing as a trusted global hub underpinned by the rule of law.

    Credit: Tom’s Hardware

  • Anthropic restores Claude Fable 5 as US lifts export controls — single filter now blocks prompt that could identify software vulnerabilities and write code to exploit them

    Featured image Anthropic restores Claude Fable 5 as US lifts export controls single filter now blocks prompt that could identify software vulnerabili

    The Great AI Standoff: How a Single Safety Filter Unlocked the Future of Claude

    After an 18-day diplomatic standoff, Anthropic has successfully restored global access to its flagship model, Claude Fable 5. The release, which occurred just a day after the U.S. Department of Commerce lifted export controls imposed on the model on June 12th, marks more than just a technical fix; it signals a fascinating and often contentious negotiation between technological capability and geopolitical regulation.

    The ability to deploy advanced AI models worldwide is inherently complicated by questions of national security and origin. In this case, Anthropic faced restrictions that barred foreign nationals, including its own non-citizen staff, from using Fable 5 and its more powerful Mythos 5. The core dilemma was verification: without a way to confirm the nationality of users, the company had been forced to pull both models globally.

    The resolution required surgical precision. The breakthrough came in the form of a single safety filter, meticulously tuned to block one specific, highly contentious technique flagged by Amazon researchers. This intervention ended the stalemate and allowed the models to return to their users across Claude.ai, the Claude Platform, Claude Code, and Claude Cowork.

    What was the trigger? Amazon researchers discovered a method to prompt Fable 5 into revealing software vulnerabilities and demonstrating how those flaws could be exploited in code. To prevent this kind of dangerous demonstration, Anthropic developed a new classifier that targets this specific request with greater than 99% accuracy. When such prompts are detected, the system reroutes the request, often sending it to the older Opus 4.8 model.

    This move highlighted a key tension in AI safety: the difference between restricting what a model can do and restricting how it is allowed to be prompted. The new classifier was designed to block the dangerous technique, not strip Fable 5 of its underlying analytical capabilities. This subtle distinction reveals that detection-based safeguards—the very mechanisms that initially triggered the export ban—were also part of the challenge.

    The ongoing reality remains complex. Anthropic acknowledged that no model can be made entirely immune to ‘jailbreaks,’ suggesting that new methods for bypassing safety measures will continue to emerge. This realization underscores a crucial lesson: AI development is not about achieving absolute robustness, but about continuous, dynamic adaptation in response to evolving threats.

    Meanwhile, the broader landscape of AI performance saw shifts. Fable 5’s return helped reclaim benchmark positions that had previously been held by other leading models, including those developed by Chinese labs, demonstrating the power and relevance of this new generation of models on a global scale.

    To foster further transparency and safety, Anthropic has also initiated community engagement. They opened a HackerOne program inviting researchers to report newly discovered Fable 5 jailbreaks. Furthermore, they committed to granting designated government partners earlier access to test future frontier models before they are publicly released, positioning the company at the forefront of collaborative AI governance.

    Credit: Tom’s Hardware

  • Chinese Z.ai’s latest model tops AI ranking charts amid Anthropic Fable 5 ban — blacklisted China firm’s popular open-weight GLM-5.2 AI model powered by Huawei silicon

    Fable 5 and Mythos 5. This move immediately ignited a fascinating technological counter-response from Beijing.

    Just two days later, a rival AI entity, Z.ai, began rolling out GLM-5.2. Crucially, this new model was reportedly trained entirely on domestic hardware—specifically Huawei Ascend chips—completely bypassing reliance on Nvidia silicon. This development signaled that the global race for cutting-edge AI is not just about software, but a fierce battle over foundational technology and supply chains.

    Within a week of this digital standoff, GLM-5.2 quickly became a force, climbing to the top of openly available leaderboards and achieving remarkable performance metrics that put it squarely in competition with the U.S. giants. It rapidly escalated Z.ai’s market value, pushing its valuation past the one-trillion-dollar mark in Hong Kong.

    The competitive landscape proved incredibly nuanced. While Fable 5 faced regulatory restrictions, GLM-5.2 demonstrated strong capability across various tests. In tasks focusing on human preference coding, GLM-5.2 edged ahead of its competitors. Furthermore, in multi-week knowledge tasks—where models test their ability to synthesize vast amounts of fragmented information—Fable 5 still led in one specific assessment, underscoring the complexity of measuring true intelligence.

    However, when looking at raw performance and accessibility, the picture shifts dramatically. GLM-5.2’s open-weight nature means anyone can download, fine-tune, and self-host its weights, a stark contrast to closed systems. Despite this openness, the sheer scale of the model remains demanding: it requires substantial enterprise GPU clusters, emphasizing that creating frontier AI is less about code and more about infrastructure.

    The hardware story adds another layer of complexity. Z.ai’s success relies on its proprietary stack, demonstrating that domestic silicon can indeed handle training-class jobs. Nevertheless, a significant gap remains between the performance of these specialized chips and the leading Nvidia technology. While reports suggest certain Chinese clusters can handle full-parameter post-training of models like DeepSeek’s V4, it underscores a critical point: model parity does not automatically equate to hardware parity.

    The restrictions imposed by the U.S. government on advanced AI chips are designed to control access to high-end computing resources. Yet, as industry observers note, the potential for domestic silicon to close this gap is real. Meanwhile, the market is buzzing with anticipation. As investment lock-ups expire and whispers circulate about a Chinese equivalent to Fable 5 arriving sooner than expected, the stakes for the future of AI dominance are reaching a fever pitch.

    Credit: Tom’s Hardware

  • OpenAI’s ChatGPT-5.6 gets the same banhammer treatment as Anthropic’s Mythos from the federal government — source says that Washington cautioned OpenAI against releasing the model without receiving approval

    The world of artificial intelligence is currently navigating a fascinating intersection where groundbreaking innovation meets intense governmental scrutiny. At the heart of this dynamic tension is the race to develop powerful new models, shadowed by critical questions about safety, security, and global competition.

    OpenAI CEO Sam Altman recently addressed these concerns during a staff meeting, announcing that their newest model, GPT-5.6, is currently available only in a limited preview to a select group of customers chosen by the U.S. government. This move signals an immediate shift toward cautious deployment, as federal bodies, specifically the Office of the National Cyber Director and the Office of Science and Technology Policy, have requested that OpenAI stagger the release of the model.

    While Altman expressed hope for a swift timeline—suggesting a potential general release within a couple of weeks—the reality is that access to GPT-5.6 remains strictly case-by-case. This controlled rollout reflects a growing realization that the speed of AI development must be balanced against robust security protocols.

    The push for caution is not new in the AI landscape. Earlier attempts by other major labs have required significant strategic maneuvering. For instance, when Anthropic released its Claude Mythos Preview, they opted to provide access first to key institutions. This strategy allowed them time to implement necessary safeguards before eventually releasing a refined version, Fable 5.

    However, this cautious approach did not always go smoothly. The U.S. government subsequently placed both Fable 5 and Mythos on an export control list just days after their release, effectively restricting access for foreign nationals. This experience underscored the high stakes involved when powerful AI models cross international boundaries.

    As the AI arms race intensifies, the focus shifts sharply to geopolitical strategy. The competition with rival nations, particularly China, adds another layer of complexity. While the U.S. has implemented export controls to slow Beijing’s progress, industry experts anticipate that rapid advancements mean the East Asian country will soon catch up in achieving a similar level of AI supremacy.

    Adding another layer of regulatory complexity, President Donald Trump recently signed an executive order mandating that U.S. AI labs provide the government with access to their latest models 30 days before a general public release. This move attempts to create a structured pathway for government oversight in the development process.

    Yet, this escalation of government intervention has sparked debate among experts. Some worry that the constant deployment of export controls and regulatory mandates creates an environment where companies slow down innovation unnecessarily. Experts caution that arbitrary interventions can hinder the broader AI ecosystem, creating uncertainty about when and how powerful new tools will be released to the public.

    Credit: Tom’s Hardware

  • Anthropic’s powerful Mythos AI reportedly breached ‘almost all’ NSA classified systems within a few hours during red-team test — report sheds more light on the U.S. government’s sudden ban on the flagship models

    The Ghost in the Machine: How AI Security Tests Blurred the Line Between Innovation and Espionage

    In the high-stakes arena of artificial intelligence, the power of models like Anthropic’s Mythos has sparked a fascinating, if slightly alarming, conversation about security boundaries. The story emerged when reports surfaced regarding the model’s capability to penetrate highly classified systems, forcing a public reckoning over how rapidly AI technology is evolving and the controls put in place to govern it.

    The controversy centered on a controlled security evaluation where the Mythos AI model reportedly accessed almost all classified systems belonging to the National Security Agency (NSA) in a matter of hours. This claim, initially circulating as sensational news, quickly ignited debate about the limits of governmental cybersecurity protocols and the very nature of AI threats.

    The initial narrative suggested a dramatic breach, but subsequent clarifications revealed a more nuanced reality. The event was not an autonomous offensive intrusion, but rather an authorized internal red-team test. During this controlled exercise, Mythos was paired with defensive tools under specific simulated environmental conditions, allowing researchers to observe how the model might react to vulnerabilities.

    The context for this test is layered with government directives. Earlier, the U.S. government had implemented export controls that barred foreign nationals, including Anthropic’s own non-citizen employees, from accessing the Fable 5 and Mythos 5 models, citing national security concerns. In response, Anthropic globally disabled access to these models.

    This regulatory push marked a pivotal moment: the United States became one of the first nations to apply export controls directly to an AI model rather than just the underlying hardware. The specifics of the initial concerns remained vague, with Anthropic citing only verbal evidence related to a potential narrow “jailbreak” that could expose software vulnerabilities.

    The context surrounding the alleged breach further illuminated the complexities of the situation. The security evaluation took place just one day before the export ban was issued, leading some observers to question the timeline and the focus of the reported incident. However, Anthropic maintained that the flagged behavior amounted only to identifying minor, already known bugs within a codebase, not an actual hostile takeover.

    Beyond the technical details, the incident highlighted the unique, close relationship between the AI developers and intelligence agencies. While public discourse on platforms like Reddit was split—some viewing the event as an indictment of government security capabilities, others dismissing it as marketing spin—a different dynamic is at play. Anthropic continues to work closely with the NSA under specialized arrangements known as Project Glasswing.

    This partnership involves embedding roughly six Anthropic engineers directly inside the agency, customizing Mythos for specific operational applications. This arrangement suggests a deep integration of cutting-edge AI development into sensitive national security operations, pushing the boundaries of what it means to deploy advanced technology in the public sphere.

    Credit: Tom’s Hardware