White House Office of Science and Technology Policy director Michael Kratsios circulated a memo on April 23 accusing Chinese entities of conducting deliberate, industrial-scale campaigns to distil and steal US frontier AI systems, using tens of thousands of proxy accounts and jailbreaking techniques, days before a scheduled Trump-Xi summit.

White House OSTP director Michael Kratsios circulated a memo to US government agencies on April 23 stating that foreign entities “principally based in China” are “engaged in deliberate, industrial-scale campaigns to distil US frontier AI systems.” The accusation represents the most formal and direct statement the Trump administration has made about Chinese AI intellectual property theft from American laboratories and arrives days before a scheduled Trump-Xi summit.

White House China AI Theft Accusation Names Distillation as the Attack Vector

Distillation is the technical process by which smaller AI models are trained on the outputs of larger, proprietary ones, allowing a lab to approximate frontier performance without paying the full training cost associated with building from scratch. According to the Kratsios memo, Chinese actors are using tens of thousands of proxy accounts to make large-scale queries to American AI systems, evading the rate limits and detection systems that companies have built to flag suspicious usage patterns. Jailbreaking techniques are then applied to strip safety filters and expose the proprietary behavior of models, which is captured as training data for distilled copies. Both OpenAI and Anthropic have previously accused DeepSeek publicly of distilling their models. The Financial Times reported that the memo is addressed to US government agencies and signals that the administration will begin sharing intelligence about active distillation campaigns directly with US AI companies. As crypto.news reported, Nvidia CEO Jensen Huang had warned days earlier that China possesses ghost datacenters with enormous computing infrastructure capable of matching US frontier AI capabilities.

DeepSeek’s V4 Launch Sharpened the Timing of the Accusation

DeepSeek released preview versions of its V4 model on April 24, built to run on Huawei’s Ascend chips, one day after the Kratsios memo surfaced. Chinese rivals Zhipu AI and MiniMax fell 9% and 7% respectively on the V4 launch day, reading the release as a competitive threat. DeepSeek has denied allegations of illegal use of synthetic training data, insisting its data is collected through organic web searches. The Chinese Embassy in Washington called the White House accusations “baseless” and said Beijing “attaches great importance to the protection of intellectual property rights.” The Foreign Ministry urged Washington to “abandon biases.” As crypto.news documented, US AI companies including Nvidia, Microsoft, and Meta have been navigating compounding geopolitical exposure to China throughout 2026, with the IRGC designation of American tech companies as targets earlier in April adding a second layer of pressure on top of the theft accusations.

What the Administration Has Said It Will Do

The memo says the administration will share intelligence about distillation campaigns with American AI companies and “explore a range of measures to hold foreign actors accountable.” Separately, Commerce Secretary Howard Lutnick confirmed on April 23 that no Nvidia advanced AI chip shipments had actually cleared through to China despite the conditional January approval, adding a chip supply dimension to a dispute already being fought at the model and data levels. The question of whether the Trump-Xi summit scheduled for the coming days will address AI intellectual property directly is being watched closely by industry executives, given that both the distillation accusations and the chip supply dispute are now active points of contention between the two governments. As crypto.news tracked, Nvidia’s revenue has already been materially affected by the chip export restriction tightening that escalated through 2025 and into 2026, making a diplomatic resolution on AI trade highly consequential for the US semiconductor industry.

The White House memo did not name specific Chinese companies or individuals, and it did not announce any immediate sanctions, penalties, or executive actions against the entities accused of the distillation campaigns.

Google‑parent Alphabet will commit $10 billion in cash to Anthropic now and up to $30 billion more over time, in a package that Bloomberg reports could reach $40 billion in total if the AI lab hits aggressive performance and usage milestones. Anthropic said the initial tranche comes at a $350 billion valuation — the same mark as its February round — cementing its status as one of the most richly valued startups on earth.

This is not just an equity trade; it is an infrastructure lock‑in. Google has already committed to providing Anthropic with access to as many as one million Tensor Processing Units under a cloud agreement “worth tens of billions,” a build‑out expected to add more than a gigawatt of AI computing capacity by 2026. SiliconANGLE notes that at its Next conference, Google pitched itself as the platform for “agentic AI” — autonomous systems that will run across finance, compliance, and trading — and the Anthropic deal is the financial spine of that strategy.

In Q1 2026 alone, global VC funding hit a record $297 billion, and roughly 81% of it went to AI companies, with just four names — OpenAI, Anthropic, xAI, and Waymo — pulling in about $188 billion between them. If you raised anything that was not AI‑adjacent, you were fighting over the remaining 19 cents of every venture dollar. Crypto founders feel that squeeze directly: the marginal dollar that might have gone into a new L2, DEX, or stablecoin protocol is instead being spent on GPU farms and frontier labs like Anthropic.

For crypto markets, this cuts both ways. On one side, capital and talent concentration around labs like Anthropic and platforms like Google Cloud makes it harder for pure‑play crypto projects to raise at scale, especially outside real‑world asset, stablecoin, or tokenization narratives. On the other, the future of on‑chain finance is going to be run increasingly by AI agents — smart order routers, risk engines, compliance bots — that sit on exactly the infrastructure this $40 billion package is building.

If Google and Anthropic succeed in turning agentic AI into a reliable, commodity service, the winners in crypto will be the protocols that plug into that compute: on‑chain derivatives venues using AI market makers, DeFi protocols with real‑time AI risk controls, tokenization stacks that use models to price and monitor collateral. In that sense, the biggest “crypto” deal of the year might not be a token round at all, but a cloud‑plus‑AI mega‑investment that decides who owns the rails every serious crypto protocol will eventually run on.

Chinese AI startup DeepSeek released a preview of its V4 model on April 24, explicitly optimized for Huawei’s Ascend chip platform, on the same day the White House accused China of running industrial-scale campaigns to copy and steal American frontier AI systems.

DeepSeek released preview versions of its V4-Pro and V4-Flash models on April 24, both explicitly designed to run on Huawei’s Ascend AI chip platform. Huawei confirmed that its full Ascend supernode product line now supports the DeepSeek V4 series. The launch came one day after a White House memo accused China of coordinated, industrial-scale intellectual property theft from American AI laboratories, with DeepSeek cited repeatedly in prior US accusations as a company that distilled proprietary OpenAI and Anthropic models.

DeepSeek V4 Huawei Chips Launch Inverts the Nvidia Dependency Narrative

DeepSeek’s earlier V3 model was trained on Nvidia hardware, which drew accusations from Washington that the company breached US export controls to acquire advanced Nvidia chips. The V4 announcement inverts that positioning entirely. Huawei is front and center as both collaborator and primary deployment target, while Nvidia is absent from the technical documentation. V4-Pro operates on a mixture-of-experts architecture with 1.6 trillion total parameters and 49 billion active parameters. V4-Flash uses a smaller 284 billion total parameter configuration with 13 billion active parameters designed for cost efficiency. DeepSeek said V4-Pro outperforms every other open-source model on world knowledge benchmarks and trails only Google’s closed-source Gemini-Pro-3.1. The launch also includes a lower-cost flash variant, the same two-tier pricing pattern DeepSeek used to undercut Western labs with V3. As crypto.news reported, Nvidia CEO Jensen Huang had warned the week prior that China possesses the infrastructure and hardware necessary to rival the capabilities of US frontier AI models.

The White House Memo That Landed the Day Before

Michael Kratsios, director of the White House Office of Science and Technology Policy, circulated a memo on April 23 accusing Chinese entities of running “industrial-scale campaigns to distil US frontier AI systems” using tens of thousands of proxy accounts to evade detection and jailbreaking techniques to expose proprietary model information. Distillation is the process of training smaller AI models on the outputs of larger ones, a method that allows a lab to approximate frontier capability without paying the full training cost. The memo stated the administration will share information with US AI companies about distillation campaigns and explore accountability measures against foreign actors. As crypto.news documented, the US AI industry has been navigating compounding geopolitical pressures in 2026, with Chainlink and other AI-adjacent crypto assets already showing sensitivity to US-China tech tensions earlier in April.

What DeepSeek V4 on Huawei Means for the US Chip Export Strategy

The V4 Huawei pivot is a direct strategic response to US export controls. By demonstrating that a model approaching frontier performance can now be trained and served on Huawei Ascend silicon, DeepSeek and China are challenging the assumption that semiconductor restrictions can meaningfully slow Chinese AI development. Commerce Secretary Howard Lutnick confirmed separately on April 23 that no Nvidia advanced AI chip shipments had actually gone through to China despite the January conditional approval, a disclosure that deepens the question of what hardware DeepSeek actually used for V4 training. DeepSeek has not confirmed this, and detailed technical specifications for V4 have not been independently verified. As crypto.news tracked, Nvidia disclosed $5.5 billion in expected charges in April 2025 when the US government required export licenses for H20 chips sold to China, a policy backdrop that made Huawei’s role in V4 commercially inevitable for DeepSeek’s long-term supply chain.

The Chinese Embassy in Washington called the White House distillation accusations “baseless,” and Beijing’s foreign ministry urged the US to “abandon biases” and said more scientific exchange, not less, was needed.

The Ethereum Foundation finalized terms of a 10,000 ETH sale at an average price of $2,387 per token via over-the-counter (OTC) trading on April 24, 2026. BitMNR served as the OTC counterparty for the transaction, which valued the sale at approximately $23.87 million based on the stated average price.

The sale represents a significant asset movement by the Ethereum Foundation, which maintains treasury holdings to fund ecosystem development and research. OTC trades of this scale typically indicate institutional-grade transactions executed outside public order books to minimize market impact.

Sources: Ethereum Foundation | Degenerate News

This article was generated automatically by The Defiant’s AI news system from publicly available sources.

The U.S. Treasury Department said Friday that a $344 million cryptocurrency freeze is part of its latest effort to disrupt financial networks tied to Iran.

Treasury Secretary Scott Bessent said in an X post that the Treasury’s Office of Foreign Assets Control (OFAC) is sanctioning multiple crypto wallets linked to Iran, resulting in the freeze of $344 million in cryptocurrency.

“We will follow the money that Tehran is desperately attempting to move outside of the country and target all financial lifelines tied to the regime,” Bessent said, adding the effort is part of a broader campaign dubbed “Economic Fury.”

The post follows action taken Thursday by stablecoin issuer Tether blacklisting two blockchain addresses on Tron holding $344 million in USDT altogether.

The company did not return a request for comment.

A U.S. official told CoinDesk that the sanctioned wallets showed material links to the Iranian regime, including transactions with Iranian exchanges and routing through intermediary addresses connected to wallets associated with the Central Bank of Iran. According to the Treasury Department, Iran’s central bank has been leaning into digital assets to try to mask its cross-border transactions.

Authorities said Iran has increasingly turned to crypto to bypass restrictions, using more complex transaction patterns to obscure its involvement in cross-border payments and support trade flows under sanctions pressure.

Treasury’s OFAC is trying to turn up the pressure by moving aggressively against both the traditional front companies and the use of digital assets, the official said. Meanwhile, it sanctioned Hengli Petrochemical (Dalian) Refinery Co. on Friday, accusing the China-based independent refineries of playing a major role in Iran’s oil economy.

The U.S. agency said it continues to work with blockchain analytics firms and maintains coordination with financial institutions, including crypto exchanges, as it tracks illicit flows tied to sanctioned entities.

Key Takeaways

• MaxLinear shares skyrocketed approximately 80% to $61.52 during Friday’s trading session
• First quarter adjusted earnings per share of $0.22 exceeded Wall Street’s $0.18 projection
• Quarterly revenue of $137.2 million marked a 43% increase from the prior year period
• Infrastructure division revenue jumped 136% year-over-year, becoming the dominant business segment
• Second quarter revenue forecast of $160M–$170M significantly surpassed analyst expectations of $137.1M

MaxLinear experienced an extraordinary trading session on Friday that few semiconductor companies ever witness. Shares rocketed approximately 80% higher to close at $61.52, marking what appears to be the company’s most significant one-day percentage increase on record and representing its strongest closing price since 2022.

MaxLinear, Inc., MXL

The explosive rally was triggered by first quarter financial results that exceeded analyst projections on every major metric. The company posted adjusted earnings of $0.22 per share, comfortably surpassing the Street’s $0.18 estimate. Total revenue climbed to $137.2 million, representing a robust 43% jump compared to the year-ago quarter.

However, the metric that truly captured investor enthusiasm was the infrastructure division’s explosive 136% year-over-year revenue expansion. This segment, primarily fueled by optical data-center platform sales, has now emerged as MaxLinear’s dominant revenue contributor — surpassing the broadband business for the first time in company history.

During Thursday’s earnings conference call, CEO Kishore Seendripu highlighted that the company’s Keystone optical transceiver platform is “ramping at multiple major high-scale customers across both the U.S. and Asia.”

Forward Outlook Crushes Consensus

Looking ahead to the second quarter, MaxLinear issued revenue guidance ranging from $160 million to $170 million. This projection substantially exceeds the $137.1 million consensus estimate that analysts had been modeling. Additionally, management boosted its full-year 2026 optical data-center revenue projection by $40 million, now anticipating between $150 million and $170 million for the year.

Needham analyst N. Quinn Bolton suggested that the company’s infrastructure transformation will likely justify premium valuation multiples. “We expect this gap to widen over the next few years on robust data center demand,” Bolton noted in a research report.

Needham elevated its rating on MXL to Buy while establishing a $60 price objective, calculated using 25 times the firm’s 2028 non-GAAP earnings estimate of $2.35 per share. Susquehanna increased its target price to $45 from $30, though maintained its Neutral stance. Stifel confirmed its Buy recommendation while raising its target from $34 to $49.

Susquehanna’s Christopher Rolland characterized the quarterly update as “the constructive update that many had been hoping for.”

Current Market Position

Friday’s dramatic surge propels MXL to gains of roughly 250% year-to-date and approximately 500% over the trailing twelve-month period. The stock is now hovering close to its 52-week peak.

Based on current market pricing, MXL is valued at roughly 43.6 times forward twelve-month earnings projections. While this represents double the valuation multiple from twelve months earlier, it remains below larger industry competitors such as Lumentum and Ciena.

According to InvestingPro analytics, ten Wall Street analysts have recently increased their earnings forecasts for upcoming reporting periods. The current consensus projects fiscal 2026 earnings of $0.91 per share — representing a dramatic reversal from the $1.58 per share loss recorded over the previous twelve months.

Stifel observed that first quarter revenue exceeded its internal forecast by 1.6%, reinforcing the firm’s confidence in maintaining its Buy recommendation.