DD
MM
YYYY

PAGES

DD
MM
YYYY

spot_img

PAGES

Home Blog Page 129

Anthropic Alleges Massive Distillation Campaign by Chinese AI Labs, Escalating Fight Over Chips and Safeguards

0

Anthropic’s claim that 16 million Claude interactions were siphoned through 24,000 fake accounts reframes the AI race as a battle over inference access, safeguards, and chip supply — not just model training.


U.S. artificial intelligence firm Anthropic has accused three Chinese AI developers, DeepSeek, Moonshot AI, and MiniMax, of orchestrating what it describes as a coordinated, large-scale “distillation” campaign to extract capabilities from its Claude AI system.

Anthropic said the three labs created more than 24,000 fraudulent accounts that generated over 16 million interactions with Claude. The queries, it alleged, were designed to systematically replicate some of Claude’s most advanced features, including agentic reasoning, tool use, and coding — areas considered differentiators among frontier AI systems.

The accusations land amid heightened geopolitical tension over artificial intelligence, particularly as Washington reassesses export controls on advanced semiconductors to China and as Chinese AI labs close the performance gap with U.S. counterparts.

Distillation as a competitive shortcut

Distillation is a common technique in machine learning in which a large, high-performing model acts as a “teacher” to train a smaller “student” model. Within a single company, it is used to compress models for lower-cost deployment while retaining much of their capability.

Across companies, however, the method becomes controversial. By querying a rival’s model at scale and using the responses as training data, a competitor can approximate performance without replicating the comprehensive research, compute expenditure, or alignment work that went into the original system.

Anthropic said the alleged campaigns varied in focus and scale. It tracked more than 150,000 exchanges linked to DeepSeek that appeared aimed at strengthening foundational reasoning and alignment, including workarounds for policy-sensitive prompts. Moonshot AI allegedly generated more than 3.4 million exchanges focused on agentic reasoning, tool integration, coding, data analysis, and computer vision. MiniMax accounted for roughly 13 million exchanges targeting agentic coding and orchestration, with Anthropic claiming it observed traffic being redirected toward the latest Claude release shortly after launch.

The scale matters since sixteen million interactions represent not casual usage but what Anthropic characterizes as industrialized extraction.

DeepSeek, in particular, has drawn scrutiny since releasing its open-source R1 reasoning model last year, which analysts said approached the performance of leading U.S. frontier labs at a fraction of the cost. The company is reportedly preparing DeepSeek V4, a new model said to outperform both Claude and OpenAI’s ChatGPT in certain coding benchmarks. Earlier this month, OpenAI also accused DeepSeek in a memo to U.S. lawmakers of using distillation techniques to mimic its systems.

Export controls and compute leverage

The dispute is unfolding alongside debate over access to high-end chips. Last month, the Trump administration allowed U.S. firms, including Nvidia, to export advanced AI processors such as the H200 to China, loosening earlier restrictions.

Anthropic linked the alleged distillation campaigns to computing power. “The scale of extraction… requires access to advanced chips,” the company said in a blog post.

It argued that export controls serve a dual function: limiting direct model training and constraining the compute needed for high-volume distillation.

This framing shifts the chip debate. Historically, export controls focused on preventing Chinese firms from training large frontier models from scratch. Anthropic’s claim suggests that even if direct training is restricted, access to sufficient inference compute could enable large-scale replication via API querying.

Policy analysts say this complicates enforcement. Distillation occurs through legitimate product interfaces — paid or public APIs — rather than through overt hacking. That creates a grey zone between normal usage and systematic extraction.

National security dimension

Anthropic also framed the issue as one of security. The company said U.S. developers build safeguards into frontier systems to prevent misuse in areas such as bioweapons design or malicious cyber operations.

“Models built through illicit distillation are unlikely to retain those safeguards,” Anthropic wrote, warning that dangerous capabilities could proliferate if protections are stripped out during replication.

It pointed to the possibility of authoritarian governments deploying advanced AI for offensive cyber operations, disinformation campaigns, and mass surveillance — risks that increase if such systems are open-sourced without embedded safety layers.

Dmitri Alperovitch, chairman of the Silverado Policy Accelerator and co-founder of CrowdStrike, told TechCrunch the allegations were unsurprising.

“It’s been clear for a while now that part of the reason for the rapid progress of Chinese AI models has been theft via distillation of U.S. frontier models. Now we know this for a fact,” he said. “This should give us even more compelling reasons to refuse to sell any AI chips to any of these [companies], which would only advantage them further.”

Anthropic said it will continue investing in defensive measures to make distillation harder to execute and easier to detect, while calling for “a coordinated response across the AI industry, cloud providers, and policymakers.”

Such coordination could include tighter API rate limits, improved anomaly detection, contractual enforcement mechanisms, and shared threat intelligence among AI labs. Cloud providers — which host the infrastructure underpinning both U.S. and Chinese AI workloads — may also face pressure to monitor and flag high-volume extraction patterns.

The broader stakes extend beyond one company. If frontier AI capabilities can be replicated rapidly through sustained querying, the competitive moat built on research expenditure and chip access narrows. In that scenario, advantage may hinge less on breakthrough architecture and more on distribution control, access management, and compute governance.

At the same time, aggressive restrictions carry trade-offs. Limiting chip exports could affect U.S. semiconductor revenues and accelerate domestic Chinese chip development. Restricting API access could constrain legitimate global customers and developers.

Anthropic’s allegations therefore crystallize a central tension in the global AI race: openness versus control. The tools that make AI widely usable, APIs, cloud access, and scalable inference, also create vectors for replication. As Chinese labs close the performance gap with U.S. peers, the contest increasingly revolves not just around building the most advanced model, but protecting it.

Dangote Refinery Set for Public Listing in Four Months as Retail Investors Promised Dollar Dividend Option

0

Aliko Dangote says ordinary Nigerians will be able to buy shares in the $20 billion Dangote Refinery within five months, with the option to receive dividends in either naira or U.S. dollars.


Chairman of Dangote Group, Aliko Dangote, has confirmed that ordinary Nigerians will be able to acquire shares in the Dangote Refinery within the next four to five months, marking a significant step toward opening up ownership of one of Africa’s largest industrial projects.

Dangote made the disclosure on Saturday, February 21, 2026, during a media tour of the refinery alongside the Group Chief Executive Officer of Nigerian National Petroleum Company Limited (NNPC), Bayo Ojulari, and other senior executives of the state-owned oil firm.

He described the visit as symbolic, emphasizing that NNPC is not merely a collaborator but an equity participant in the refinery.

“Today is really our best day ever. At least he (Ojulari) is not just a guest, he is a shareholder and you know NNPC invested in us when we ourselves were not even sure that the refinery will be successful,” Dangote said.

According to him, NNPC holds a 7.25% equity stake in the refinery on behalf of Nigerians. “They are holding 7.25% of the shares that we have here, which is more than the shares that Elon Musk has in Tesla, and they are holding that on behalf of Nigerians,” he said, adding that individual Nigerians will soon have the opportunity to directly participate in ownership.

“But individually Nigerians too will have an opportunity in the next maximum four or five months they will actually be able to buy their shares.”

Structure of the Planned Offer

The refinery, valued at about $20 billion, is expected to list roughly 10% of its equity on the Nigerian Exchange Limited (NGX) in 2026. Dangote had earlier disclosed that discussions were ongoing with the Securities and Exchange Commission of Nigeria and the NGX to finalize the framework for the initial public offering (IPO).

A distinctive feature of the proposed listing is the dual-currency dividend structure. Dangote said investors would be allowed to receive dividends either in naira or in U.S. dollars, citing the refinery’s foreign currency earnings profile.

“People will have a choice either to get their dividends in naira or to get their dividends in dollars because we earn in dollars,” he stated.

Projected export earnings of about $6.4 billion, largely from petrochemical products such as polypropylene and fertilizer, are expected to underpin this dollar dividend model. That revenue stream provides the refinery with a steady flow of hard currency, which in turn supports the feasibility of foreign currency payouts to shareholders.

Capital Market and Currency Implications

If executed within the stated timeline, the IPO would rank among the most consequential listings in Nigeria’s capital market history. A company of this scale entering the exchange is likely to boost market capitalization on the NGX and attract renewed domestic and international investor interest.

The structure also introduces a new dynamic into Nigeria’s equity market. Dollar-denominated dividends are uncommon on the local exchange, and the option could appeal to investors seeking protection against exchange rate fluctuations. For retail investors who typically have limited access to dollar-based assets, such a structure may represent a rare opportunity to hold an instrument tied to foreign currency earnings.

Beyond liquidity gains, the listing could deepen retail participation in equities at a time when household savings are often concentrated in fixed-income instruments. By offering exposure to an asset with a strong export orientation and integrated value chains, the refinery may broaden the investment landscape for individual Nigerians.

The Dangote Refinery occupies a central position in Nigeria’s downstream petroleum industry. Designed to reduce reliance on imported refined products, it has implications for fuel supply stability, trade balances, and industrial feedstock availability.

Its petrochemical outputs, including polypropylene, are critical inputs for manufacturing sectors such as plastics, packaging, and consumer goods. Fertilizer production also links directly to agricultural productivity, which has wider food security and export implications. The project seeks to capture more value domestically rather than exporting crude and importing finished products by combining refining and petrochemicals in a single integrated complex.

Opening up ownership to the public embeds this industrial infrastructure more directly into Nigeria’s financial system. It creates a bridge between capital markets and large-scale manufacturing, potentially setting a precedent for how future infrastructure and industrial projects are financed.

Relationship with NNPC

Dangote expressed optimism about collaboration with the new NNPC leadership under Ojulari, describing the relationship as forward-looking.

“I think the sky is the limit and we would cooperate and also make sure we work together to make Nigerians proud,” he said.

NNPC’s 7.25% stake positions the state oil company as both a shareholder and a strategic partner. That arrangement may facilitate alignment on crude supply agreements, product distribution channels, and broader policy coordination within the energy sector.

Following the announcement of the listing, Nigerians have excitement and readiness to bet on the refinery. Dangote indicated that priority would be given to Nigerian retail investors to ensure broad-based participation.

Amazon Unveils $12bn Louisiana Data Center Expansion Amid Intensifying AI Infrastructure Race

3

Amazon plans to invest $12 billion in new AI-focused data center campuses in Louisiana as part of a broader $200 billion capital expenditure push.


Amazon on Monday announced plans to invest $12 billion in new data center campuses in Louisiana, reinforcing its aggressive expansion of artificial intelligence and cloud infrastructure even as investors weigh the financial implications of surging capital expenditures.

The facilities will be located in Caddo and Bossier Parishes in northwestern Louisiana. Amazon said it expects to create 540 full-time roles tied directly to the data centers and support approximately 1,700 additional jobs connected to the campuses, including electricians, HVAC technicians, construction workers, and security specialists. Beyond direct employment, such projects often generate secondary economic effects through local suppliers, logistics providers, and service industries.

The investment forms part of Amazon’s broader capital spending strategy. Earlier this month, the company projected roughly $200 billion in capital expenditures for the year — a figure that outpaces other hyperscale cloud operators and underscores the scale of infrastructure required to support next-generation AI workloads. Combined, major technology firms have forecast nearly $700 billion in 2026 expenditures as the race to build AI capacity accelerates.

Amazon’s announcement follows a volatile period for its stock. Shares declined for nine consecutive trading sessions after the company’s Feb. 5 earnings report, wiping out more than $450 billion in market value during that stretch. Investors have questioned whether the pace of capital deployment may weigh on near-term profitability, particularly as AI monetization models are still evolving.

A substantial portion of Amazon’s spending is directed toward AI-related assets, including advanced data centers, proprietary chips, networking equipment, and specialized computing clusters. These investments are largely channeled through Amazon Web Services (AWS), the company’s cloud division and primary profit engine. AI services, including model training and inference, are increasingly becoming core growth drivers for AWS, but they require enormous upfront infrastructure commitments.

Amazon’s expansion mirrors similar moves by competitors such as Microsoft and Google, both of which have committed hundreds of billions of dollars to expanding their global data center footprints. Meta has also selected Louisiana for a major project — the $27 billion Hyperion data center joint venture with Blue Owl Capital.

The clustering of hyperscale facilities in Louisiana signals the state’s growing appeal as an infrastructure hub. Factors likely include land availability, grid capacity, regulatory conditions, and economic development incentives. For technology firms, site selection balances proximity to energy sources, transmission infrastructure, fiber connectivity, and tax considerations.

AI models require vast computational power not only for initial training but also for ongoing inference — the process of generating responses or predictions in real time. As enterprises integrate AI into customer service, analytics, healthcare diagnostics, and industrial automation, demand for high-performance computing continues to rise. Hyperscalers are effectively building the backbone for this digital transformation.

Energy, Water, and Community Considerations

The rapid buildout of data centers nationwide has prompted environmental and community scrutiny. Large-scale facilities can consume significant amounts of electricity and water, raising concerns about grid strain and local resource allocation.

Last year, Microsoft withdrew from a planned site in rural Wisconsin after residents raised environmental and financial concerns, illustrating the political and regulatory sensitivity surrounding such projects.

Amazon said it worked with the local utility, Southwestern Electric Power Company, to ensure that it would bear the full cost of energy infrastructure tied to the campuses. The company stated it will pay “100% of the costs” associated with new and upgraded energy infrastructure.

To mitigate electricity demand, Amazon plans to incorporate natural air cooling when feasible, reducing reliance on energy-intensive mechanical cooling systems. The company also said the Louisiana campuses will use only surplus water from the surrounding area, asserting there will be no strain on local water supplies. In addition, Amazon plans to invest up to $400 million in public water infrastructure to support the sites.

Amazon is partnering with data center developer Stack Infrastructure on the project, leveraging the firm’s expertise in constructing and operating hyperscale facilities.

The Louisiana investment points to the structural shift underway in the technology sector. AI has moved from an experimental capability to a foundational computing paradigm, requiring sustained capital outlays comparable to utilities or transportation networks. Data centers are increasingly viewed as critical infrastructure.

For Amazon, the long-term thesis rests on AI-driven revenue growth offsetting the near-term pressure on free cash flow. The company is effectively wagering that enterprises’ migration toward AI-powered services will generate durable demand across cloud storage, compute, machine learning platforms, and related services.

At the state level, Louisiana stands to gain from increased tax revenues, workforce development opportunities, and enhanced infrastructure. However, sustained oversight of environmental impact and grid resilience will likely remain central to public discourse.

Nigeria Targets 5% of GDP for Industrial Financing Under Sweeping 2025 Policy Drive

0

By earmarking up to 5% of GDP for industrial financing, Abuja is signaling one of the most aggressive state-backed industrial funding commitments in Nigeria’s recent economic history.


The Federal Government has unveiled plans to allocate up to 5% of Nigeria’s Gross Domestic Product to industrial financing under the Nigeria Industrial Policy (NIP) 2025, marking a decisive shift toward large-scale production, export expansion and structured job creation.

Contained in the framework released by the Federal Ministry of Industry, Trade and Investment, the policy lays out a coordinated strategy to reposition the economy toward mass manufacturing, deeper value addition, and reduced import dependence.

At the heart of the plan is a strengthened development finance architecture anchored on recapitalizing the Bank of Industry and scaling intervention funds in collaboration with the Central Bank of Nigeria.

“We recognize that no policy succeeds without financing,” the document stated. “By setting aside up to 5% of GDP for industrial financing and leveraging public–private partnerships, this government demonstrates its commitment to matching ambition with resources.”

The 5% allocation is designed to crowd in private capital through structured public–private partnerships while lowering financing constraints that have historically hampered Nigeria’s manufacturing sector.

  • Under the framework:
  • The Bank of Industry is expected to be recapitalized to N3 trillion by 2026.
  • Sector-specific intervention funds are projected to rise to N3 trillion.
  • Credit guarantees for MSMEs will be mainstreamed to de-risk lending.
  • New instruments such as interest-drawback schemes and equity-based financing will be introduced.

This approach represents a shift from fragmented credit schemes toward a consolidated industrial financing ecosystem with clearer institutional responsibilities and performance metrics.

If implemented at scale, the allocation could materially alter capital access conditions for manufacturers, especially in agro-processing, textiles, pharmaceuticals, petrochemicals, and light engineering.

President Bola Tinubu formally unveiled the Nigeria Industrial Policy 2025 last week, directing ministries, departments, and agencies to ensure swift execution.

The policy aligns with Tinubu’s “Renewed Hope” agenda, particularly its emphasis on domestic production, import substitution, and industrial self-sufficiency.

A key feature is the enforcement of a “Nigeria First” procurement stance, prioritizing locally manufactured goods in public spending. The framework also seeks to reduce reliance on imported raw materials by encouraging backward integration and local value chains.

By consolidating fiscal, monetary, trade, and industrial measures into a unified strategy, the government is attempting to address longstanding coordination gaps that have limited the impact of previous industrial programmes.

Target: Manufacturing at 25% of GDP

Nigeria’s manufacturing contribution to GDP has historically remained in the low double digits. The NIP 2025 sets an ambitious target: raising manufacturing’s share to between 20% and 25% of GDP by 2030.

Achieving that would require sustained growth in output, infrastructure upgrades, and improved logistics efficiency, particularly in power supply and transport networks.

Industrial expansion at that scale would also carry employment implications. Manufacturing is typically more labor-intensive than extractive industries, suggesting potential for broad-based job creation, particularly among youth.

Beyond domestic production, the policy underscores export competitiveness as a strategic priority. Nigeria’s export base remains heavily concentrated in crude oil. Industrial diversification is positioned as essential to stabilizing foreign exchange earnings and reducing vulnerability to commodity price shocks.

Strengthening non-oil exports — from processed agricultural goods to finished consumer products — would require improvements in standards compliance, logistics, trade facilitation, and market access.

The policy’s integrated design seeks to align export promotion with financing mechanisms, rather than treating them as separate policy domains.

Implementation Risks and Structural Constraints

While the financing commitment is substantial on paper, execution will determine outcomes.

Key variables include:

  • Fiscal sustainability: Allocating up to 5% of GDP requires disciplined budget management and efficient capital deployment.
  • Governance oversight: Transparent allocation and monitoring of funds will be critical to prevent leakages.
  • Infrastructure readiness: Industrial financing must be matched by reliable electricity, transport corridors, and digital infrastructure.
  • Private sector participation: Public–private partnerships depend on investor confidence and regulatory stability.

Recapitalizing development finance institutions alone will not guarantee productivity gains unless funds are channeled into sectors with clear competitive advantages.

Overall, the Nigeria Industrial Policy 2025 represents one of the most comprehensive industrial blueprints in recent years. By embedding financing targets within a structured implementation framework — complete with timelines and measurable performance indicators — the government is attempting to move from policy declaration to institutional execution.

If the 5% GDP allocation is realized and effectively deployed, it could mark a structural pivot away from oil dependence toward diversified industrial growth.

While the scale of ambition is clear, economists reiterate that the test will lie in translating capital commitments into functioning factories, expanded export lines, and sustainable employment across the country.

U.S. Customs to Halt IEEPA Tariff Collections at Midnight Tuesday

0

U.S. Customs and Border Protection (CBP) announced late Monday that it will cease collecting tariffs imposed under the International Emergency Economic Powers Act (IEEPA) effective 12:01 a.m. EST (0501 GMT) on Tuesday.

This comes more than three days after the U.S. Supreme Court declared those duties illegal in a landmark 6-3 ruling on Friday.

In a message posted to its Cargo Systems Messaging Service (CSMS), CBP stated it would deactivate all tariff codes associated with President Donald Trump’s prior IEEPA-related orders as of Tuesday. The agency provided no explanation for the delay in halting collections despite the immediate legal effect of the Supreme Court decision, nor did it offer details on the process for refunds to importers who paid duties under the now-invalidated regime.

CBP emphasized that the halt applies only to IEEPA tariffs and does not affect other duties imposed by Trump, including those under Section 232 (national security), Section 301 (unfair trade practices), antidumping/countervailing measures, or the new 15% global tariff enacted under Section 122 of the 1974 Trade Act.

“CBP will provide additional guidance to the trade community through CSMS messages as appropriate,” the agency said.

Supreme Court Ruling and Immediate Aftermath

The Supreme Court’s Friday decision invalidated Trump’s use of IEEPA — a 1977 law granting emergency economic powers — to impose broad “reciprocal” tariffs (10–50%) and fentanyl-related duties since February 2025. Chief Justice John Roberts, writing for the majority, held that IEEPA does not authorize unilateral import taxes absent a specific, imminent foreign threat, ruling the president exceeded congressional intent and violated the separation of powers.

The ruling dismantled the legal foundation for tariffs that had generated an estimated $175–$179 billion in revenue since February 2025, according to the Penn-Wharton Budget Model (PWBM). PWBM’s ground-up model, using Census Bureau import data across 11,000 product categories and 233 countries, calculated roughly $500 million in daily IEEPA-based collections, leading to the cumulative $179 billion figure.

A cross-check with historical CBP assessment data as a share of total Treasury customs receipts yielded a similar $175–$176 billion range. CBP’s last published IEEPA assessment (December 14, 2025) stood at $133.5 billion, with net collections typically lower after adjustments, protests, and refunds.

Trump reacted swiftly to the ruling. On Saturday, he imposed a temporary 10% global levy under Section 122 of the 1974 Trade Act, then raised it to 15% — the maximum rate allowable for 150 days without congressional approval.

“Effective immediately,” he declared on Truth Social, framing the action as necessary to maintain leverage despite the court decision.

U.S. Trade Representative Jamieson Greer defended the continuity of existing trade deals, insisting the ruling affected only IEEPA-based tariffs.

“Our partners have been responsive and engaged in good-faith negotiations and agreements despite the pending litigation, and we are confident that all trade agreements negotiated by President Trump will remain in effect,” Greer said Sunday on CBS’ Face the Nation.

Refund Process and Fiscal Implications

Importers who paid IEEPA duties since February 2025 are now eligible to seek refunds from CBP. The process will involve filing protests or refund claims, subject to administrative review, potentially stretching over months or years. A $175–$179 billion refund would represent a massive one-time cash outflow for the Treasury — exceeding the combined fiscal 2025 outlays of the Department of Transportation ($127.6 billion) and Department of Justice ($44.9 billion).

Treasury Secretary Scott Bessent told Reuters in January that the Treasury could “easily cover” any repayments through planned cash balances ($850 billion at end-March 2026, $900 billion at end-June). The administration has signaled contingency plans to restore tariffs under alternative authorities (Section 232, Section 301) if needed, though these may face their own legal and procedural hurdles.

The refund potential could provide unintended stimulus to businesses and consumers, though administrative bottlenecks at CBP may delay payouts. Importers in affected sectors (steel, aluminum, autos, consumer goods) stand to recover substantial duties, potentially improving cash flow and margins.

The ruling significantly curtails executive authority to impose broad tariffs under emergency powers, reinforcing congressional primacy over trade policy. It may force the administration to rely more heavily on Section 232, Section 301, and antidumping/countervailing mechanisms — processes requiring more evidentiary findings and procedural steps.

For trading partners, the decision offers temporary relief from broad emergency tariffs while signaling that targeted, evidence-based actions under other laws remain likely. The EU, U.K., Japan, and South Korea — early deal-makers with preferential rates — now face uncertainty over whether those concessions survive the transition to new tariff frameworks.

The administration’s pivot to Section 301 probes (covering pharmaceuticals, industrial overcapacity, forced labor, digital services taxes, and more) indicates a shift toward a more targeted, legally durable approach — albeit one that could still provoke retaliation from affected countries.

Trade partners welcomed the ruling. China’s Ministry of Commerce called it “a step toward fairer trade,” while the EU expressed hope for reduced transatlantic tensions. However, if the U.S. reimposes duties under new authorities, retaliatory measures could escalate — potentially reigniting global trade frictions.

For importers, the ruling unlocks a path to refunds but introduces short-term uncertainty as CBP processes claims. Legal experts predict a surge in filings, with class actions possible for smaller importers.