Economic Archives | Page 447 of 709 | Digital Watch Observatory

Meta has developed an AI chip to cut reliance on Nvidia, Reuters reports

Meta, the owner of Facebook, Instagram, and WhatsApp, is testing its first in-house chip designed for training AI systems, sources told Reuters.

The social media giant has started a limited rollout of the chip, planning to scale up production if testing delivers positive results. The move represents a crucial step in Meta’s strategy to lessen dependence on external suppliers like Nvidia and lower substantial infrastructure costs.

The company has projected expenses between $114 billion and $119 billion for 2025, with up to $65 billion dedicated to AI infrastructure.

The chip, part of Meta’s Meta Training and Inference Accelerator (MTIA) series, is a dedicated AI accelerator, meaning it is specifically designed for AI tasks rather than general processing. This could make it more power-efficient than traditional GPUs.

Meta is collaborating with Taiwan-based chip manufacturer TSMC to produce the new hardware. The test phase follows Meta’s first ‘tape-out’ of the chip, a crucial milestone in silicon development where an initial design is sent to a chip factory.

However, this process is costly and time-consuming, with no guarantee of success, and any failure would require repeating the tape-out step.

Meta has previously faced setbacks in its custom chip development, including scrapping an earlier version of an inference chip after poor test results. However, the company has since used another MTIA chip for AI-powered recommendations on Facebook and Instagram.

The new training chip aims to first enhance recommendation systems before expanding to generative AI applications like the chatbot Meta AI.

Meta executives hope to implement their own chips for AI training by 2026, although the company continues to be one of Nvidia’s biggest customers, investing heavily in GPUs for its AI operations.

The development comes as AI researchers increasingly question whether scaling up large language models by adding more computing power will continue to drive progress. The recent emergence of more efficient AI models, such as those from Chinese startup DeepSeek, has intensified these debates.

While Nvidia remains a dominant force in AI hardware, fluctuating investor confidence and broader market concerns have caused turbulence in the company’s stock value.

For more information on these topics, visit diplomacy.edu.

Google revises AI team’s mission statement, removing equity focus

Google has quietly updated the webpage for its Responsible AI and Human-Centred Technology team, removing references to diversity and equity

Terms such as ‘marginalised communities’ and ‘underrepresented groups’ have been replaced with more neutral language. The changes were first spotted by watchdog group The Midas Project, which previously reported similar edits to Google’s Startups Founders Fund page.

The company’s move comes amid a broader rollback of diversity, equity, and inclusion (DEI) initiatives across the tech industry. Google announced in February that it would end its diversity hiring targets and reassess its DEI programmes.

Other companies, including Amazon and Meta, have also scaled back diversity policies in response to legal and political pressures from the Trump administration, which has criticised such initiatives.

Federal contracts could be influencing these decisions, as many of the affected companies, including Google, work closely with United States agencies.

While some firms, such as OpenAI, have removed diversity language from hiring pages, Apple recently rejected a shareholder proposal to eliminate its DEI programmes. The changes suggest a shifting landscape for corporate diversity efforts in the US tech sector.

For more information on these topics, visit diplomacy.edu.

US drops AI investment proposal against Google

The US Department of Justice (DOJ) has decided to drop its earlier proposal to force Alphabet, Google’s parent company, to sell its investments in AI companies, including its stake in Anthropic, a rival to OpenAI.

The proposal was originally included in a wider initiative to boost competition in the online search market. The DOJ now argues that restricting Google’s AI investments might lead to unintended consequences in the rapidly changing AI sector.

While this move represents a shift in the government’s approach, the DOJ and 38 state attorneys general are continuing their antitrust case against Google. They argue that Google holds an illegal monopoly in the search market and is distorting competition.

The government’s case includes demands for Google to divest its Chrome browser and implement other measures to foster competition.

Google has strongly opposed these efforts, stating that they would harm consumers, the economy, and national security. The company is also planning to appeal the proposals.

As part of the ongoing scrutiny, the DOJ’s latest proposal mandates that Google notify the government of any future investments in generative AI, a move intended to curb further concentration of power in the sector.

This case is part of a broader wave of antitrust scrutiny facing major tech companies like Google, Apple, and Meta, as US regulators seek to rein in the market dominance of Big Tech.

For more information on these topics, visit diplomacy.edu.

Authors challenge Meta’s use of their books in AI training

A lawsuit filed by authors Richard Kadrey, Sarah Silverman, and Ta-Nehisi Coates against Meta has taken a significant step forward as a federal judge has ruled that the case will continue.

The authors allege that Meta used their books to train its Llama AI models without consent, violating their intellectual property rights.

They further claim that Meta intentionally removed copyright management information (CMI) from the works to conceal the alleged infringement.

Meta, however, defends its actions, arguing that the training of AI models qualifies as fair use and that the authors lack standing to sue.

Despite this, the judge allowed the lawsuit to move ahead, acknowledging that the authors’ claims suggest concrete injury, specifically regarding the removal of CMI to hide the use of copyrighted works.

While the lawsuit touches on several legal points, the judge dismissed claims related to the California Comprehensive Computer Data Access and Fraud Act, stating that there was no evidence of Meta accessing the authors’ computers or servers.

Meta’s defence team has continued to assert that the AI training practices were legally sound, though the ongoing case will likely provide more insight into the company’s stance on copyright.

The ruling adds to the growing list of copyright-related lawsuits involving AI models, including one filed by The New York Times against OpenAI. As the debate around AI and intellectual property rights intensifies, this case could set important precedents.

For more information on these topics, visit diplomacy.edu.

US takes aim at Google’s web browser amid legal fight

The US Department of Justice is maintaining its push for Google to sell its Chrome web browser as part of an ongoing antitrust case.

A recent court filing reaffirmed the department’s stance, arguing that Google’s dominance in online search has created an unfair advantage.

While earlier proposals called for divesting all AI investments, the DOJ is now only requiring prior notification of future deals.

Legal action follows a ruling by Judge Amit P. Mehta, who found that Google illegally maintained its search monopoly. The United States DOJ also seeks to prohibit payments to distribution partners, a key practice that has helped Google secure its search dominance.

Google has criticised the proposals, claiming they would harm consumers and national security, and is preparing an appeal against the ruling.

Arguments from both sides will be heard in court this April. The case is expected to have far-reaching implications for competition in the tech industry, with potential changes to how Google operates its search and browser business.

For more information on these topics, visit diplomacy.edu.

Saylor proposes a Bitcoin strategy to unlock trillions for the US economy

Michael Saylor has unveiled an ambitious cryptocurrency strategy at the White House Digital Assets Summit, arguing that the US could unlock up to $100 trillion in economic value over the next decade.

He called for clear regulations, the removal of barriers to innovation, and a strategic acquisition of bitcoin to strengthen the financial system.

Saylor proposed a four-part framework for digital assets, categorising them into digital tokens, securities, currencies, and commodities like Bitcoin.

His plan aims to reduce regulatory uncertainty, ensure the US dollar remains dominant in global trade, and integrate cryptocurrencies into mainstream finance.

He also urged the government to support major banks in holding and trading Bitcoin while ending what he described as ‘hostile’ tax policies towards the industry.

The summit, seen as a shift towards a more crypto-friendly stance under Trump’s administration, gathered industry leaders from Coinbase, Ripple, Kraken, and others.

A key part of Saylor’s vision includes a US Bitcoin reserve, with the country acquiring 5% to 25% of the total Bitcoin supply by 2035. He projected this could generate up to $81 trillion by 2045, offering a long-term solution to national debt challenges.

For more information on these topics, visit diplomacy.edu.

Tusk warns against arrogance after US-Poland social media clash

Poland’s Prime Minister, Donald Tusk, has urged allies to show respect and avoid arrogance in a recent post on X, following a heated social media exchange between Polish and US officials. The remarks came after a disagreement over the role of Starlink satellites in Ukraine’s war effort. Radosław Sikorski, Poland‘s foreign minister, had suggested Ukraine may need an alternative to Starlink if its reliability becomes an issue. Poland funds the satellite service for Ukraine, which is crucial for military communications.

The dispute escalated when Marco Rubio, the US Secretary of State, accused Sikorski of being ungrateful, stating that ‘no one has made any threats about cutting Ukraine off from Starlink.’ Rubio emphasised the importance of Starlink in Ukraine’s success, saying the war could have been lost without it. Sikorski responded by thanking Rubio for reaffirming the collaboration between the US and Poland in providing the service.

The controversy deepened when Elon Musk, the founder of SpaceX, which operates Starlink, labelled Sikorski a “small man” and told him to ‘be quiet’ after the suggestion that Poland may seek alternatives. Musk reiterated his commitment to keeping Starlink operational in Ukraine, despite political disagreements, and denied using the service as a bargaining chip.

The ongoing debate highlights growing tensions surrounding the role of private companies in international conflict and the geopolitical importance of satellite technology. Meanwhile, the Franco-British operator Eutelsat saw a surge in stock prices, as speculation grows that it could potentially replace Starlink in providing services to Ukraine.

For more information on these topics, visit diplomacy.edu.

Manus AI platform sparks hype but struggles with performance

Manus, an AI platform launched in preview last week, has quickly gained widespread attention, with tech influencers and Chinese media calling it a breakthrough.

The platform, developed by Chinese company The Butterfly Effect, claims to be an advanced ‘agentic’ AI capable of handling complex tasks autonomously. Early adopters scrambled for invite codes, some of which were reportedly resold for thousands of dollars on Chinese marketplaces.

Despite the hype, users report significant performance issues. The platform relies on existing AI models, including Anthropic’s Claude and Alibaba’s Qwen, rather than proprietary technology.

Early testing revealed frequent errors, with some users encountering endless loops and failures when attempting tasks such as booking flights or ordering food. Even seemingly straightforward requests resulted in incomplete or broken responses, raising doubts about Manus’ readiness for real-world use.

Critics argue that exclusivity and marketing have fueled Manus’ popularity more than its actual capabilities. Viral videos misrepresented the platform’s abilities, while comparisons to Chinese AI firm DeepSeek exaggerated its technological depth.

The developers acknowledge the shortcomings, describing the closed beta as a stress test and promising improvements. For now, the platform’s rapid rise appears to be driven more by perception than performance.

For more information on these topics, visit diplomacy.edu.

ECB plans digital euro launch by 2028

Christine Lagarde, President of the European Central Bank (ECB), has confirmed the bank’s target to finalise preparations for the digital euro by October 2025. However, the project’s actual launch remains uncertain, as it hinges on legislative approvals and cooperation from various stakeholders. Despite the urgency, the ECB is facing delays due to the complexity of the legislative process.

The digital euro will consist of two components: a retail version for public use and a wholesale version for financial institutions. The retail version promises privacy protections, free transactions, and offline functionality, while the wholesale arm aims to streamline interbank settlements and cross-border payments. Although preparations are underway, experts predict that a full launch may not occur until 2028.

Privacy concerns and potential impacts on commercial banks are among the challenges the ECB is addressing. To reassure the public, the ECB has committed to strong privacy standards and is exploring blockchain technologies like Ethereum to underpin the digital euro. The project comes as global competitors, such as China’s digital yuan and the rise of US stablecoins, intensify the pressure on Europe to maintain its monetary sovereignty.

While the ECB is making significant strides, the final approval and launch of the digital euro will depend on future legislative decisions and overcoming technical hurdles. The timeline remains uncertain, but the ECB’s preparations signal that the eurozone is keen to remain competitive in the digital currency space.

For more information on these topics, visit diplomacy.edu.

New York MTA partners with Google to detect track problems

The Metropolitan Transportation Authority (MTA) in New York City has partnered with Google Public Sector on a pilot program designed to detect track defects before they cause significant disruptions. Using Google Pixel smartphones retrofitted onto subway cars, the system captured millions of sensor readings, GPS locations, and hours of audio to identify potential problems. The project aimed to improve the efficiency of the MTA’s response to track issues, potentially saving time and money while reducing delays for passengers.

The AI-powered program, called TrackInspect, analyses the sounds and vibrations from the subway to pinpoint areas that could signal defects, such as loose rails or worn joints. Data collected during the pilot, which ran from September 2024 to January 2025, showed that the AI system successfully identified 92% of defect locations found by human inspectors. The system was trained using feedback from MTA inspectors, helping refine its ability to predict track issues.

While the pilot was considered a success, the future of the program remains uncertain due to financial concerns at the MTA. Despite this, the success of the project has sparked interest from other transit systems looking to adopt similar AI-driven technologies to improve infrastructure maintenance and reduce delays. The MTA is now exploring other technological partnerships to enhance its track monitoring and maintenance efforts.

For more information on these topics, visit diplomacy.edu.