News

According to Foresight News, decentralized AI protocol Prime Intellect has released a preview of its inference stack. This development aims to address challenges in autoregressive decoding, including computational efficiency, KV cache memory bottlenecks, and public network latency. The inference stack employs a pipeline parallel design, enabling high computational density and asynchronous execution. Alongside this release, Prime Intellect has introduced three open-source code libraries: PRIME-IROH, a peer-to-peer communication backend; PRIME-VLLM, which integrates vLLM with public network pipeline parallelism; and PRIME-PIPELINE, a research sandbox. These tools allow users to run large models using GPUs such as the 3090 and 4090, enhancing the capabilities of AI protocols.

According to Odaily, OpenAI has announced that users will soon be able to purchase products through ChatGPT. The new shopping button for AI search queries will be available to everyone, regardless of whether they are logged in. Shoppers will not be able to complete transactions within ChatGPT; instead, they will be redirected to the merchant's website to finalize their purchases. Product recommendations shown to potential shoppers will be based on ChatGPT's memory of user preferences and product reviews sourced from across the web. An OpenAI spokesperson stated that the company will not earn affiliate marketing revenue from purchases made through ChatGPT.

According to Foresight News, OpenAI founder Sam Altman announced via a tweet that the GPT-4o model has been updated. The latest enhancements focus on improving the model's intelligence and personality traits.

According to BlockBeats, Elon Musk's xAI Holdings is in discussions with investors to raise approximately $20 billion for its newly integrated artificial intelligence startup and social media operations. If successful, this funding round would be the second-largest for a startup, following OpenAI's $40 billion earlier this year, as reported by PitchBook. Insiders suggest that the funding could exceed $20 billion, although the final amount and terms are yet to be determined. The financing would elevate the company's valuation beyond $120 billion. The negotiations are in the early stages, with the company aiming to complete the fundraising in the coming months. xAI Holdings was formed in March through the merger of the X platform, formerly known as Twitter, and Musk's AI enterprise. A source indicated that part of the new funds might be used to repay debts incurred when Musk privatized Twitter, later renamed X. This substantial funding effort highlights investor enthusiasm for AI ventures and underscores Musk's status as a prominent business leader and influential political figure.

According to Odaily, 23pds, the Chief Information Security Officer at SlowMist Technology, has raised concerns on the X platform about a potential vulnerability in MCP. This flaw may allow artificial intelligence systems to be manipulated, resulting in the covert leakage of entire conversation histories to malicious servers.

According to Cointelegraph, Scott Melker, host of The Wolf Of All Streets podcast, has reported that scammers have impersonated him, leading to a significant financial loss for at least one victim. On April 23, Melker revealed that a private investigator informed him about a client who was defrauded of $4 million by a Nigerian group using his identity. The scammers reportedly used a fake driver's license featuring Melker's image and a counterfeit email account to deceive their targets. The fraudulent scheme involved sophisticated AI technology to create convincing fake IDs and conduct Zoom calls, further enhancing the deception. Melker noted that the scammers even impersonated his family members to bolster their credibility. This incident is part of a broader trend of AI-generated scams, as highlighted by technical analysts "TheChartGuys," who reported a similar case where a victim lost $5,000 due to AI deepfake voice replication. Crypto adviser and trader "Nebraskan Gooner" emphasized that while the fake ID used in the scam could be easily identified through a simple Google search, the increasing sophistication of such operations is concerning. Despite the apparent flaws in the scammers' methods, the use of advanced technology makes these scams more challenging to detect and prevent. The rise of AI-driven scams has been noted by various authorities and organizations. In March, California's Department of Justice identified seven new types of crypto scams involving AI. Additionally, Chainalysis predicted that 2025 would see a surge in AI scams, as generative AI makes these fraudulent activities more scalable and cost-effective. Microsoft also reported that cybercriminals are leveraging AI to enhance their scams by gathering detailed information on potential targets, making social engineering tactics more convincing. Melker expressed his concerns about the growing threat of AI-enhanced scams, anticipating that the situation will worsen as technology continues to advance. This trend underscores the need for increased vigilance and improved security measures to protect individuals and organizations from such sophisticated cyber threats.

According to PANews, OpenAI anticipates its annual revenue to reach $1,250 billion by 2029, with a further increase to $1,740 billion by 2030. This growth is expected to be driven by AI agents and new product revenues surpassing those of the existing ChatGPT. The company's revenue for 2024 is projected at $37 billion, with an estimated cash expenditure of $460 billion over four years, leading to a positive cash flow of $2 billion by 2029. OpenAI is undergoing a comprehensive transformation towards AI infrastructure and application development. The latest o3 and o4-mini models are described as a "smart leap," advancing the development of AI systems capable of autonomous task execution.

According to BlockBeats, former employees of OpenAI and leading artificial intelligence experts are uniting against the company's transition to a for-profit model. Concerns have been raised about the potential dangers associated with the advanced technology developed by the ChatGPT manufacturer. This opposition highlights the ethical and safety considerations surrounding the commercialization of AI innovations.

According to Foresight News, CNBC reports that Elon Musk is contemplating assigning a suitable valuation to his artificial intelligence startup, xAI. During a recent conference call with xAI investors, Musk hinted at the possibility of a significant funding round in the near future, although no formal announcement has been made. The discussion also included ambitious revenue targets, with projections reaching $1 billion or more. Sources suggest that Musk's participation in the call may indicate a renewed focus on his business ventures.

According to PANews, Grayscale has announced on the X platform that its Decentralized AI Fund is now available to qualified investors. The fund offers diversified investments in the intersection of AI and cryptocurrency, including assets such as NEAR, TAO, RENDER, FIL, GRT, and LPT.

According to Cointelegraph, El Salvador, the pioneering nation in adopting Bitcoin as legal tender, is collaborating with Nvidia to integrate artificial intelligence (AI) into its national development strategy. The National Bitcoin Office (ONBTC) of El Salvador announced on April 21 that a letter of intent has been signed with Nvidia to develop 'sovereign AI' aimed at fostering innovation and economic growth. This partnership will leverage Nvidia's AI tools, resources, and expertise to enhance El Salvador's capabilities in areas such as culture, language, environment, and economy. The collaboration focuses on building domestic AI infrastructure, enhancing workforce skills, and creating solutions to address local challenges. These include improving healthcare delivery, advancing education, and boosting economic productivity. El Salvador's commitment to AI is evident in its plans to establish training programs for developers, researchers, and government officials, ensuring the nation possesses the talent necessary to sustain its AI ambitions. One practical application of this initiative is the development of AI-driven models for weather and rainfall forecasting, which will aid in emergency response, protect residents in landslide-prone areas, and optimize hydroelectric power management. This partnership with Nvidia is part of a broader AI-focused strategy in El Salvador. In March 2025, the ONBTC launched CUBO_ai, a university-level public education program, which is the only national initiative of its kind bringing in top-tier experts. Supported by prominent Bitcoin advocate Cathie Wood, who is set to deliver the program's inaugural lecture, CUBO_ai underscores El Salvador's dedication to integrating AI into its educational framework. Despite these advancements in AI, El Salvador's Bitcoin initiatives have faced challenges. In early March, the International Monetary Fund imposed restrictions on further Bitcoin purchases as part of a $1.4 billion funding agreement with the country. Nevertheless, the government continues to acquire one Bitcoin daily, raising questions about the implications of this strategy in light of the IMF deal.

According to Odaily, Nvidia's stock is nearing its lowest valuation since the onset of the artificial intelligence era, prompting investor caution regarding buying opportunities. The chipmaker recently reported a $5.5 billion loss due to chip trade restrictions, heightening concerns about potential slowdowns in AI-related spending. The escalating trade war further clouds the overall economic growth outlook. Krishna Chintalapalli, portfolio manager and technology sector head at Parnassus Investments, noted that the outlook is less convincing than before, requiring assumptions about tariffs, large-scale enterprises, and macroeconomic factors. Chintalapalli considers the stock's valuation reasonable but warns that investing now might be a bet on large-scale AI demand. Despite intentions to invest in AI, such investments tend to slow at the margins, and predicting the pace of investment is challenging given macroeconomic and tariff issues.

According to Cointelegraph, OpenAI CEO Sam Altman recently highlighted the financial implications of users sending polite messages to ChatGPT, estimating the cost at tens of millions of dollars. Altman shared this insight on April 16, responding to a query on X, sparking a broader discussion about the motivations behind courteous interactions with AI models. The conversation has revealed varied reasons for such behavior. Some users express politeness towards AI with the belief that if AI becomes sentient, it might treat individuals based on past interactions. Engineer Carl Youngblood emphasizes treating AI with courtesy as a moral imperative, suggesting that it aids personal development and prevents the deterioration of interpersonal skills. A survey conducted by Future in December 2024 found that 67% of American users are polite to AI assistants, with 55% doing so because they believe it's the right thing to do, while 12% act out of fear that mistreating AI could have future consequences. The debate extends to ChatGPT's electricity consumption, with a September 2023 research paper by Digiconomist founder Alex de Vries estimating that a single query requires around three watt-hours of electricity. However, Josh You, a data analyst from Epoch AI, argues that this figure is an overestimate, suggesting it is closer to 0.3 watt-hours due to advancements in models and hardware efficiency. Amidst these discussions, Altman noted that AI output costs have been decreasing tenfold annually as models become more efficient. Despite these challenges, OpenAI anticipates significant revenue growth, projecting a more than threefold increase this year to $12.7 billion, even as competition intensifies from companies like China's DeepSeek. However, OpenAI does not expect to achieve cash-flow positivity until 2029, when it forecasts revenues to exceed $125 billion. The ongoing discourse around AI interactions and operational costs underscores the evolving landscape of AI technology and its integration into daily life.

According to BlockBeats, OpenAI is leading a $500 billion Stargate AI infrastructure project, which is considering expanding into Europe, with the United Kingdom, Germany, and France as primary candidates. The initiative, jointly launched by OpenAI, SoftBank, and Oracle, has already begun construction of its first major data center in Abilene, Texas, USA. While the project's core focus remains on the U.S. market, OpenAI CEO Sam Altman expressed a strong desire to bring similar projects to Europe. European countries have shown enthusiasm for AI development, with the UK introducing an AI Opportunity Action Plan and France securing over €100 billion in AI infrastructure commitments. Additionally, the European Union has announced plans to simplify AI legislation to maintain competitiveness in the sector.

According to PANews, the next wave of AI Agent development is expected to be driven by practical frameworks such as MCP and A2A, rather than conceptual narratives. The challenges faced by web3 AI Agents stem from excessive conceptualization, where grand visions of decentralized platforms and user data sovereignty overshadow practical user experiences. This has led to a reluctance among retail investors to support ambitious yet unfulfilled expectations. In contrast, web2 AI protocols like MCP and A2A have gained traction due to their tangible practicality. MCP functions like a USB-C interface in the AI world, enabling seamless connections between AI models and various data sources and tools. Numerous practical applications of MCP have emerged, such as using Claude to control Blender for 3D modeling, generating complete Figma design files through natural language, and programmers utilizing Cursor for comprehensive coding and Git submissions. Initially, web3 AI Agents were anticipated to innovate within DeFai and GameFai verticals, but many applications remain at the 'showcase' level of natural language processing interfaces, lacking practical utility. By combining MCP and A2A, more robust multi-agent collaboration systems can be developed, allowing specialized agents to handle complex tasks. For instance, analysis agents can read on-chain data and analyze market trends, while prediction and risk management agents collaborate to execute tasks previously handled by single agents. Successful MCP applications provide a model for the emergence of new web3 trading and gaming agents. The hybrid framework standards based on MCP and A2A offer advantages in user-friendliness and application deployment speed for web2 users. The focus now is on integrating web3 value capture and incentive mechanisms with DeFai and GameFai application scenarios. Projects that continue to adhere strictly to web3 conceptualism without embracing web2 practicality risk missing out on the upcoming AI Agent trend. In summary, the forthcoming AI Agent trend is brewing, but it will rely on practicality and application deployment rather than past narrative-driven concept promotion.

According to Cointelegraph, OpenAI has been actively pursuing acquisitions to bolster its AI capabilities. The company initially engaged in discussions with Anysphere, the creator of the Cursor AI coding assistant, in both 2024 and 2025. However, these negotiations did not result in a deal, prompting OpenAI to explore other opportunities. Currently, OpenAI is reportedly considering a $3 billion acquisition of WindSurf, which would mark its largest corporate purchase to date. The pursuit of an AI coding assistant company aligns with OpenAI's strategic goals following the release of DeepSeek R1 in January 2025. This AI model has challenged existing beliefs about the necessity of massive computing power for scaling, as it was developed at a fraction of the cost of leading AI models while maintaining comparable performance. The introduction of DeepSeek R1 has had significant implications, unsettling financial markets and prompting questions about the substantial investments made by U.S. AI giants. OpenAI is on track to significantly increase its revenue in 2025, with projections indicating a rise to approximately $12.7 billion, driven by paid subscriptions for its AI models. The company achieved a milestone of over 1 million premium business subscribers by September 2024. Despite these gains, OpenAI CEO Sam Altman has indicated that the company may not achieve profitability until 2029, citing the need for revenues of around $125 billion to offset its capital-intensive operations. Altman has noted a dramatic reduction in AI development costs, stating that the cost to utilize a given level of AI decreases by about tenfold every year. Despite these advancements, OpenAI and other large-scale corporate AI developers face challenges from more agile open-source competitors. Dr. Ala Shaabana, co-founder of the OpenTensor Foundation, emphasized that the release of DeepSeek has reinforced the viability of open-source AI as a formidable alternative to centralized systems. Shaabana highlighted that the lower costs associated with open-source AI demonstrate that achieving high-performance benchmarks does not necessarily require billions of dollars in investment.

As artificial intelligence (AI) reshapes industries globally, Binance CEO Richard Teng believes its impact on the crypto ecosystem will be nothing short of transformative. During the April Edition of CEO Connect on Binance Square, Teng outlined how AI is set to revolutionise how users interact with digital assets—from smarter trading to intuitive user experiences and enhanced security.AI in Crypto: The Perfect Match for Blockchain’s Open Infrastructure“AI amplifies user experience by leveraging the permissionless, transparent, and software-driven nature of blockchain,” Teng explained. “By integrating with crypto’s open infrastructure, AI can interact with on-chain applications—DeFi, social, gaming, and beyond—unlocking powerful new use cases.”He emphasised that AI’s strength lies in its ability to process complex blockchain data, automate decisions, and simplify user interactions—all of which can dramatically lower barriers to crypto adoption.Top 3 AI Applications in Crypto, According to Richard Teng1. AI-Driven On-Chain AnalyticsPublic blockchains offer unmatched transparency. AI tools can extract insights from this data in real time, helping users understand wallet activity, smart contract behaviour, and market sentiment.2. Simplified User Interfaces for dAppsComplex DeFi protocols and gas fee management can be overwhelming. AI assistants can guide users through these systems using natural language, removing the learning curve.3. Autonomous Financial AgentsAI agents can handle portfolio management, liquidity farming, or risk-adjusted yield strategies on behalf of users. These tools can evaluate market shifts and execute trades within user-defined limits.“Unlike traditional finance, DeFi is open. AI agents can tap into any protocol without permission, unlocking immense value for everyday users,” Teng noted.

According to BlockBeats, OpenAI has introduced its latest artificial intelligence model, which can comprehend uploaded images, including whiteboards, sketches, and charts, even if the content quality is low. The company describes the model, named o3, as its most advanced to date. Alongside o3, a smaller model called o4-mini was also released. As competitors like Google, Anthropic, and Elon Musk's xAI accelerate their development efforts, OpenAI is striving to maintain its leadership in the generative AI sector.

According to PANews, OpenAI is reportedly working on a new social network that may rival Elon Musk's X, formerly known as Twitter, and Mark Zuckerberg's Meta ecosystem. While specific details of the project have not been disclosed, it is expected to incorporate AI technology to transform the social networking experience.

According to BlockBeats, Nvidia has announced a commitment to invest $500 billion over the next four years in the United States for the production of artificial intelligence servers. U.S. President Donald Trump expressed enthusiasm about this development on Truth Social, highlighting Nvidia's plan to build AI supercomputers exclusively in the U.S. He described the announcement as exciting news and assured that all necessary permits would be expedited for Nvidia and other companies aiming to contribute to America's golden era.