Pulse

The U.S. Pentagon has signed classified deals with several major AI companies, including Google, OpenAI, and Nvidia, to utilize their AI models for "any lawful government purpose" on classified networks. This move aims to establish the U.S. military as an "AI-first fighting force" and enhance intelligence operations. Notably, Anthropic was excluded from these new agreements due to its refusal to remove safety guardrails concerning autonomous weapons and domestic surveillance, leading to a dispute and its designation as a supply-chain risk. AI

Google and Amazon reported significant profits in the last quarter, largely driven by accounting markups on their stakes in Anthropic. Alphabet's profit increased by 81% to $62.6 billion, with $28.7 billion attributed to the revaluation of its Anthropic equity. Similarly, Amazon's net income surged 77% to $30.3 billion, including a $16.8 billion pre-tax gain from its Anthropic investment. This accounting practice, while standard, highlights how increased investment and business transactions with Anthropic can inflate the reported value of their stakes. AI

Alphabet reported a 22% year-over-year revenue increase to $109.9 billion for its first quarter, with Google Cloud revenue surging 63%. Microsoft also saw strong performance, with an 18% revenue increase to $82.9 billion in its third fiscal quarter. The reports highlight significant growth in cloud services and overall revenue for both tech giants. Separately, the news mentions AI coding tools like Copilot and Claude are increasing prices, prompting a discussion about the cost-effectiveness of AI versus human programmers. AI

OpenAI has made its GPT models, Codex, and Managed Agents accessible through Amazon Web Services (AWS). This integration allows businesses to develop and deploy AI applications securely within their existing AWS infrastructure. The partnership aims to broaden the reach of OpenAI's technology to a wider enterprise audience. AI

Wiz researchers discovered a critical vulnerability in GitHub's git infrastructure, enabling remote attackers to gain full read/write access to private repositories. They utilized AI tools, specifically Claude Code and IDA MCP, to accelerate the reverse-engineering process, reducing the time from idea to exploit from months to under 48 hours. GitHub responded rapidly, patching the vulnerability within six hours and awarding Wiz one of its largest bug bounty payouts. AI

China's National Development and Reform Commission (NDRC) has blocked Meta's acquisition of the AI startup Manus, demanding the deal's cancellation and the return of Manus's assets within China. This intervention highlights China's increasing scrutiny of AI acquisitions and its efforts to prevent sensitive technology and talent from moving overseas, a practice sometimes referred to as 'Singapore-washing'. Despite Meta's claims of compliance and potential difficulties in unwinding the deal, China aims to set a precedent against bypassing national regulations. AI

Taiwan Semiconductor Manufacturing Company (TSMC) is accelerating its expansion plans to meet surging demand for AI and high-performance computing. The company is advancing its production capacity at double the usual pace, with five 2nm wafer fabrication plants entering their ramp-up phase this year. This aggressive expansion is projected to result in a 45% increase in 2nm output during its first year compared to the 3nm process. AI

OpenAI has reportedly missed key revenue and user growth targets, sparking concerns about its ability to fund future compute agreements as it approaches an IPO. This news has led to a decline in the stock prices of major partners including Oracle, Nvidia, AMD, and CoreWeave. While OpenAI has disputed the report, some analysts suggest the slowdown is a natural consequence of increased competition from rivals like Anthropic and Google's Gemini models. AI

Meta is launching a new AI model called "Meta Muse Spark" as part of a significant investment to bolster its AI capabilities. This initiative aims to challenge existing AI dominance by leveraging advanced visual processing power to integrate AI more deeply into daily life. The company is also exploring a private operational model to establish its own revenue streams. AI

Sony's Project Ace, an advanced AI system, has demonstrated its capabilities by defeating professional ping-pong players. This achievement marks a significant advancement in the field of physical AI, suggesting a future where AI systems can compete and excel in real-world physical activities. AI

OpenAI is reportedly developing its own smartphone, aiming for mass production by 2028, potentially integrating ChatGPT features directly. Separately, Anthropic's valuation is soaring, with traders pushing its on-chain market cap towards $1 trillion ahead of a potential IPO. New AI models are also emerging, including exaBase's GPT Images 2.0 and Claude Opus 4.7, which promise enhanced performance and image generation capabilities. AI

Together, an open-source AI lab, has announced its participation in the Genesis Mission, a project aimed at doubling American scientific productivity over the next decade. The initiative connects supercomputers, experimental facilities, and datasets across national laboratories and academic institutions. Together's platform supports over a million developers globally and is designed to provide inspectable and adaptable AI infrastructure, fostering scientific research without vendor lock-in. AI

The European Commission has issued proposed measures to Google, requiring the company to grant AI competitors access to core Android functionalities. This move aims to ensure compliance with the EU's Digital Markets Act, as Google currently reserves key capabilities like voice activation and background processes for its own Gemini AI services. Separately, Microsoft and OpenAI have adjusted their partnership terms, with Microsoft no longer receiving revenue share from OpenAI but maintaining IP licensing until 2032, while continuing as OpenAI's primary cloud partner. AI

NTT is planning a significant expansion of its data center infrastructure, focusing on AI-native facilities. The company intends to construct AI data centers in central Tokyo that will feature liquid cooling as a standard. This initiative, branded as "AIOWN," aims to provide the necessary infrastructure to support the growing demands of AI development and deployment. AI

OpenAI has achieved FedRAMP Moderate authorization for its ChatGPT Enterprise and API services. This accreditation allows U.S. federal agencies to securely utilize OpenAI's AI capabilities. The authorization is a significant step towards broader government adoption of advanced AI technologies. AI

Over 600 Google employees, including those from DeepMind and Google Cloud, have signed a letter to CEO Sundar Pichai. They are urging the company to refuse classified work for the Pentagon. The employees cite concerns about lethal autonomous weapons and mass surveillance. AI

The European Commission has proposed new measures under the Digital Markets Act (DMA) that would require Google to grant third-party AI services deeper access to Android's core functionalities. Currently, features like sending emails or sharing photos are largely exclusive to Google's own AI offerings, such as Gemini. The Commission aims to ensure rival AI assistants can integrate with applications and perform tasks on user devices with the same ease as Google's native tools, fostering greater competition and user choice. AI

Google has introduced two new AI chips, the TPU 8t for training and the TPU 8i for inference, marking a split in its custom silicon strategy. This move comes as major AI players like OpenAI, Anthropic, and Meta are reportedly purchasing significant quantities of these chips. Concurrently, Pony.ai has unveiled a new autonomous driving domain controller powered by NVIDIA's DRIVE platform, aimed at large-scale L4 deployment. AI

OpenAI is evolving ChatGPT into a comprehensive "super app" integrating chat, coding, and browsing functionalities, alongside an upgraded image generation tool that exhibits improved reasoning capabilities. Concurrently, Google is set to embed its Gemini AI deeply into Android 17, aiming to automate tasks across applications and make AI an invisible, integral part of user interaction. This strategic shift by both tech giants highlights an intensifying competition and a global race to position AI at the core of daily digital activities. AI

Core Scientific, a company previously focused on Bitcoin mining, is pivoting to become a major AI datacenter operator. The company plans to convert its 300-megawatt facility in Pecos, Texas, into a 1.5-gigawatt AI campus, with the first phase expected online in early 2027. This strategic shift involves significant investment, including plans to issue $3.3 billion in junk bonds to finance the transition. Core Scientific is also exploring various power solutions to meet the immense energy demands of AI operations. AI

DeepSeek has significantly reduced its API prices by up to 90% following the release of its V4 model. The company attributes these price cuts, which establish a new industry low, to its sparse attention architecture. This new architecture reportedly lowers per-token compute needs and supports context windows of up to 1 million tokens. AI

Google is investing $10 billion into the AI firm Anthropic, a significant move in the competitive AI landscape. Additionally, Hershey is exploring the use of AI agents to address its business challenges, and Singtel is partnering with Mistral AI to develop GPU and AI-as-a-Service offerings. These developments highlight the increasing integration of AI across various industries, from tech giants to consumer goods. AI

In August 2023, the Biden administration issued Executive Order 14105 to limit US investment in China's advanced technology sectors, including AI and semiconductors. This order was intended to prevent China's military from acquiring critical technologies. Subsequently, Manus, a company that had established a subsidiary in China, faced US scrutiny after its funding rounds were led by Benchmark. AI

A recent survey by NASCIO and Deloitte reveals a significant drop in confidence among State Chief Information Security Officers (CISOs), falling from 48% to 26% over two years. This decline is attributed to two primary factors: an increase in AI-powered cyber threats and concurrent budget reductions. These challenges are impacting the ability of state governments to maintain robust cybersecurity postures. AI

Ineffable Intelligence, a new AI lab founded by former DeepMind researcher David Silver, has secured $1.1 billion in funding. The company aims to develop a "superlearner" that can acquire knowledge and skills autonomously through reinforcement learning, without relying on human-generated data. Silver, known for his work on AlphaZero, seeks to create an AI capable of discovering all intelligence from its own experience. The funding round was led by Sequoia Capital and Lightspeed Venture Partners, with participation from Google and Nvidia. AI

OpenAI and Microsoft have renegotiated their partnership, significantly reducing Microsoft's exclusivity over OpenAI's products. OpenAI can now offer its services across all cloud providers, including rivals like Amazon Web Services (AWS) and Google Cloud, while Microsoft retains its status as OpenAI's primary cloud partner and a significant shareholder. The revised agreement also removes the controversial "AGI clause" that previously dictated terms based on the achievement of artificial general intelligence, simplifying their financial arrangements and extending Microsoft's IP rights through 2032. AI

South Africa has suspended two officials from its Department of Home Affairs after an AI tool generated fictitious citations in a revised white paper on citizenship, immigration, and refugee protection. Separately, the Department of Communications and Digital Technologies withdrew its draft national AI policy due to similar AI-generated fabricated sources. These incidents highlight the risks of AI hallucinations in policy drafting and have prompted reviews of government documents and the implementation of AI checks. AI

Anthropic's new Claude Mythos Preview model has exposed significant vulnerabilities in the UK's financial institutions, prompting urgent meetings between regulators and experts. Separately, a collaboration between Matterhorn and ASI Alliance aims to enhance security in Web3 by introducing a mathematically verifiable process for smart contract creation, mitigating risks associated with AI-generated code. AI

Texas is poised to become the world's largest data center market by 2030, with over 400 facilities either operational or under construction. These data centers are projected to consume a significant amount of water, estimated between 29 and 161 billion gallons annually by the same year. This rapid expansion highlights growing concerns about the environmental impact of the burgeoning tech infrastructure. AI

David Silver, a key figure behind DeepMind's AlphaGo and other AI projects, has launched a new research lab called Ineffable Intelligence. The lab aims to create a "superlearner" that acquires knowledge through direct experience rather than pre-training on existing data. This approach, rooted in reinforcement learning, seeks to enable AI to make novel discoveries in science and mathematics. AI

Opening arguments have begun in the legal battle between Elon Musk and Sam Altman, stemming from Musk's lawsuit against OpenAI. The trial is expected to delve into the foundational disagreements regarding the company's direction and commercialization. While the case could touch upon AI safety, its core appears rooted in personal grievances and financial disputes. AI

Supermicro has announced a new campus in Silicon Valley, signaling a significant change in how enterprises are purchasing AI infrastructure. Companies are now prioritizing vendors based on domestic production speed, integration capabilities, and operational accountability, moving beyond just securing access to AI chips. This shift indicates a growing demand for more robust and domestically-supported AI hardware solutions. AI

The U.S. Justice Department has filed a lawsuit against the State of Colorado, seeking to invalidate the state's recently enacted artificial intelligence law. This intervention challenges Colorado's regulations concerning algorithmic discrimination, arguing they are preempted by federal law. The department's action aims to prevent the enforcement of these state-level AI rules. AI

Signull Labs has secured over $3.58 million in pre-seed funding for its upcoming iPhone app, Skye, which aims to create an "agentic homescreen." The app utilizes iOS widgets to provide ambient intelligence, offering personalized insights and performing tasks like drafting emails and flagging suspicious transactions. Despite being in private testing, Skye has garnered significant user interest, with tens of thousands joining a waitlist, suggesting a demand for more AI-integrated smartphone experiences. AI

Taylor Swift has filed trademark applications for her voice and image to protect against AI misuse, following a similar strategy by Matthew McConaughey. The applications cover specific spoken phrases and a distinctive photograph from her Eras Tour. This move aims to provide a legal framework to combat AI-generated deepfakes and unauthorized imitations, which copyright law alone cannot address. AI

Meta is exploring innovative energy solutions to power its expanding AI data centers, announcing a partnership with Overview Energy to potentially beam solar power from space. The company has also secured a significant energy storage deal with Noon Energy, aiming for 100 hours of backup power. These initiatives are driven by the immense energy demands of AI development and the limitations of current clean energy technologies. AI

Companies are experiencing significant pressure on their AI budgets, with many exhausting their full-year allocations within just a few months. This rapid depletion is driven by surging costs associated with compute power and the overall boom in AI adoption. The situation highlights a growing challenge for businesses as they navigate the escalating expenses of integrating advanced AI technologies. AI

Microsoft has reportedly reduced its revenue share from OpenAI, signaling a potential shift in their strategic alliance. This move comes as Microsoft increasingly focuses on its own AI development and integration into its products. The exact details of the new revenue-sharing agreement have not been disclosed, but it suggests a move towards a more balanced partnership. AI

Chinese AI lab DeepSeek is reportedly nearing a significant funding round that could value the company at approximately $45 billion. China's state-backed "Big Fund" is expected to lead this investment, with tech giants Tencent and Alibaba also in advanced talks to participate. This valuation marks a substantial increase from previous reports and signals Beijing's strategic focus on developing domestic AI capabilities. AI

Microsoft and OpenAI are entering the next phase of their partnership, focusing on accelerating AI innovation. This collaboration aims to combine OpenAI's cutting-edge AI models with Microsoft's extensive cloud infrastructure and distribution channels. The goal is to bring advanced AI capabilities to more businesses and consumers worldwide. AI

Polish voice actors and narrators have formed a coalition to combat the unauthorized use of their voices by generative AI. This initiative, named "My Voice. My Property," was prompted by a lawsuit where a narrator's digital voice model was used in an advertisement without consent. The coalition is demanding legal regulations that require explicit consent for AI voice usage, transparency in AI training data, the right to remove voice data from AI systems, and clear labeling of AI-generated content. AI

Microsoft and OpenAI have restructured their partnership, ending Microsoft's exclusive rights to distribute OpenAI's models. While Microsoft will continue to be a primary cloud partner and hold IP licenses until 2032, its access to OpenAI's technology will now be non-exclusive. This change allows OpenAI to offer its models through other cloud providers, potentially broadening its reach and infrastructure options. AI

David Silver, the AI researcher behind AlphaGo, has founded Ineffable Intelligence to pursue a different path to superintelligence. His company will focus on reinforcement learning, where AI learns through trial and error, contrasting with the large language model (LLM) approach favored by most labs. Silver believes LLMs, which learn from human data, will ultimately fail to achieve true superintelligence, unlike systems that can learn from the world independently. Ineffable Intelligence has secured $1.1 billion in seed funding and aims to create AI capable of independent scientific discovery. AI

SpaceX successfully launched its Falcon Heavy rocket, carrying the ViaSat-3 F3 communications satellite into orbit. This marks the Falcon Heavy's first flight in over 18 months, with SpaceX aiming to recover both side boosters upon landing. The mission is significant as it initiates a period of increased activity for the heavy-lift vehicle, with further launches planned for a lunar lander and a space telescope. AI

The Royal Dutch Military Police collaborated with Palantir, a technology firm known for its data analysis capabilities. This partnership was reportedly kept from the House of Representatives by a minister. The collaboration has raised concerns regarding surveillance and privacy. AI

Mistral AI has established a significant presence in the AI landscape, reportedly building a $14 billion empire by focusing on open-source models and diverging from US-centric development. This approach leverages European data ethics and collaborative efforts to set a benchmark for responsible AI. Meanwhile, discussions are emerging about the future AGI economy, exploring concepts like self-organizing agent ecologies and superintelligence arcologies that could reshape civilization. AI

Microsoft and OpenAI have amended their partnership agreement to provide long-term clarity and support continued AI innovation. This revised deal simplifies their collaboration, aiming to foster scaled AI development. The specifics of the amended agreement were not fully detailed, but it signifies a deepening of the relationship between the two tech giants. AI

Utimaco's integration with VAST Cosmos signals a shift in enterprise AI infrastructure, driven by customer demands for enhanced security. Businesses are increasingly scrutinizing encryption key ownership, data residency, and accountability measures before expanding their AI operations. This move highlights a growing focus on robust security and data sovereignty within the AI sector. AI

China has blocked Meta's proposed acquisition of a Chinese AI company, reportedly valued at over a billion dollars. The decision prevents Meta from integrating the company's AI technology into its own platforms. This move highlights ongoing geopolitical tensions and regulatory scrutiny surrounding AI development and cross-border tech investments. AI

The UK Home Office is deploying AI tools in asylum decision-making without adequate safeguards, transparency, or clear accountability structures. Critics argue that AI is not neutral and can lead to discriminatory or unfair outcomes, especially in life-changing asylum assessments. Organizations like Open Rights Group are urging the public to contact their MPs to oppose the use of these tools until proper governance and human oversight are established. AI