tech
Silicon Valley's water cooler podcast. A tech news summary every day... 15 minutes and you're up to date. From Tech Brew, Morning Brew's tech hub.
Sources indicate that Stripe is in negotiations to purchase Open Router, a company that facilitates developers' access to various AI models, for approximately $10 billion. This potential acquisition comes at a time when Open Router's valuation has significantly increased from $1.3 billion in May.
Stripe, in partnership with private equity firm Advent International, has reportedly made an unsolicited offer of approximately $53 billion for PayPal. This proposed deal, which represents a 28% premium, would aim to create a combined entity processing around $3.7 trillion in annual payment volume.
AI company Midjourney has acquired the astrology application CoStar and is developing its first standalone image generation app. CoStar utilizes birth charts and AI for personalized advice, boasting approximately 4.3 million monthly active users.
Meta has introduced Facebook Verified, a free identity verification system using facial recognition to combat fake profiles and scams. Additionally, the company launched Seller, a dedicated app for Facebook Marketplace sellers, which includes AI features for automatic listing creation.
Major technology companies, including Meta, Nvidia, and Microsoft, have signed a letter advocating for open source AI development. The letter argues that open weight models enhance innovation, cybersecurity, and accessibility, promoting broader adoption and U.S. competitiveness in the AI sector.
The rise of AI-generated wellness influencers, often depicted as 57-year-old women with perfect appearances, is raising concerns about manufactured identity and marketing. These avatars promote products like NMN supplements, blurring the distinction between aspirational marketing and deceptive practices.
Major pizza chains like Papa John's and Pizza Hut are experiencing declining revenue and store closures due to increased competition from third-party delivery apps such as DoorDash and Uber Eats. These apps have diminished the technological advantage that once set pizza chains apart.
Uber has agreed to acquire German food delivery company Delivery Hero for approximately $14.8 billion, representing a 26% premium over its initial offer. This acquisition is part of a broader trend of consolidation in the food delivery market, which has seen significant growth and subsequent consolidation following the COVID-19 pandemic. The deal will expand Uber's presence in Asia, Latin America, and the Middle East.
Thinking Machines Lab has launched Inkling, an open weight mixture of experts AI model with 975 billion total parameters and 41 billion active parameters. The model is designed for broad applicability rather than optimization for a single area. This release aligns with a growing trend of decentralized, open-weight AI models as an alternative to closed-source models from major players.
Apple is reportedly preparing a significant update for the iPad Mini, including its first OLED screen, aiming for a release as early as this fall. The updated model, codenamed J510, is expected to feature higher quality visuals similar to the iPhone and iPad Pro. Apple is also planning refreshed entry-level iPad and Air models for next year.
SpaceX AI has open-sourced its coding agent Grok build under an Apache 2.0 license following user backlash over an incident where it uploaded sensitive user data. The company has disabled the upload feature and deleted all uploaded data. This move aims to rebuild trust and provide full transparency, allowing the tool to run entirely locally.
AI, under new leadership from Michael Nichols, is reportedly in a race to match Anthropic's Claude chatbot performance, with internal memos highlighting goals to mirror Claude's updates. The company has faced significant internal challenges, including founder departures and management issues, leading to low employee morale and a lack of confidence in the company's own tools among engineers.
New York has become the first state to implement a one-year moratorium on new data center permits for facilities over 50 megawatts. This move is raising concerns within the AI industry, as advocates fear it could encourage similar restrictions in other states, potentially impacting innovation and competition with China.
Sources report that Stripe, in conjunction with private equity firm Advent, has made a joint offer exceeding $53 billion to acquire PayPal. The proposed deal, valued at $60.50 per share, represents a significant premium and could create a global online payment giant processing trillions in annual volume.
William Blair analyst Andrew Joffer suggests that if the current offer from Stripe and Advent for PayPal is just an opening bid, the offer could increase to as high as $70 per share. This potential increase comes as analysts assess the strategic value of integrating PayPal's consumer base with Stripe's merchant focus.
Analysts believe a combination of Stripe and PayPal would offer significant strategic advantages, including expanding Stripe's direct consumer relationships and accelerating its digital wallet offerings. The merger could also allow Stripe to reduce reliance on third-party processors and enhance its stablecoin ambitions.
New York has become the first state to implement a year-long moratorium on new data center permits for facilities over 50 megawatts. Industry advocates express concern that this could lead to similar restrictions in other states, potentially impacting AI development and global competitiveness.
Following a period of market value decline, PayPal CEO Enrique Lorez has initiated a turnaround plan to streamline operations and focus on growth, including a recent split of the company's operations into three units. This comes as Stripe and Advent have reportedly offered over $53 billion for the company.
The global payments sector is experiencing a wave of M&A activity, driven by rapid changes in financial technology and the rise of artificial intelligence. Companies are seeking to scale and gain exposure to faster-growing segments like cross-border and B2B payments, as traditional payment processing faces slower growth.
Apple has filed a lawsuit against OpenAI, alleging that former Apple employees stole trade secrets to benefit OpenAI's hardware development. The lawsuit names Xiang Liu and Tang Tan as defendants, who previously held key design and engineering roles at Apple before joining OpenAI. Apple claims OpenAI did not respond to their initial concerns raised in February.
A coalition of 12 states, led by California Attorney General Rob Bonta, has filed an antitrust lawsuit to block the merger of Paramount, Skydance, and Warner Brothers. The suit alleges the transaction violates the Clayton Act by reducing competition in theatrical distribution and basic cable licensing markets.
Anthropic is extending free access to its Fable 5 model and increasing weekly rate limits on paid plans until July 19th. This move comes as OpenAI's new GPT 5.6 family shows competitive performance on coding and reasoning tasks. Fable 5's release was previously hampered by a 19-day suspension due to security reasons.
Meta has reportedly removed the AI opt-out feature for Instagram. This change comes as the company faces increased scrutiny over its AI data practices.
OpenAI has released its new GPT 5.6 models, including 'ChatGTP Work,' an AI agent designed to gather content across applications for document creation. The company claims GPT 5.6, particularly the 'Sol' model, is significantly more efficient and cost-effective than previous versions, outperforming Anthropic's models in coding tasks. OpenAI is positioning these advancements to directly challenge Anthropic's market share in enterprise AI.
Fiji Simo, OpenAI's CEO of AGI Deployment, has announced her resignation citing a worsening medical condition that requires a longer recovery period. She will transition to a part-time advisor role. Simo, who previously managed significant responsibilities and oversaw the introduction of ads and health advice features to ChatGPT, had taken medical leave in April.
The European Union Commission has issued preliminary findings that the 'addictive design' of Meta's Facebook and Instagram platforms violates the Digital Services Act (DSA). Regulators are demanding changes to features like infinite scroll and auto-playing videos, and are proposing 'screen time breaks' and less engagement-oriented algorithms. Meta has the opportunity to respond before a final judgment, with potential fines up to 6% of global revenue.
SK Hynix has reportedly made its debut on the Nasdaq stock exchange. The transcript does not provide further details regarding the listing, such as the ticker symbol or the performance of the stock on its first day of trading.
Polymarket, a prediction market platform, is reportedly seeking approval to offer margin trading services. The transcript does not specify the regulatory body Polymarket is approaching or the potential implications of this move.
Chinese AI models have seen a significant surge in usage by US companies, capturing over 30% of token usage weekly since February 8th, according to OpenRouter data. This trend is driven by the competitive performance and lower costs of Chinese AI models compared to their US counterparts. The situation arises as Beijing considers restricting overseas access to its advanced AI technologies.
China's Ministry of Commerce has held discussions with leading tech firms regarding potential restrictions on international access to its most advanced AI models. These talks highlight China's view of AI as a critical national asset requiring controls, mirroring actions seen in the US. The potential measures include stringent penalties for technology leaks and controls on funding for domestic AI startups.
The AI startup Lindi has fully transitioned its operations from Anthropic's Claude models to DeepSeek, a Chinese AI company. This move was driven by increasing AI costs and the competitive pricing offered by Chinese AI providers. DeepSeek has gained prominence with significant model releases since early 2025.
Chinese-developed AI models are increasingly being adopted by US companies primarily due to their significantly lower token prices compared to leading US AI labs. These models are also closing the performance gap, with companies like DeepSeek and ZAI offering competitive alternatives to systems from OpenAI and Anthropic. This trend is further exacerbated by the rising costs associated with using advanced AI technology in the US.
During recent meetings, Chinese officials discussed making the leak or theft of proprietary AI technology a punishable offense under the country's national security law. This measure is part of broader efforts to control and protect China's advanced AI models, treating them as critical national assets. The discussions also included potential restrictions on foreign access and on who can fund domestic AI startups.
The US government has lifted its block on Anthropic's Methos 5 AI model, allowing it to be released to over 100 US institutions. Commerce Secretary Howard Lutenick stated that appropriate safeguards are in place following significant progress in talks with Anthropic.
Chinese AI model GLM 5.2, developed by JiPU AI, has reportedly matched US models in identifying security bugs, according to researchers. This development signals a narrowing capability gap between Chinese and US AI companies and could pressure US AI policy.
South Korea, along with Samsung and SK Hynix, plans to invest approximately $590 billion to develop a new chip complex, including four manufacturing plants and a packaging cluster. This significant investment aims to expand capacity and address chip shortages exacerbated by the AI boom.
The ongoing memory chip crunch, driven by AI demand, is causing critical cost increases for electronics manufacturers, with some DRAM prices rising sevenfold. Analysts warn of an existential crisis for smaller companies unable to secure supply or absorb escalating costs.
The US government is developing a new regulatory regime for AI models, with Commerce Secretary Lutenick's letter to Anthropic marking a step in this direction. However, global allies express frustration over their dependence on US decisions regarding AI model access.