July 28, 2025
AI data labelling startup Micro1 raising funds at $500M valuation
San Francisco-based Micro1, a competitor to Scale AI, is finalizing a Series A fundraising round that values the startup at about $500 million. Micro1 provides AI data-labeling services and has rapidly grown its revenue from $10 million to $50 million annualized, projecting $100 million by September. The startup has benefited from turmoil at rival Scale AI – after Meta hired Scale’s CEO – causing clients like Google and OpenAI to leave Scale, which in turn boosted Micro1’s business. Former Twitter COO Adam Bain recently joined Micro1’s board as it attracts prominent investors 01A and LG Tech Ventures. Why it matters: Micro1’s rise underscores how demand for high-quality AI training data is spawning new players and how big tech’s talent moves (like Meta poaching Scale’s CEO) are reshaping the AI supply chain.
Source: Reuters
Tesla taps Samsung in $16.5B deal to make next-gen AI chips in Texas
Tesla has signed a $16.5 billion agreement with Samsung Electronics to supply chips for Tesla’s forthcoming "AI6" self-driving and robotics chips, to be produced at Samsung’s new Taylor, Texas factory. Samsung’s foundry business, which has struggled to win customers, saw its stock jump nearly 7% on hopes the deal will bolster its push into AI chips. Tesla CEO Elon Musk confirmed the partnership and noted the Texas plant’s proximity to his home, saying Tesla will help maximize the fab’s efficiency. While the chips won’t be ready for years, Musk says AI6 will power Tesla’s autonomous vehicles and Optimus robots, though analysts caution Tesla often misses timelines. Why it matters: This huge deal highlights Tesla’s ambition to custom-build AI hardware for cars and robots, and it gives Samsung’s U.S. foundry a critical win as nations vie to localize cutting-edge chip production.
Source: Reuters
China’s Zhipu releases open-source GLM-4.5 AI model
Beijing-based AI startup Zhipu unveiled an open-source large language model called GLM-4.5, geared toward creating intelligent agents. Zhipu – one of China’s so-called “AI tigers,” with backing from local governments – made headlines in June when OpenAI noted its rapid progress in winning Chinese government contracts. The new model adds to a flood of Chinese AI models: China has released 1,509 large-language models as of July, far more than any other country (out of ~3,755 globally), according to state media. The launch reflects intensifying competition among Chinese firms to match U.S. advances while adhering to open-source principles. Why it matters: Zhipu’s open-source strategy showcases China’s drive to foster homegrown AI alternatives and indicates Beijing’s support for proliferating AI models even as U.S. export curbs push Chinese firms to become more self-reliant.
Source: Reuters
Microsoft adds AI “Copilot Mode” to Edge browser
Microsoft rolled out a new "Copilot Mode" in its Edge web browser that uses AI to enhance how users search and navigate the web. The Copilot sidebar can perform tasks like aggregating results from all open tabs, conducting multi-tab searches, and providing comparisons, without users having to manually switch tabs. Microsoft said voice commands will also be supported, and that Copilot can optionally access users’ browser history and credentials (if granted permission) to take actions like booking tickets. The feature, free for now on Windows and Mac, comes as rivals like Google have launched AI-only search modes and startups such as Perplexity AI debut their own AI-powered browsers. Why it matters: Edge’s Copilot reflects Big Tech’s race to infuse browsers and search with generative AI assistants, a bid to attract users and advertisers by making web browsing more intuitive and integrated with AI capabilities.
Source: Reuters
U.S. firm Auterion sending 33,000 AI-guided drone kits to Ukraine
Defense tech company Auterion announced it will supply Ukraine with 33,000 AI-driven drone guidance kits under a $50 million Pentagon contract. The kits enable human-launched drones to autonomously track and strike targets up to one kilometer away, even in environments with heavy GPS and signal jamming. Auterion’s CEO said the shipment, which is over ten times larger than previous deliveries of its “AI strike” systems to Ukraine, will greatly expand Ukraine’s fleet of AI-augmented First Person View drones. The move comes as Ukraine plans to procure 4.5 million small drones through 2025 and is urgently adopting AI to improve drone accuracy against Russian forces’ jamming efforts. Why it matters: This massive deployment illustrates how AI is rapidly being integrated into warfare: the U.S. is backing Ukraine with smarter, autonomously targeting munitions to overcome electronic warfare and improve strike precision on the battlefield.
Source: Reuters
July 29, 2025
China forges AI alliances to cut reliance on U.S. tech
At China’s World Artificial Intelligence Conference in Shanghai, leading Chinese AI companies announced two new industry alliances aimed at building a domestic AI ecosystem amid U.S. export curbs. One group, the “Model-Chip Ecosystem Innovation Alliance,” links major large-language model developers (like StepFun and MiniMax) with Chinese chipmakers including Huawei, Biren, Enflame and others hit by U.S. Nvidia chip restrictions. A separate AI committee under Shanghai’s Chamber of Commerce will push deeper AI integration into industry, with members such as SenseTime (also under U.S. sanctions) shifting from face recognition to AI models. The conference showcased new Chinese AI products despite sanctions – Huawei revealed a high-end AI computing cluster (384 of its 910C chips) claimed to rival top Nvidia systems, and tech giants like Tencent, Baidu and Alibaba debuted consumer AI gadgets like smart glasses and virtual human generators. Why it matters: Facing U.S. chip sanctions, China is mobilizing collaborative alliances between its AI software firms and chipmakers to ensure progress continues – even touting Huawei’s latest AI hardware as competitive with Nvidia’s – underscoring Beijing’s determination to achieve AI self-sufficiency and innovation despite export controls.
Source: Reuters
Amazon-backed Skild AI launches ‘Skild Brain’ model for versatile robots
Robotics startup Skild AI – funded by Amazon and SoftBank – unveiled a foundational AI model called "Skild Brain" that can run on almost any type of robot, from factory arms to humanoids. The Skild Brain model enables robots to navigate, balance, and respond more like humans: demo videos showed robots climbing stairs, recovering from shoves, and picking cluttered objects by reasoning about their environment. The model incorporates safety limits on force and is trained with simulated scenarios and human-action videos, then fine-tuned on real robot feedback – a method Skild’s founders (ex-Meta robotics lab leads) say tackles robotics’ lack of large online datasets. Having raised $300 million at a $1.5B valuation last year, the two-year-old Skild is hiring talent from Tesla, Nvidia and Meta and counts enterprise clients like LG’s logistics arm, as it positions itself to quickly add new robotic capabilities across industries. Why it matters: Skild’s launch signals a push to bring generative AI’s breakthroughs into the physical world – a recognition that making robots smarter and more adaptable could unlock huge new markets, with major tech investors betting that general-purpose AI brains will transform everything from warehouses to home robotics.
Source: Reuters
Microsoft seeks new deal to secure OpenAI tech even if it achieves AGI
Microsoft is in advanced talks with OpenAI to rewrite parts of their partnership so that Microsoft retains access to OpenAI’s most advanced models in the future, Bloomberg reported. Under the current contract, OpenAI’s attainment of “artificial general intelligence” (AI surpassing human capabilities) could cut off some of Microsoft’s rights to OpenAI technology. The negotiations aim to remove that clause and set new terms for Microsoft’s ongoing access and equity stake as OpenAI transitions into a public-benefit corporation. Talks have been ongoing for months, spurred by OpenAI’s moves to diversify its cloud partners (it’s now also working with Google, Oracle and others) and a lawsuit from Elon Musk accusing OpenAI of straying from its nonprofit roots. An agreement could be reached within weeks, though neither company commented publicly on the report. Why it matters: This highlights how indispensable OpenAI’s cutting-edge models are to Microsoft’s strategy – so much so that Microsoft is willing to renegotiate its multi-billion dollar investment to ensure it won’t be shut out, especially as OpenAI explores outside cloud providers and even faces legal pressure from its co-founder.
Source: Reuters
Microsoft’s AI cloud lead questioned as OpenAI inks deals beyond Azure
Ahead of Microsoft’s earnings, analysts noted that OpenAI – which spurred big growth in Microsoft’s Azure cloud – has begun shifting some workloads to rivals like Oracle, CoreWeave and Google Cloud. Microsoft’s exclusive licensing of OpenAI tech fueled a ~35% jump in Azure revenue last quarter, but OpenAI’s push for independence (including a planned $40B SoftBank-led funding round contingent on a corporate restructuring) has led to renegotiations over how much access Microsoft keeps and how its stake might be diluted. OpenAI recently deepened its tie-up with Oracle, planning 4.5 GW of new cloud capacity there, and started using Google’s cloud – moves that put pressure on Microsoft’s “AI edge” narrative. Still, many investors believe Microsoft holds leverage and will secure terms favoring its shareholders, as AI-driven optimism has already added over 20% to Microsoft’s stock this year. Why it matters: As OpenAI’s hunger for compute outgrows any single cloud, Microsoft must navigate a delicate balance: protecting its early advantage from the OpenAI partnership while the startup courts other backers and cloud providers – a sign that the AI boom is straining even the most closely aligned alliances.
Source: Reuters
Freshworks ups revenue outlook as businesses flock to its AI tools
Customer-software maker Freshworks raised its 2025 annual revenue forecast to ~$823–829 million, a slight increase, citing strong demand for its new AI-powered features. The California-based company – whose helpdesk and CRM products compete with Salesforce and ServiceNow – said over 5,000 customers now pay for its “Freddy AI” solutions that automate customer support queries and other routine tasks. Freshworks beat second-quarter expectations with $204.7M revenue (vs. $198.8M est.) and improved profitability, attributing the performance in part to clients adopting AI to onboard employees and serve users more efficiently. Government agencies are also using its AI software, with Freshworks’ CEO noting it has over 1,000 public-sector entities on its platform as it pushes into that market. Why it matters: Freshworks’ modest forecast bump shows that even mid-sized enterprise software firms are getting a measurable lift from the AI wave – validating the idea that embedding AI into business tools can drive revenue growth and competitive momentum against bigger rivals.
Source: Reuters
Meta’s billion-dollar AI gambits weigh on profits amid talent war
Meta (Facebook’s parent) has spent lavishly over the past year on what CEO Mark Zuckerberg calls the race for "superintelligence" – poaching top researchers (including paying $14.3 billion for a stake in data-labeling firm Scale AI and hiring its CEO) and pledging hundreds of billions for new AI data centers. Despite these bold bets, Meta’s core financials are slowing: Q2 profit was expected to rise just ~11.5%, the weakest in two years, as costs jumped ~9%. While investors have cheered Meta’s AI vision (its stock is up over 20% this year) and support Zuckerberg’s push to open-source AI research, there is skepticism about when these investments – including a new “Superintelligence Lab” alongside Meta’s AI lab – will pay off. Meta’s advertising business also faces headwinds from an uncertain economy and competition (e.g. TikTok), so questions loom over how its expensive AI quest fits with its long-term strategy and shareholder returns. Why it matters: Meta’s aggressive AI expansion underscores both Big Tech’s commitment to leading on next-gen AI (even at enormous cost) and the market’s patience for now – but it also highlights the gamble involved, as Meta must prove its pricey pursuit of “superintelligence” can eventually translate into consumer products or financial gains.
Source: Reuters
U.S. lets Nvidia send AI chips to China to avoid spurring Chinese chipmaking
President Trump’s economic adviser Kevin Hassett revealed that the administration decided to allow Nvidia to proceed with sales of its H20 AI chips to China – chips that are slightly less powerful than those sold elsewhere due to prior export limits. Nvidia had applied for U.S. licenses to resume these shipments and was assured it will get them, after earlier rules (from the Biden era) curtailed exports of advanced AI semiconductors to China. Hassett said on Fox News that if the U.S. blocked Nvidia completely, China might accelerate developing its own advanced chips, so letting Nvidia sell a constrained version was intended to prevent Beijing from “jumping ahead” in the chip race. The comments signal a shift in U.S. strategy: rather than blanket bans, the administration aims to maintain a technological edge by keeping China dependent on U.S. chip suppliers (even with trimmed-down products). Why it matters: This move reveals a calibrated U.S. approach to AI tech exports – easing some chip restrictions to keep American firms like Nvidia in China’s market, thereby slowing China’s incentive to develop indigenous chips, all while ostensibly preserving a technological gap. It’s a significant adjustment in the U.S.–China tech rivalry over AI hardware.
Source: Reuters
July 30, 2025
Voice actors in Europe urge AI dubbing rules to save their jobs
Veteran voice actors across Europe are lobbying regulators as AI-generated voices begin threatening the dubbing industry. Performers in France, Germany, Italy and elsewhere have launched campaigns (like France’s “Touche Pas Ma VF” and a German TikTok that went viral with 8.7M views) calling for laws to prevent AI from cloning voices without consent and to ensure human dubbers are compensated and credited. They warn that streaming platforms’ experiments with AI – for example, a Polish series on Viaplay was partly AI-dubbed (it had to be pulled after viewers panned its flat delivery) – could sacrifice quality for cost savings. Industry giants like Netflix admit to testing AI for tasks like matching lip movements, but so far still rely on human voice actors in the loop; the actors argue dubbing is an art requiring emotional nuance that AI alone can’t achieve. Why it matters: This is a flashpoint in creative industries: voice actors are effectively fighting for their livelihood against generative AI replacements. Their push for EU legislation (on consent, labeling, and payment for AI-generated performances) highlights broader cultural and ethical debates about protecting human creativity in the age of cheap synthetic media.
Source: Reuters
Google to endorse EU’s voluntary AI code despite qualms over rules
Alphabet’s Google confirmed it will sign onto the European Union’s new AI Code of Practice – a voluntary set of guidelines intended to help companies comply with the upcoming EU AI Act – even as it expressed reservations about the regulations. Kent Walker, Google’s global affairs chief, said in a blog post that Google supports the code’s aims for safe, high-quality AI but is concerned that some AI Act provisions (like stricter copyright rules, trade-secret disclosures and slower approvals) could hinder Europe’s AI development and competitiveness. The EU’s code covers transparency, copyright, and safety/security for AI systems; signing it gives firms some legal clarity ahead of the binding law. Microsoft has signaled it will likely sign as well, whereas Meta Platforms has publicly declined, arguing the code’s requirements overstep and create legal uncertainty for AI model builders. Why it matters: Google’s move to sign the EU code – while simultaneously warning that Europe’s AI Act might over-regulate – reflects how tech giants are navigating Europe’s assertive tech governance. It shows a willingness to cooperate on AI safety and transparency, but also foreshadows tension between innovation and regulation as the world’s first comprehensive AI law looms.
Source: Reuters
Palo Alto Networks to buy CyberArk for $25B, bolstering AI-era cyber defenses
U.S. cybersecurity firm Palo Alto Networks agreed to acquire Israel’s CyberArk Software for roughly $25 billion in cash and stock, marking Palo Alto’s largest-ever deal. The takeover expands Palo Alto’s offerings into identity and access management – crucial for countering AI-driven cyber threats – at a time when companies seek one-stop security suites after high-profile breaches. The deal follows a consolidation trend in the sector (Google’s $32B purchase of cloud security firm Wiz in March) as rising ransomware and data theft attacks, some aided by AI, drive demand for integrated defenses. Palo Alto’s CEO Nikesh Arora said the “explosion of machine identities” and AI risks mean every user’s privileges must be tightly managed. Palo Alto shares fell ~8% on the news amid investor concern over digesting such a large merger, since the company usually buys much smaller firms. Why it matters: This is one of the largest cybersecurity acquisitions on record, and it underscores how the advent of AI – which can both supercharge cyberattacks and require new protections – is reshaping the industry. Major players are combining forces to offer comprehensive, AI-resistant security platforms to big customers, even if it means hefty price tags and short-term investor skittishness.
Source: Reuters
Italy opens antitrust probe into Meta’s AI assistant on WhatsApp
Italy’s competition authority (AGCM) has launched an investigation into Meta Platforms over potential abuse of dominance by integrating its new AI chatbot into WhatsApp without user consent. Regulators raided Meta’s Italian offices, concerned that bundling “Meta AI” with WhatsApp – the country’s most popular messaging service – could unfairly steer millions of users to Meta’s AI and hinder rival AI assistants. Meta began rolling out the chatbot on WhatsApp in March 2025, inserting it into the app’s interface (the search bar) for Italian users. Meta argues the AI feature, offered free, “benefits customers” by giving them optional new services in a familiar app. The probe is being coordinated with EU authorities and could lead to fines up to 10% of Meta’s global turnover if wrongdoing is found. Why it matters: This case shows how regulators are scrutinizing Big Tech’s deployment of AI for competition harms. Even as companies rush to embed AI in their platforms, authorities in Europe are prepared to intervene if dominant firms leverage their user base to crowd out competitors – setting a precedent for balancing AI innovation with antitrust principles.
Source: Reuters
July 31, 2025
Elon Musk’s xAI to sign onto EU’s AI safety code (but not all of it)
Musk’s new AI startup xAI announced it will endorse the “Safety and Security” chapter of the European Union’s voluntary AI Code of Practice, a set of guidelines for complying with the coming AI Act. By signing that section, xAI aligns itself with EU principles on AI safety, but the company pointedly criticized other parts of the Code and Act as “profoundly detrimental to innovation,” especially provisions on copyright and transparency. It’s unclear if xAI will sign the remaining chapters (Transparency and Copyright) of the EU code; the firm’s post on Musk’s X platform praised the code’s safety intent but blasted its broader requirements. Google and likely Microsoft have agreed to the full code, while Meta has refused to sign any of it, citing legal uncertainties and overreach in the EU’s approach. Why it matters: Musk’s selective participation highlights the tech industry’s split stance on regulation: even avowed "AI safety" proponents may balk at concrete rules they see as too strict. xAI’s move is largely symbolic – it gets some goodwill by signing the safety pledge while joining other U.S. firms in pushing back on the tougher transparency and IP mandates in Europe’s landmark AI law.
Source: Reuters
August 1, 2025
France’s Mistral AI seeks $1B funding at massive $10B valuation – FT
French AI startup Mistral is in talks with venture capital investors and Abu Dhabi’s MGX fund to raise about $1 billion in new financing that would value the company at roughly $10 billion, according to the Financial Times. Founded just last year, Mistral in June released what it called Europe’s first AI reasoning model as it races to keep up with U.S. and Chinese AI labs. The huge prospective funding would help commercialize Mistral’s chatbot (dubbed “Le Chat”) and further develop its large language models. Mistral, which previously raised €600M at a €5.8B valuation, is seen by some analysts as Europe’s best chance to produce an AI leader on par with OpenAI – though it hasn’t yet achieved comparable scale or revenue. Neither Mistral nor MGX commented on the report. Why it matters: The sheer size of the mooted deal signals investor confidence that even relatively young AI startups can be worth tens of billions if they show promise. It also reflects Europe’s urgency to back its own AI champions in a field currently dominated by US and Chinese players.
Source: Reuters
Meta offloads $2B in data center assets to share soaring AI costs
Meta Platforms revealed plans to sell about $2.04 billion worth of its data center assets to third-party partners as part of a strategy to co-finance the huge infrastructure build-out needed for AI. In a quarterly filing, Meta said it reclassified land and construction projects as “held-for-sale” and aims to contribute them to an outside investor within 12 months to jointly develop new data centers. The move marks a shift for tech giants, which historically funded their own growth – Meta’s CFO noted that while the company will still primarily self-fund, some projects could get “significant external financing” to add flexibility as its AI compute needs evolve. Meta’s capital expenditures for 2024 are now projected at $66–72 billion (up $2B from prior guidance), as CEO Mark Zuckerberg embarks on building AI “supercluster” facilities he says could each span a large chunk of Manhattan. Why it matters: Even cash-rich Meta is feeling the pinch from the astronomical costs of AI hardware. By effectively mortgaging part of its data center expansion, Meta is acknowledging that the race to build AI supercomputers is financially straining – a signal that the industry is entering a phase where partnerships and creative financing replace the old ‘all in-house’ model to support AI growth.
Source: Reuters
Delta Air tells Congress it won’t use AI for personalized ticket pricing
Delta Air Lines assured U.S. lawmakers that it has “no fare product” that sets individualized ticket prices based on personal data and that it has never used AI to do so. The airline was responding to Senators who voiced fears that AI could let airlines charge each passenger the maximum they’d tolerate – so-called “surveillance pricing” that exploits personal info like browsing history or urgent needs. Delta acknowledged working with an AI pricing startup (Fetcherr) to optimize fares on 20% of its routes by year-end, but stressed this system uses aggregate market conditions, not buyer-specific targeting. Executives from Delta and rival American Airlines have publicly rejected the idea of personalized airfare pricing, saying it would erode customer trust and is not something they will pursue. Meanwhile, some in Congress have introduced a bill to preemptively ban AI-driven price discrimination in airlines and other industries. Why it matters: As AI tools spread in commerce, regulators are on high alert for unfair practices. Delta’s pledge not to use AI for personal pricing – and lawmakers pushing legislation – show a proactive effort to draw ethical lines around AI’s use in consumer markets, aiming to prevent “Big Brother” pricing before it starts.
Source: Reuters
Apple ready to spend big – even acquire – to catch up in AI, Tim Cook says
On Apple’s earnings call, CEO Tim Cook signaled a strategic shift: Apple will “increase spending” on artificial intelligence, including building more of its own data centers and even potentially buying larger AI companies if needed. This marks a departure from Apple’s usual frugal, mostly in-house approach – historically it has only made small acquisitions and relied on third-party clouds for some services. Cook noted Apple acquired seven tiny AI firms so far this year and is “very open” to bigger mergers to accelerate its AI roadmap. The urgency comes as Apple has lagged rivals Google and Microsoft, whose AI chatbots and assistants have attracted hundreds of millions of users (albeit with huge spending – Google plans $85B and Microsoft over $100B on AI/data centers). Apple is reportedly exploring an AI revamp of its Safari browser and has delayed improving Siri until 2026 as it works on more advanced models, following a partnership with OpenAI for some iPhone features. Why it matters: Apple’s acknowledgement that it must pour significantly more resources into AI – and its rare openness to large acquisitions – underscores how critical AI has become in tech competition. The company that famously prioritized privacy and incremental innovation is effectively saying it’s willing to spend and even break tradition (with major M&A) to avoid falling behind in the AI era.
Source: Reuters
Reddit stock jumps 15% as AI-powered ads drive surprise revenue growth
Shares of social media site Reddit soared after it reported quarterly results that beat expectations and issued an upbeat outlook, crediting its AI-driven advertising tools for much of the success. Reddit’s Q3 revenue forecast of $535–$545 million handily topped analysts’ $474M estimate. Executives said advertisers are flocking to platforms that use artificial intelligence to target ads more effectively and craft personalized campaigns – an approach Reddit has embraced by leveraging insights from its niche community discussions. The company saw daily active users climb 21%, and analysts noted that features like ads embedded within subreddit conversations are resonating with marketers, showing resilience in ad spending despite economic worries. Reddit’s strong performance mirrored positive results from Meta and Google, suggesting the digital ad market is being buoyed by new AI capabilities in content targeting. Why it matters: Reddit’s rally indicates that AI personalization isn’t just a buzzword – it’s translating into real financial gains. By using AI to better match ads to its passionate user communities, Reddit is boosting revenue and investor confidence, illustrating how even a mid-tier platform can thrive by harnessing AI where it uniquely excels (in this case, mining authentic user conversations for ad insights).
Source: Reuters
McDonald’s to expand AI from drive-thrus to global operations by 2027
Fast-food giant McDonald’s plans to “double down” on AI investment through 2027, and is ramping up its tech hubs in markets like India to support that push. One major initiative is rolling out AI order verification – currently used in 400 U.S. restaurants to catch mistakes before food is handed over – to all 40,000+ McDonald’s locations worldwide in the next few years. McDonald’s is also using AI for demand forecasting, dynamic pricing decisions, product performance analysis, and building a personalized global mobile app experience for customers. A senior executive said it’s early to quantify the investment but emphasized spending will go into technology and tools rather than a huge headcount, with India’s new AI center (and another planned in Poland) focusing on data governance, engineering and platform development. Why it matters: McDonald’s strategy shows that AI is not just for Silicon Valley – large traditional companies are betting big on it to streamline operations and boost sales. By automating order accuracy and tailoring menus and promotions through AI, McDonald’s aims to improve efficiency and customer experience at massive scale, signaling AI’s transformative potential even in the low-margin, high-volume world of fast food.
Source: Reuters