Tech Patrol

Etiket: artificial intelligence

  • # AI Investment Boom: US Startups Attract Billions in First Half of 2025

    ## AI Investment Boom: US Startups Attract Billions in First Half of 2025

    The artificial intelligence sector continues to be a magnet for investment, with U.S.-based AI startups securing substantial funding rounds in the first half of 2025. Building on the momentum of a record-breaking 2024, which saw 49 startups raise $100 million or more, 2025 is shaping up to be another landmark year. Multiple billion-dollar rounds have already closed, and the total number of AI megarounds in the U.S. during Q1 2025 surpassed that of Q1 2024, signaling sustained investor confidence.

    Here’s a breakdown of the U.S. AI companies that have raised $100 million or more so far in 2025, categorized by month:

    **April**

    * **SandboxAQ:** This AI model company closed a $450 million Series E round on April 4, valuing it at $5.7 billion. Notable investors included Nvidia, Google, and Ray Dalio, founder of Bridgewater Associates.
    * **Runway:** Known for its AI models for media production, Runway secured a $308 million Series D round announced on April 3. The round, led by General Atlantic, valued the company at $3 billion, with participation from SoftBank, Nvidia, and Fidelity.

    **March**

    * **OpenAI:** Leading the pack by a significant margin, OpenAI raised a staggering $40 billion funding round on March 31, valuing the startup at $300 billion. SoftBank led the round, with contributions from Thrive Capital, Microsoft, and Coatue, among others.
    * **Nexthop AI:** This AI infrastructure company announced a $110 million Series A round on March 25, led by Lightspeed Venture Partners. Kleiner Perkins, Battery Ventures, and Emergent Ventures also participated.
    * **Insilico Medicine:** Based in Cambridge, Massachusetts, Insilico Medicine raised $110 million for its generative AI-powered drug discovery platform, announced on March 13. The Series E round, co-led by Value Partners and Pudong Chuangtou, valued the company at $1 billion.
    * **Celestial AI:** Another AI infrastructure company, Celestial AI, closed a $250 million Series C round on March 11, valuing the company at $2.5 billion. Fidelity led the round, with participation from Tiger Global, BlackRock, and Intel CEO Lip-Bu Tan.
    * **Lila Sciences:** Aiming to create a science superintelligence platform, Lila Sciences raised a substantial $200 million seed round led by Flagship Pioneering. Other investors included March Capital, General Catalyst, and ARK Venture Fund. The company is based in Cambridge, Massachusetts.
    * **Reflection.Ai:** This Brooklyn-based startup, focused on building superintelligent autonomous systems, secured a $130 million Series A round. The round, led by Lightspeed Venture Partners and CRV, valued the company at $580 million, impressive for a company only a year old.
    * **Turing:** An AI coding startup that partners with LLM companies, Turing closed a $111 million Series E round on March 7, valuing it at $2.2 billion. Khazanah Nasional led the round, with participation from WestBridge Capital, Gaingels, and Sozo Ventures.
    * **Shield AI:** This AI defense tech startup raised $240 million in a Series F round that closed on March 6. Co-led by L3Harris Technologies and Hanwha Aerospace, with participation from Andreessen Horowitz and the US Innovative Technology Fund, the round valued Shield AI at $5.3 billion.
    * **Anthropic:** An AI research and large language model company, Anthropic, raised $3.5 billion in a Series E round announced on March 3. The round, led by Lightspeed with participation from Salesforce Ventures, Menlo Ventures, and General Catalyst, valued the startup at $61.5 billion.

    **February**

    * **Together AI:** Focused on open-source generative AI and AI model development infrastructure, Together AI raised a $305 million Series B round on February 20, valuing the company at $3.3 billion. Prosperity7 and General Catalyst co-led the round, with participation from Salesforce Ventures, Nvidia, and Lux Capital.
    * **Lambda:** This AI infrastructure company announced a $480 million Series D round on February 19, valuing the startup at nearly $2.5 billion. The round was co-led by SGW and Andra Capital, with participation from Nvidia, G Squared, and ARK Invest.
    * **Abridge:** An AI platform that transcribes patient-clinician conversations, Abridge was valued at $2.75 billion in a Series D round announced on February 17. The $250 million round was co-led by IVP and Elad Gil, with participation from Lightspeed, Redpoint, and Spark Capital.
    * **Eudia:** An AI legal tech company, Eudia, raised $105 million in a Series A round led by General Catalyst, closing on February 13. Floodgate, Defy Ventures, and Everywhere Ventures, along with numerous angel investors, also participated.
    * **EnCharge AI:** Focused on AI hardware, EnCharge AI raised a $100 million Series B round, also closing on February 13. Tiger Global led the round, with participation from Scout Ventures, Samsung Ventures, and RTX Ventures. The Santa Clara-based company was founded in 2022.
    * **Harvey:** Another AI legal tech company, Harvey, raised a $300 million Series D round announced on February 12, valuing the 3-year-old company at $3 billion. Sequoia led the round, with participation from OpenAI Startup Fund, Kleiner Perkins, and Elad Gil.

    **January**

    * **ElevenLabs:** This synthetic voice startup raised a $180 million Series C round announced on January 30, valuing the company at more than $3 billion. ICONIQ Growth and Andreessen Horowitz co-led the round, with participation from Sequoia, NEA, and Salesforce Ventures.
    * **Hippocratic AI:** Developing large language models for the healthcare industry, Hippocratic AI announced a $141 million Series B round on January 9. This round, led by Kleiner Perkins with participation from Andreessen Horowitz, Nvidia, and General Catalyst, valued the company at more than $1.6 billion.

    The substantial influx of capital into these AI startups highlights the ongoing excitement and potential within the industry. From infrastructure and healthcare to media production and defense, AI is attracting significant investment across diverse applications, suggesting a bright future for innovation in the field.

  • # OpenAI Aims for Top Spot with New ‘Open’ AI Model

    ## OpenAI Aims for Top Spot with New ‘Open’ AI Model

    OpenAI is planning to release its first truly “open” language model since GPT-2 later this year, and new details are emerging about the company’s ambitions for the project. According to sources within the AI developer community, OpenAI is targeting an early summer release for the model, spearheaded by VP of Research Aidan Clark. The company’s goal is nothing short of creating the best-in-class open reasoning model, surpassing existing open alternatives.

    This new model will reportedly be a “text in, text out” system, designed to run on high-end consumer hardware. Developers may also have the option to toggle the model’s “reasoning” capabilities on and off, similar to features seen in models recently released by Anthropic. While reasoning enhances accuracy, it can also increase latency, offering users a trade-off between speed and precision.

    Furthermore, OpenAI is considering a highly permissive license for the model, aiming to avoid the “onerous requirements” that have drawn criticism toward other open models like Llama and Google’s Gemma. This move signals a potential shift in OpenAI’s approach to open source, especially as the company faces increasing competition from rivals like Chinese AI lab DeepSeek, which have successfully leveraged open strategies to build large user bases. Meta’s Llama models, for example, have amassed over 1 billion downloads.

    OpenAI CEO Sam Altman has publicly acknowledged that the company may have been on the “wrong side of history” regarding open source, suggesting a renewed commitment to a more open approach. This shift includes plans for thorough red-teaming and safety evaluations of the upcoming model, with the release of a detailed model card outlining internal and external benchmarking and safety testing results.

    Altman emphasized the importance of safety, stating that the model will be evaluated according to OpenAI’s preparedness framework, with extra attention given to its post-release modifiability. This focus on safety comes after criticism surrounding the reported rushing of safety testing on recent models and the failure to release model cards, underscoring OpenAI’s intent to prioritize responsible AI development with its new open model. The company was contacted for comment but has yet to respond.

  • # ChatGPT’nin Cevapları Artık Washington Post Makalelerini İçerecek

    ## ChatGPT’nin Cevapları Artık Washington Post Makalelerini İçerecek

    OpenAI ve The Washington Post, ChatGPT’nin cevaplarında Post’un orijinal haberlerini özetleyip bağlantı verecek yeni bir içerik ortaklığını duyurdu. Bu, yapay zeka devinin The Guardian ve Axios gibi yayınlar da dahil olmak üzere şimdiye kadar 20’den fazla haber yayıncısıyla yaptığı anlaşmalarla OpenAI’nin en son medya ortaklığı.

    Washington Post, 500 milyondan fazla kullanıcısı olan sohbet robotunun geniş kitlesinden yararlanacağını söylüyor. Bu arada OpenAI, basın açıklamasına göre Post’un “zamanında, iyi kaynaklı haberlerine” dayalı daha yüksek kaliteli cevaplardan yararlanacak.

    Anlaşmanın mali şartları açıklanmadı ve Washington Post sorulduğunda paylaşmayı reddetti. OpenAI ise bir yorum talebine hemen yanıt vermedi.

    OpenAI bazı haber merkezleri tarafından benimsendi, ancak hepsi değil: Örneğin The New York Times, telif hakkıyla korunan çalışmalarını kullandığı iddiasıyla OpenAI’ye dava açıyor; OpenAI ise bu suçlamayı reddediyor.

  • # Washington Post Joins Forces with OpenAI to Power ChatGPT Responses

    ## Washington Post Joins Forces with OpenAI to Power ChatGPT Responses

    ChatGPT users can now expect more informed and comprehensive answers, thanks to a new partnership between OpenAI and The Washington Post. The collaboration, announced today, will integrate the Post’s original reporting into ChatGPT’s responses, offering summaries and links to the source articles.

    This move expands OpenAI’s growing network of media partnerships. The AI giant has already forged similar deals with over 20 news publishers, including prominent names like The Guardian and Axios. By incorporating high-quality journalism, OpenAI aims to improve the accuracy and reliability of ChatGPT’s output.

    The Washington Post anticipates significant benefits from tapping into ChatGPT’s massive user base, which currently exceeds 500 million. The partnership provides the Post with a new avenue to reach a wider audience and showcase its reporting. OpenAI, in turn, gains access to the Post’s extensive archive of “timely, well-sourced reporting,” enhancing the quality and depth of ChatGPT’s knowledge base.

    While the financial details of the agreement remain undisclosed, the potential impact on both organizations is substantial. The integration of trusted news sources could solidify ChatGPT’s position as a reliable source of information, while simultaneously driving traffic and readership to The Washington Post.

    This partnership arrives amidst ongoing debate about the use of copyrighted material in AI training. Notably, The New York Times is currently suing OpenAI, alleging the unauthorized use of its copyrighted content. OpenAI refutes these claims, highlighting the complex legal and ethical considerations surrounding AI development and content ownership in the media landscape. Despite this controversy, OpenAI continues to pursue partnerships like the one with The Washington Post, signaling its commitment to incorporating high-quality journalism into its AI models.

  • # Noxtua Secures $92M to Pioneer Sovereign AI for German Legal Landscape

    ## Noxtua Secures $92M to Pioneer Sovereign AI for German Legal Landscape

    Noxtua, formerly known as Xayn, has announced a successful $92.2 million (€81.2 million) Series B funding round, signaling a significant leap forward in its mission to deliver sovereign AI solutions tailored for the German legal system. The company, which started as a privacy-focused, on-device AI startup for smartphones back in 2020, pivoted to address the specific needs of the legal sector.

    The funding round was spearheaded by C.H. Beck, Germany’s leading legal publisher, a strategic investor with unparalleled access to the country’s vast legal database. This partnership will provide Noxtua with access to C.H. Beck’s extensive archive of legal cases, judgments, and news, fueling the development of their flagship legal AI product, Beck-Noxtua.

    Joining C.H. Beck as new investors are Northern Data Group, a specialist in high-performance computing, CMS, Germany’s largest business law firm, and global law firm Dentons. Existing investors like Global Brain, KDDI Open Innovation Fund, and Dominik Schiener also participated in the round.

    The involvement of Northern Data Group is particularly noteworthy. Beck-Noxtua will be deployed as a sovereign AI solution on Northern Data’s cloud infrastructure, ensuring that all data and processing remain within Germany. This addresses growing concerns about data sovereignty and security, especially when handling sensitive legal information.

    Noxtua asserts that its AI is uniquely positioned to research legal matters, analyze complex documents, and even draft legal content, all while adhering to Germany’s stringent legal compliance standards. The company emphasizes the critical role of C.H. Beck’s extensive training data, comprising 55 million documents – the largest legal database in the German-speaking world – in ensuring the accuracy and reliability of its AI.

    In a landscape increasingly shaped by geopolitical considerations, Noxtua’s decision to partner with Frankfurt-based Northern Data reinforces its commitment to keeping German legal AI infrastructure within German borders. This move reflects a growing trend towards “sovereign AI,” where nations seek to control and secure their AI infrastructure and data.

    Dr. Leif-Nissen Lundbæk, CEO and Co-Founder of Noxtua, explained to TechCrunch that the company utilizes its own custom transformer AI model, meticulously trained on legal contracts. “We’ve already rolled it out to a lot of law firms and legal departments and partnered with C.H. Beck,” he stated, drawing a parallel to “Thompson Reuters for law” in Germany.

    Lundbæk further highlighted the limitations of U.S.-based foundational models for the German legal context, noting that they are often trained on American and UK data, which differ significantly from German legal frameworks. “Those models really fail in precision,” he added. “Plus, government services are essentially legal, right? So you cannot just use an American AI model in a German legal context.”

    Noxtua’s technology stems from research initiated by its founders at the University of Oxford and Imperial College London, further refined in collaboration with CMS.

    Professor Dr. Klaus Weber, a member of the executive board at C.H. Beck, emphasized the alignment of values between the two companies: “Noxtua’s vision of a sovereign European legal AI aligns hand in hand with our values … Noxtua is a cornerstone of our innovation strategy.” The substantial Series B funding underscores the growing demand for specialized AI solutions that respect national legal frameworks and prioritize data sovereignty. Noxtua is now poised to become a key player in the evolution of legal technology in Germany and potentially beyond.

  • # Alman Hukuk Sistemine Özel Yapay Zeka: Noxtua 92 Milyon Dolar Yatırım Aldı

    ## Alman Hukuk Sistemine Özel Yapay Zeka: Noxtua 92 Milyon Dolar Yatırım Aldı

    Akıllı telefonlar için gizlilik odaklı, cihaz üzerinde çalışan yapay zeka çözümleri sunan Xayn, 2020’deki deneyimlerinin ardından hukuk alanına yönelik “bağımsız yapay zeka” geliştirmeye yöneldi. Şimdi ise Noxtua adıyla yeniden markalaşan şirket, Seri B yatırım turunda 92.2 milyon dolar (yaklaşık 81.2 milyon Euro) yatırım almayı başardı.

    Yatırım turuna Almanya’nın önde gelen hukuk yayıncısı ve stratejik yatırımcısı C.H. Beck liderlik etti. Bu yatırımın Noxtua için önemi büyük. C.H. Beck, Almanya’daki tüm hukuki dava ve kararların arşivi üzerinde neredeyse tekel sahibi. Bu sayede Noxtua, “Beck-Noxtua” adını taşıyan yeni hukuki yapay zeka ürünü için bu devasa arşive ve hukuki haber akışına erişim sağlayacak.

    Yeni yatırımcılar arasında yüksek performanslı bilişim uzmanı Northern Data Group, Almanya’nın en büyük ticari hukuk firması CMS ve küresel hukuk firması Dentons da yer alıyor. Global Brain, KDDI Open Innovation Fund ve Dominik Schiener gibi önceki yatırımcılar da bu turda yerlerini korudu.

    Northern Data’nın katılımı tesadüf değil. Beck-Noxtua, şirketin Almanya sınırları içindeki bulut altyapısı üzerinde “bağımsız yapay zeka” olarak çalışacak.

    Noxtua, son derece uzmanlaşmış yapay zekasının hukuki konuları araştırabildiğini, hukuki belgeleri analiz edip taslaklarını hazırlayabildiğini ve tüm bunları Almanya merkezli müşterileri de dahil olmak üzere hukuki mevzuata uygun bir şekilde yapabildiğini iddia ediyor. Almanya’daki hukuki uyumluluk standartlarının son derece yüksek olması, C.H. Beck’ten elde edilen eğitim verilerinin doğruluğu açısından kritik önem taşıyor. Bu veri seti, Almanca konuşulan dünyadaki en büyük hukuk veri tabanı olan 55 milyon belgeden oluşuyor.

    Jeopolitik gelişmelerin, Almanya merkezli yapay zeka modellerinin ABD merkezli altyapılar üzerinde çalıştırılması fikrine gölge düşürmesiyle birlikte Noxtua, Frankfurt merkezli Northern Data ile bir barındırma ortaklığına imza attı.

    Noxtua CEO’su ve Kurucu Ortağı Dr. Leif-Nissen Lundbæk, TechCrunch’a yaptığı açıklamada, Noxtua’nın kendi dönüştürücü yapay zeka modeli versiyonunu kullandığını, ancak bu modelin özellikle hukuki sözleşmeler üzerinde eğitildiğini belirtti. Lundbæk, “Zaten birçok hukuk firmasına ve hukuk departmanına sunduk ve Almanya’daki ‘hukukun Thompson Reuters’ı’ olan C.H. Beck ile ortaklık kurduk,” dedi.

    Lundbæk, ABD merkezli temel modellerin Amerikan verilerine, kısmen de İngiltere verilerine dayandığını ve bu nedenle Noxtua’nın bu yolu izlemek zorunda kaldığını söyledi.

    Lundbæk, “Almanya veya Fransa gibi ülkelerde bu durum çok farklı. Bu modellerin hassasiyeti gerçekten yetersiz kalıyor. Ayrıca, devlet hizmetleri esasen hukuki nitelikte, değil mi? Bu nedenle Alman hukuki bağlamında Amerikan yapay zeka modeli kullanamazsınız,” diye ekledi.

    Noxtua’nın teknolojisi, kurucular tarafından Oxford Üniversitesi ve Imperial College London’da yürütülen araştırmalar sonucunda geliştirildi ve daha sonra CMS ile birlikte geliştirilmeye devam edildi.

    C.H. Beck Yönetim Kurulu Üyesi Profesör Dr. Klaus Weber yaptığı açıklamada, “Noxtua’nın bağımsız bir Avrupa hukuki yapay zekası vizyonu, değerlerimizle örtüşüyor… Noxtua, inovasyon stratejimizin temel taşıdır,” dedi.