Yazılar

Oracle and Google Cloud Partner to Offer Gemini AI Models via Oracle Cloud

/in /tarafından

Oracle (ORCL.N) and Alphabet (GOOGL.O) announced on Thursday that their cloud computing divisions have reached an agreement to make Google’s Gemini artificial intelligence models available through Oracle’s cloud services and business applications. The collaboration allows developers to leverage Google’s models for generating text, video, images, and audio.

The deal extends to businesses using Oracle applications for corporate finance, human resources, and supply chain management, enabling them to integrate Google’s AI capabilities directly into these tools. While the AI models will run on Google’s servers, Oracle will integrate access through Google’s Vertex AI platform, allowing customers to use Oracle cloud credits to pay for the services. Financial terms between the companies were not disclosed.

This partnership follows a similar arrangement Oracle made with Elon Musk’s xAI in June and aligns with Oracle’s strategy to offer customers multiple AI options instead of focusing solely on its own technology. For Google, the collaboration broadens the reach of its cloud services and aims to attract corporate clients from competitors like Microsoft (MSFT.O).

AI Startup Cohere Valued at $6.8 Billion in Latest Fundraising, Hires Meta Exec

/in /tarafından

Canadian AI startup Cohere raised $500 million in its latest funding round, pushing its valuation to $6.8 billion as it targets enterprise AI expansion. The round was led by Radical Ventures and Inovia Capital, with participation from AMD Ventures, Nvidia (NVDA.O), PSP Investments, Salesforce Ventures, and other existing investors.

Unlike broad AI models from OpenAI or Meta’s Llama, Cohere focuses on enterprise-specific AI solutions. Co-founder Nick Frosst said the new funding will help the company expand globally, explore different AI modalities—including its recently launched Command Vision model—and continue building secure AI tools for businesses.

Cohere also appointed two senior executives: Joelle Pineau, former VP of AI Research at Meta (META.O), as chief AI officer, and Francois Chadwick, ex-Uber and Shield AI executive, as chief financial officer. Pineau led Meta’s Fundamental AI Research group until May 2025.

Earlier this year, Cohere launched North, a ChatGPT-style tool designed to assist knowledge workers with tasks like document summarization. The company plans to use the funding to develop agentic AI aimed at improving operational efficiency for businesses and government agencies.

The fundraising comes amid a broader surge in AI investment, with private equity firms and Big Tech pouring capital into startups to capture returns from innovative AI technologies.

OpenAI Launches GPT-5 Amid Industry Push for AI Returns on Investment

/in /tarafından

OpenAI unveiled GPT-5 on Thursday, the latest iteration of its influential AI technology powering ChatGPT, available now to all 700 million users of the chatbot platform. This launch marks a key moment as the AI sector seeks to justify massive investments by demonstrating clear enterprise value.

While consumer enthusiasm remains strong, with users captivated by ChatGPT’s conversational capabilities, business spending on AI has yet to fully materialize. The industry’s top players—Alphabet, Meta, Amazon, and Microsoft (which backs OpenAI)—are collectively spending nearly $400 billion this fiscal year on AI infrastructure, raising expectations for significant returns.

OpenAI is currently in talks to let employees cash out at a $500 billion valuation, up from $300 billion. The company has attracted high-profile talent with signing bonuses reaching $100 million, underscoring intense competition for AI expertise.

CEO Sam Altman emphasized GPT-5’s advancements in software development, finance, and health queries, describing it as capable of expert-level responses comparable to PhD-level knowledge. A standout feature demonstrated was “vibe coding”—the ability to generate fully functioning software from natural language prompts, highlighting a new era of software on demand.

However, early reviewers suggest that the leap from GPT-4 to GPT-5, while impressive, may not be as large as previous model upgrades. Altman acknowledged GPT-5 still cannot autonomously learn, a critical step toward matching human-like intelligence.

A new capability called “test-time compute” enables GPT-5 to spend extra computational effort on harder problems, improving its reasoning and problem-solving. This technology is now publicly accessible for the first time and is central to OpenAI’s mission to build broadly beneficial AI.

Despite GPT-5’s launch, OpenAI faces challenges, including a shortage of new training data and the complexity and cost of running large-scale AI training. The company sees infrastructure expansion worldwide as essential for broader AI adoption.

Altman remains optimistic about AI’s future impact but stresses that current investments in infrastructure are still insufficient to realize its full potential globally.