AIreview

The daily pulse of AI, technology, and innovation

ΕΛ

Subscribe

Category: Core

The technology layer behind AI — infrastructure, APIs, agents, MLOps, and developer tools

  • Investors Are Betting Big on the Energy Transition

    Investors Are Betting Big on the Energy Transition

    Despite political headwinds in the United States, global investors are signaling strong confidence that the shift toward clean energy is here to stay. Two major funds announced this week underscore that sentiment, with a combined $21 billion committed to projects and companies driving the transition.

    Brookfield revealed it has raised $20 billion for its second energy transition fund, 33% more than its first in 2021. The firm has already deployed $5 billion into renewable power developers focused on solar, wind, and battery storage. Raising a larger fund in a less exuberant economic climate suggests investors see long-term value in the sector rather than a passing trend.

    Energy Impact Partners also closed its third flagship fund at $1.36 billion, about 40% larger than its predecessor. The venture investor, known for backing growth-stage climate companies, typically participates in rounds averaging $26 million. Recent investments include GridBeyond, which manages distributed energy resources, and Quilt, a manufacturer of residential heat pumps.

    The broader picture supports this optimism. Since 2014, institutional investors such as pension funds and endowments have allocated nearly $1 trillion to the energy transition. Climate tech venture funds, while smaller in scale, are outpacing the rest of the venture industry. They now account for 3.8% of all venture capital raised globally, nearly double their share from 2020.

    Political uncertainty in the U.S. remains a drag. Proposed policy reversals threaten tax credits and grants, and the International Energy Agency has cut its forecast for U.S. renewable deployment by 45% through 2030. Even so, global renewable capacity is expected to double by the end of the decade, driven by strong growth in China, India, Europe, and Sub-Saharan Africa.

    Analysts at DNV project that renewables will supply 65% of global electricity by 2040 and almost all of it by 2060. It may fall short of achieving net-zero by 2050, but the momentum clearly points toward continued expansion. The short-term turbulence in the U.S. does little to alter the global trajectory: investors are still betting that the energy transition isn’t going away.

  • Intel Unveils First PC Chip Built on Its New 18A Manufacturing Process

    Intel Unveils First PC Chip Built on Its New 18A Manufacturing Process

    Intel has disclosed new technical details about its upcoming Panther Lake processor — the first laptop chip built using its next-generation 18A fabrication process. The announcement marks a critical milestone in Intel’s ambitious effort to regain its leadership position in semiconductor manufacturing.

    Targeted at high-performance, AI-enabled laptops, Panther Lake represents a major test of Intel’s ability to successfully scale the 18A process and recover market share lost to AMD.

    According to the company, the integrated graphics and CPU cores in Panther Lake deliver up to 50% faster performance compared to the previous Lunar Lake generation, which was largely produced by Taiwan Semiconductor Manufacturing Company (TSMC). The 18A process introduces a new transistor design and a more efficient power delivery system, enhancing performance and energy efficiency.

    Panther Lake is built on a system-on-chip (SoC) architecture that merges the CPU, GPU, and other critical components into a single integrated circuit. Production will begin ramping up later this year, with first units expected before the end of 2025 and wider availability starting January 2026.

    High Stakes for Intel’s Comeback

    “Panther Lake is extremely important to Intel on many different levels,” said Bob O’Donnell, chief analyst at Technalysis Research. He added that the chip could serve as proof of Intel’s ongoing progress in semiconductor innovation and a demonstration of what its new fabs are capable of producing.

    Intel’s new CEO, Lip-Bu Tan, has scaled back the extensive manufacturing expansion initiated by his predecessor, Pat Gelsinger. In July, the company warned it would pause development of its future 14A process unless a customer committed to adopting it.

    After months of uncertainty — including public calls by former U.S. President Donald Trump for Tan’s resignation — Intel secured new investments from SoftBank Group and Nvidia. Following Tan’s meeting with Trump and White House officials, the U.S. government converted a planned CHIPS Act grant into a 9.9% equity stake in Intel, signaling renewed government backing.

    New Technologies as Catalysts for Innovation

    “The new technologies are catalysts for innovation across our business as we build a new Intel,” Tan said.

    Intel confirmed that its Fab 52 facility in Arizona is now fully operational and will begin high-volume 18A production later this year. The same facility will also produce the company’s next-generation Clearwater Forest server processors, expected to launch in the first half of 2026.

    Although Intel still trails Nvidia in the AI graphics processor market, it expects Clearwater Forest to help it gain ground in AI data centers, thanks to its power-efficient design and integration with the 18A process.

  • Gemini Enterprise: Google’s Vision for AI Agents in Every Workplace

    Gemini Enterprise: Google’s Vision for AI Agents in Every Workplace

    Google Cloud has unveiled Gemini Enterprise, a new platform described as “the new front door for AI in the workplace.”

    Presented during a virtual press briefing, Gemini Enterprise unifies Google’s Gemini AI models, both first- and third-party agents, and the core framework formerly known as Google Agentspace into a single intelligent environment. The goal is to make AI-driven agents accessible to all businesses, enabling the automation of complex workflows and productivity improvements across entire organisations.

    Thomas Kurian, CEO of Google Cloud, introduced the initiative by noting that enterprise customers are moving beyond building basic AI applications toward constructing full-fledged intelligent agents.

    Gemini Enterprise represents Google’s strategic response to this shift, combining the company’s complete AI ecosystem into a unified experience that allows both technical and non-technical users to create agents using a no-code interface.

    The platform is structured around six foundational components.
    At its core are the Gemini models, including the newly released Gemini 2.5 Flash Image. The “workbench”, derived from Agentspace, provides the tools for building, orchestrating, and managing agents that automate tasks. Supporting these are preconfigured tools known as the “taskforce” — Google’s collection of specialised agents such as Code Assist Agent and Deep Research Agent.

    To enhance performance, Gemini Enterprise integrates deeply with corporate data systems using new connectors compatible with Microsoft Teams, Salesforce, Box, Confluence, and Jira. Kurian highlighted that the platform “remembers who you are and what you do,” allowing it to personalise interactions and maintain context across sessions with large language models.

    A unified governance framework lets organisations monitor, secure, and audit all agents from a central console, with built-in safeguards like Model Armor. The open ecosystem, encompassing over 100,000 partners, ensures broad compatibility and extensibility.

    Redefining Workplace Productivity

    To showcase Gemini Enterprise’s potential, Maryam Gholami demonstrated a practical enterprise use case.

    She emphasised that the system maintains the familiar Gemini interface, adapted for corporate workflows and offering full administrative control over data sources.

    Using a custom campaigns agent, she combined four AI agents to conduct market research, generate media content, manage internal communications, and oversee inventory. The system recognised a growing consumer trend toward sci-fi themes, detected a 25% stock shortfall, initiated a purchase order via ServiceNow, composed an email to store managers, and generated social media visuals — all within minutes.

    “Gemini Enterprise is not just a chat interface,” Gholami concluded. “It’s a comprehensive AI solution that unites your data, tools, and teams to compress weeks of work into a single, cohesive workflow.”

    Real-World Adoption: AI Fleets in Action

    Demonstrating real-world impact, Nirmal Saverimuttu, CEO of Virgin Voyages, shared insights from his company’s AI adoption journey. He emphasised that meaningful AI integration requires a cultural transformation, not just technology deployment.

    “AI will never replace our people,” he said. “It’s about empowering them and amplifying their potential.”

    Virgin Voyages now operates a fleet of over 50 specialised AI agents. One early example, nicknamed ‘Email Ellie’, increased content production speed by 40% and helped drive a 28% year-over-year sales rise in July. The company also reported a 35% reduction in agency dependency costs, granting greater creative autonomy.

    Another prominent adopter, Macquarie Bank, has extended Gemini Enterprise to its entire workforce. The Australian financial institution reports that 99% of employees have already completed training in generative AI tools.

    Google underscores that Gemini Enterprise is an open and extensible platform. Partners like Box, Salesforce, and ServiceNow have already launched compatible agents. An AI agent directory will soon allow customers to browse thousands of verified partner-built solutions.

    To accelerate learning, Google introduced Google Skills, a free training platform with 3,000 courses. Complementing this is the Gemini Enterprise Agent Ready (GEAR) program — an intensive initiative aimed at enabling one million developers to design and deploy AI agents.

    Availability and Pricing

    Gemini Enterprise is now globally available wherever Google Cloud services are sold.

    Pricing tiers include Gemini Business for small organisations at $21 per seat per month, while the Gemini Enterprise Standard and Plus plans for larger companies begin at $30 per seat per month.

    According to Kurian, the launch represents a major step toward democratising AI access:
    “Gemini Enterprise reimagines powerful AI technology for the workplace, making it easy to use and putting it directly into the hands of every company and employee.”