<div>
  <ul>
    <li>
      <p>
        EdgeRunner AI, a startup developing GenAI solutions for Edge devices, raised USD 5.5 million in a seed funding round led by Four Rivers Group, with participation from Madrona Ventures and strategic angels.
      </p>
    </li>
    <li>
      <p>
        The funding will be used to build EdgeRunner AI's platform, which will be powered by transparent models designed to operate on any device without internet access. This will enable enterprises to deploy domain-specific GenAI models.
      </p>
    </li>
    <li>
      <p>
        EdgeRunner AI aims to develop small, task-specific open models optimized for any chip, called Ultra-Efficient Language Models (UELMs), allowing organizations to run GenAI models locally for improved performance, data privacy, low latency, and reduced power consumption.
      </p>
    </li>
  </ul>
  <p>
  </p>
</div>


<div>
  <p>
    <a href="https://www.businesswire.com/news/home/20240718441467/en/EdgeRunner-AI-Emerges-from-Stealth-with-5.5M-Seed-Funding-to-Build-Generative-AI-for-the-Edge/">
      Press release
    </a>
  </p>
</div>


Generative AI Infrastructure

<div>
 <p>EdgeRunner AI, an emerging company developing GenAI solutions for Edge devices, has successfully garnered a total of 5.5 million USD in a seed funding round. This round was mainly supervised by Four Rivers Group and had added contributions from Madrona Ventures and other strategic angel investors.</p>
 <p>The funds raised will be channeled towards strengthening EdgeRunner AI's platform. The company plans on utilising transparent models that can function on any device regardless of internet availability. As such, businesses can deploy domain-specific GenAI models with convenience.</p>
 <p>EdgeRunner AI's goal is to create and implement smaller, task-oriented open models that can be optimised for any kind of chip, identified as Ultra-Efficient Language Models (UELMs). This would facilitate organizations in locally running GenAI models. This approach is likely to ramp up performance, reinforce data privacy, diminish latency, and lower power consumption.</p>
 
</div>

EdgeRunner AI raises USD 5.5 million in seed funding to build GenAI for Edge devices

Metaverse Experience Platforms

ANALYST_QUICK_TAKE

INCUMBENTS_PRESENT

PROMINENT_CATEGORY

<div>
 <p>Nintendo is creating their debut MMO (massively multiplayer online) game. A playtest for this upcoming game has been announced, with Nintendo Switch Online members being invited. The playtest, which hopes to include 10,000 participants, will run from October 23 to November 5, 2024.</p>
 <p>In the MMO game, players can collaboratively construct and explore an expansive planet. The game features discoveries of new territories, adversaries, and resources. Players will use "Beacons", special items that allow them to create and manage their building zones. Only Beacon owners can directly modify these areas. A social area named the Dev Core will allow for player interaction and progression.</p>
 <p>Nintendo has been constantly improving its Switch Online subscription service since its launch in 2018. This improved service provides exclusive access to a broader variety of games, including classic ones, and has accrued over 38 million paid subscribers globally.</p>
</div> <p><b>Analyst QuickTake:<br>
 </b> Nintendo is strategically developing its first MMO-style game to compete with popular titles such as <a href="https:<br>
//www.minecraft.net/en-us"> Minecraft </a> and <a href="https:<br>
//sp-edge.com/companies/22278"> Roblox, </a> which dominate the social gaming space. The move reflects Nintendo's desire to tap into a huge audience of school-age children and families increasingly drawn to interactive, user-generated content. By creating an expansive world for players to build and explore together, Nintendo is positioning itself to capture the attention of the millions of active users on these platforms.&nbsp;</p>

Nintendo ventures into MMOs with new Switch Online game playtest

Additive manufacturing makes objects by systematically adding material via a CAD program

Additive Manufacturing

<div>
 
 <p>HP has created a texture library for applying textures to certain surfaces of 3D-printed parts with its CoreTechnologie 4D Additive software. Users can download the library for free from the HP website and integrate it into the 4D Additive textures module conveniently.<p>
 <p>With over 100 surface patterns available in the library, users have the flexibility to modify the size, alignment, and height of each HP texture. </p>
 <p>4D Additive's textures module facilitates more accurate texture application since it works directly with the surfaces of CAD models, unlike STL models.</p>
 
</div>

HP releases texture library for 3D printing with CoreTechnologie 4D Additive

Using data and automation to achieve 24/7 production.

Smart Factory

<div>
 <p>Aerospace and industrial solutions provider Honeywell has teamed up with Google to incorporate Gemini, Google's genAI model, into its industrial applications. This partnership is intended to deliver AI-powered insights through Honeywell's datasets, which can potentially reduce maintenance costs, increase productivity, and enhance employee skills. The partnership will utilize Google Cloud's Vertex AI and large language models (LLMs) in conjunction with Honeywell's Forge IoT platform for creating AI agents trained on industrial data. The resulting AI solutions will become available to industrial customers by 2025.</p>
 <p>Honeywell is also looking into implementing Gemini Nano, an on-device AI version, for operations in diverse industrial environments, especially in areas with restricted internet availability.</p>
</div>

Honeywell partners with Google to bring Gemini AI to industrial sector

Foundation Models

<div>
 <p>Elon Musk's AI startup, xAI, has introduced a new API, currently in public beta, which allows third-party developers to build applications utilizing xAI's Grok LLMs. The pricing for the API, called "grok-beta," stands at USD 5 per million input tokens and USD 15 per million output tokens.</p>
 <p>The accessible features provided by the xAI API include Grok models that can generate text and code, conduct data extraction, summarization, and image analysis. It also has the capability to call functions, which lets the models communicate with external databases and APIs. Additional features are a usage explorer for monitoring consumption, simplified team management, and improved security measures.</p>
</div> <p><b>Analyst QuickTake</b> :<br>
 This launch marks xAI’s entry into the competitive models landscape dominated by companies like OpenAI and Anthropic. Adopting a similar business model focused on providing access to API positions xAI to leverage its expertise while competing directly with established players.</p>

xAI launches API for third-party developers to access Grok models

<div>
 <p>Nava Benefits, an employee benefits solutions provider, is working with OpenAI to unleash the Nava Benefits Assistant. It is an AI-enabled instrument for managing employee benefits.</p>
 <p>The tool proffers real-time updates regarding deductibles and out-of-pocket maximums, resolves inquiries about benefits and company rewards, and supplies information on medical, vision, and dental schemes. The assistant operates round the clock and includes privacy measures to preserve user anonymity. Its future capabilities would encompass aid in finding doctors within the network, scrutinizing bills, and support in selecting medical plans.</p>
 
</div>

Nava Benefits partners with OpenAI to launch AI-powered benefits assistant

<div>
 <p>IBM has introduced Granite 3.0, a series of AI models custom-made for enterprise usage. These models are covered by the Apache 2.0 license and are accessible via IBM's watsonx platform as well as partner platforms including Google Cloud, NVIDIA, Hugging Face, Ollama, and Replicate.</p>
 <p>Included in Granite 3.0 are general-purpose language models, guardrail models for enhanced safety, expert mixture models for reduced latency, and a speculative decoder for better inference speed. These models have been trained with over 12 trillion tokens spanning 12 natural and 116 programming languages. They are suited for tasks such as retrieval, augmented generation, classification, summarization, and entity extraction.</p>
 
</div> <p><b>Analyst QuickTake</b> :<br>
 IBM launched its open-source Granite models built for various code-related tasks <a href="https:<br>
//sp-edge.com/updates/29754"> in May 2024 </a> ; the older launch included a new class of watsonx assistants for enterprise Java applications and IBM Z systems, with code explanation and modernization functionalities. This new launch comes five months later as an AI designed for enterprise use. IBM claims that Granite 3.0 models outperform or match similarly sized models from other providers on many academic and industry benchmarks.</p>

IBM releases Granite 3.0 AI models for enterprise use

<div>
 <p>Global conglomerate Honeywell has partnered with Google to integrate Google's GenAI model, Gemini, into industrial applications. Their collaboration aims to offer AI-powered insights across Honeywell's dataset, reducing maintenance costs, enhancing productivity, and upskilling employees. The partnership will utilize Google Cloud's Vertex AI and LLMs along with Honeywell's Forge IoT platform to develop AI agents trained on industrial data. These AI solutions will be available to industrial clients starting from 2025.</p>
 <p>Honeywell is also considering the use of Gemini Nano, an on-device AI version, for operations in multiple industrial settings, especially in areas with limited internet connectivity.</p>
 
</div>

<div>
 <p>Inflection AI, a provider of enterprise AI systems, has acquired European Robotic Process Automation (RPA) solution provider, Boundaryless, for an undisclosed sum. This acquisition is expected to aid in the deployment of Inflection AI's agent capabilities within enterprises.</p>
 <p>Boundaryless offers a range of solutions including data transformation, automation, AI, and analytics to clients in various sectors such as finance, energy, healthcare, and manufacturing.</p>
 <p>Post acquisition, Inflection AI is set to open its first European office in Switzerland, where Boundaryless is based. The larger team will concentrate on sales, implementation, and partner enablement in the EMEA region, with the aim of spreading Inflection AI's AI system to more enterprises.</p>
</div> <p><b>Analyst QuickTake:<br>
 </b> This acquisition marks a strategic shift to enhance the deployment of trusted enterprise AI agents after changes to its <a href="https:<br>
//sp-edge.com/updates/27806"> business model. </a> Previously, Inflection AI redefined its focus to prioritize user-centric AI solutions that foster human-like interactions, emphasizing trust and reliability. This acquisition aligns with that vision by integrating Boundaryless' expertise in creating secure AI frameworks that facilitate enterprise integration.</p>

Inflection AI acquires Boundaryless for undisclosed sum to accelerate enterprise AI deployments

Exploring the limitless possibilities of AI

Generative AI Applications

<div>
 <p>News Corp, the parent company of The Wall Street Journal and New York Post, has initiated a legal suit against Perplexity AI on charges of copyright infringement. The allegation points towards the unauthorized copying and utilization of copyrighted content by Perplexity AI, purportedly leading to misplaced traffic and revenue that should rightfully belong to the original publishers.</p>
 <p>In the lawsuit, News Corp demands that Perplexity AI cease using its content without proper permits, and also insists on the complete deletion of any databases that contain its copyrighted material.</p>
</div> <p><b>Analyst QuickTake:<br>
 </b> This legal action follows similar moves by other media companies, including The New York Times, against AI firms for unauthorized use of their content. This highlights the ongoing tensions between traditional media companies and AI startups. As AI continues to integrate into information dissemination, the outcome of this case may influence future regulatory frameworks and the operational landscape for media and technology entities.</p>

News Corp sues Perplexity AI for alleged copyright infringement

<div>
 <p>Haiper, a London-based company specializing in video GenAI technology, has introduced Haiper 2.0, an improved version of its video generation model. This version, designed to generate short clips based on user prompts, provides cloud services for consumers to create short videos.</p>
 <p>This updated model uses a DiT (Diffusion Transformer) architecture, which blends diffusion and transformer neural network designs. Haiper 2.0 delivers quicker and more realistic video generation. A forthcoming update promises to hike video resolution to 3,840 by 2,160 pixels.</p>
 <p>In conjunction with the release of Haiper 2.0, the company is set to debut Video Templates. This feature is intended to simplify exploration of new use cases for creators without the need for complex prompts.</p>
 
</div>

All Updates

EdgeRunner AI raises USD 5.5 million in seed funding to build GenAI for Edge devices

Nintendo ventures into MMOs with new Switch Online game playtest

HP releases texture library for 3D printing with CoreTechnologie 4D Additive

Honeywell partners with Google to bring Gemini AI to industrial sector

xAI launches API for third-party developers to access Grok models

Nava Benefits partners with OpenAI to launch AI-powered benefits assistant

IBM releases Granite 3.0 AI models for enterprise use

Honeywell partners with Google to bring Gemini AI to industrial sector

Inflection AI acquires Boundaryless for undisclosed sum to accelerate enterprise AI deployments

News Corp sues Perplexity AI for alleged copyright infringement

Haiper debuts new video generation model

Generative AI Infrastructure

EdgeRunner AI raises USD 5.5 million in seed funding to build GenAI for Edge devices

Contact us