<div>
  <ul>
    <li>
      <p>
        Meta is reportedly cutting back on investments in its Reality Labs division with reports surfacing from The Information that the first version of Meta&rsquo;s AR glasses, codenamed &ldquo;Project Nazare&rdquo; and due to be released in 2024, will not be available commercially.&nbsp;
      </p>
    </li>
    <li>
      <p>
        The reports suggested that the AR glasses, designed to work independently from a mobile phone, would be used as a demonstration product instead. Meta expects to prioritize work on the second version, codenamed Artemis.&nbsp;
      </p>
    </li>
  </ul>
  <p>
    &lt;ul&gt;&lt;li&gt;
    <b>Analyst QuickTake:</b>
     Meta houses its extended reality activities, both hardware and software, under Reality Labs. Reality Labs reported a net loss of USD 10.2 billion in 2021 (on revenue of USD 2.3 billion), which likely motivated the decision to scale back on development. Furthermore, it reportedly plans to launch a total of four new VR headsets by 2024, the first of which is codenamed &ldquo;Project Cambria&rdquo; and is expected to be launched in late 2022 (initially revealed in February 2022) for work-related use cases.&lt;/ul&gt;
  </p>
</div>


<div>
  <p>
    <a href="https://www.theinformation.com/articles/meta-scales-back-ar-glasses-plan-amid-reality-labs-shakeup?utm_source=newsletter&utm_medium=email&utm_campaign=newsletter_axioslogin&stream=top">
      The Information
    </a>
  </p>
  <p>
    <a href="https://techcrunch.com/2022/06/09/meta-scaling-back-plans-ar-glasses-pivoting-portal-away-from-consumers/">
      TechCrunch
    </a>
  </p>
</div>


Extended Reality

Meta turns back on first release of AR glasses

Foundation Models

PROMINENT_CATEGORY

<div>
 <p>H2O.ai, a provider of GenAI models and Predictive AI platforms, has introduced two new multimodal foundation models, H2OVL Mississippi 2 billion and 0.8 billion, designed for OCR and Document AI use cases. They are freely accessible on the Hugging Face platform.</p>
 <p>The H2OVL Mississippi model is designed with efficiency in deployment and handling different image resolutions. It offers multimodal capabilities and can handle OCR and Document AI tasks across multiple resolutions. One of its distinct features includes multi-stage training with fine-tuning layers for optimized performance.</p>
</div>

H2O.ai launches multimodal foundation models for document AI

The toolkit to build machine learning applications at scale

Machine Learning Infrastructure

<div>
 <p>Dottxt, an LLM interaction platform, secured USD 11.9 million through pre-seed and seed funding rounds. The seed round, which accounted for USD 8.7 million, was headed by EQT Ventures, while Elaia led the pre-seed round that raised USD 3.2 million. Other contributing investors included Seedcamp, Common Magic, Kima, and FSJ.</p>
 <p>With this capital, Dottxt plans to grow its team, primarily aiming to hire additional software engineers and a Chief of Staff. The funds will also serve to broaden Dottxt's technology and deploy it to enterprises and developers on a larger scale.</p>
 <p>Dottxt's open-source library, Outlines, has seen significant use, with over 3 million downloads, including 600,000 in September 2024. This technology is being used in production by companies of diverse sizes, and big AI players such as OpenAI and Cohere are listed among its early adopters.</p>
</div>

Dottxt raises USD 11.9 million in seed funding to improve LLMs

Modern tools to train and motivate staff, and bridge emerging skill gaps.

EdTech: Corporate Learning

ANALYST_QUICK_TAKE

<div>
 <p>Eruditus, an Indian executive education platform, secured USD 150 million in a Series F financing round. The round was spearheaded by TPG's The Rise Fund and saw participation from existing investors such as SoftBank Vision Fund 2, Leeds Illuminate, and Accel.</p>
 <p>The procured funds are destined for several purposes. The company aims to enhance its global footprint, invest in advanced AI technology, expand its operational horizons, and make further investments in the Indian and APAC market. Eruditus also has plans to scout for acquisition opportunities. A significant shift in the company’s strategic operations involves relocating its headquarters from Singapore to India in the coming 12–18 months.</p>
 
</div>
 <p><b>Analyst QuickTake</b> :<br>
 This funding marks Eruditus' first major equity round since <a href="https:<br>
//www.prosus.com/news-insights/group-updates/2021/eruditus-announces-650-million-capital-raise-led-by-accel-and-softbank"> August 2021 </a> , when it raised USD 650 million. Excluding the latest round, the company has raised over USD 814 million in equity funding. Despite the challenges facing India's EdTech sector, particularly after Byju's' <a href="https:<br>
//sp-edge.com/updates/28057"> decline </a> , funding continues to flow into companies like Eruditus.</p>

Eruditus raises USD 150 million in Series F funding to expand global presence

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

Additive Manufacturing

<div>
 <p>Metal 3D printing systems provider One Click Metal has formed a partnership with Spanish additive manufacturing solutions company Dativic in order to grow its market presence in Spain. This partnership will introduce One Click Metal's technology to Spanish academic institutions, universities, and research facilities, with Dativic serving as a local partner. Dativic will offer support, technical systems, and training to these institutions for the implementation of metal 3D printing technology.</p>
 <p>One Click Metal, a German startup that was developed by Trumpf, a metal 3D-printing solutions provider, caters to small and medium-sized enterprises by offering metal additive manufacturing technology. One of their key products is the BOLDseries, designed with the MPRINT+ system, an LPBF (laser powder bed fusion) 3D printer. This technology uses a 200 W fiber laser to melt metal powder and can produce highly accurate prints with a precision of 0.07 mm. The MPRINT+ system presents a considerable build volume of 150 x 150 x 150 mm.</p>
</div>

One Click Metal partners with Dativic to expand metal 3D printing in Spain

INCUMBENTS_PRESENT

<div>
 
 <p>NVIDIA, in conjunction with MIT and Tsinghua University, have launched "Sana," an AI model specialised in creating text-to-image content. Sana has the capability to generate high-resolution images up to the size of 4,096 x 4,096 in mere seconds. Though public access to the model is granted, the prices for using Sana have not been revealed.</p>
 <p>Sana operates using a combination of a deep compression autoencoder, a linear DiT (Diffusion Transformer), and a decoder-only text encoder. The model uses Flow-DPM-Solver to minimise the number of sampling steps required, effective caption tokenization, and selection to speed up the process of image generation. NVIDIA asserts that with the use of Sana-0.6 billion, it's possible to produce 1,024 x 1,024 resolution images on a 16GB laptop GPU in under a second.&nbsp;</p>
 
</div> <p><b>Analyst QuickTake:<br>
 </b> Sana’s image synthesis technology's key differentiators lie in its ability to understand and replicate complex patterns and textures, which is critical for producing lifelike images. By incorporating advanced neural networks that can analyze and interpret various art styles and elements, Sana provides users with the tools to generate visuals tailored to their needs. This capability enhances user creativity and may reduce the time and effort typically required to produce images.</p>

NVIDIA releases Sana, an AI model for image generation

Using data and automation to achieve 24/7 production.

Smart Factory

<div>
 
 <p>Avathon (previously known as SparkCognition) is a company that offers AI analytics solutions. It has recently launched a new system-level industrial AI platform.</p>
 <p>This AI platform aims to tackle a range of industrial problems, such as maximizing the uptime of high-cost assets, facilitating efficient manufacturing increase, and maintaining industrial worker safety. The platform uses a systems-level strategy to offer solutions for critical and intricate infrastructure.&nbsp;</p>
 <p>In addition, Avathon has revealed plans to move its headquarters to Silicon Valley. This move is part of the company's strategy to position itself within the area's innovation ecosystem.</p>
 
</div>

Avathon launches system-level industrial AI platform

Exploring the limitless possibilities of AI

Generative AI Applications

<div>
 <p>Perplexity, an AI chat and search platform, now offers Internal Knowledge Search for Pro and Enterprise Pro users. This service permits clients to search across the web and their internal databases within the same platform. Its availability is exclusive to the company's premium tier subscribers.</p>
 <p>This new feature lets users upload and look through their files such as Excel spreadsheets, Word documents, and PDFs. However, it's only limited to specific files that users consider crucial and upload to the platform, capping at 500 uploads for Enterprise Pro users. In addition, Perplexity unveiled Spaces, a feature for teams to share, organize research, customize AI assistants, and regulate access to information.</p>
 
</div> <p><b>Analyst QuickTake:<br>
 </b> With these new features, Perplexity is positioning itself as a major player in the AI search engine market. Its ability to combine internal and external data queries differentiates it from traditional search engines and even some AI-driven competitors. As businesses increasingly seek integrated solutions that enhance productivity and collaboration, Perplexity's offerings could attract a growing user base, particularly among enterprise clients.</p>

Perplexity launches Internal Knowledge Search for enterprise users

<div>
 <p>UK-based text-to-speech company Neuphonic has secured GBP 3 million in pre-seed funding in a round led by Moonfire VC and included contributions from Tiny VC, Salica Oryx Fund, and Cur8 Capital.</p>
 <p>The firm intends to use this finance to grow its language capabilities and voice services, boost model performance via increased research, and to create on-device solutions.</p>
 <p>Neuphonic's unique algorithm, currently patent-pending, allows for real-time speech production with an ultra-low latency of 25 milliseconds. Its API is accessible to clientele through an exclusive closed beta program.</p>
</div>

Neuphonic raises GBP 3 million in pre-seed funding to expand capabilities

<div>
 <p>Bain &amp; Company, a global consultancy firm, has broadened its collaboration with OpenAI to hasten the deployment of AI in businesses. The enhanced partnership involves the foundation of an OpenAI Center of Excellence at Bain, the joint design of industry-tailored solutions primarily for retail and healthcare life sciences, and the application of OpenAI's latest technologies such as multi-modal, real-time, and reasoning applications. The objective of this alliance is to augment Bain's capacity in providing potent AI solutions and assisting clients in navigating their AI transformation journey.</p>
</div> <p><b>Analyst QuickTake:<br>
 </b> The partnership has already yielded results for clients such as The Coca-Cola Company and Amgen, with Bain embedding multiple AI solutions into clients' operations across various use cases. The expanded collaboration will continue to focus on delivering concrete business results and transformational impact for clients globally.</p>

Bain & Company expands partnership with OpenAI to accelerate AI solutions for clients

<div>
 <p>OpenAI has shared a preview of a unique Windows app for ChatGPT, that is currently exclusive to ChatGPT Plus, Team, Enterprise, and Edu users. A more comprehensive version is promised later in the year.</p>
 <p>The unique feature of the ChatGPT app for Windows supports users to communicate about files and photos, experience the latest model enhancements of OpenAIl, inclusive of the o1-preview model, and is compatible with most Windows 10 machines. The app's capability to condense to a small "companion" window while using other apps, permits users to upload files and photos, summarize documents, and generate images using OpenAI's DALL-E 3 image generator.</p>
</div> <p><b>Analyst QuickTake:<br>
 </b> By making ChatGPT available on Windows, OpenAI is tapping into a vast market of users who rely on this operating system for daily tasks. The app is compatible with most Windows 10 machines, which broadens its reach significantly. While the app currently has limitations, its integration into the Windows ecosystem positions it well for future growth and development.&nbsp;</p>

All Updates

Meta turns back on first release of AR glasses

H2O.ai launches multimodal foundation models for document AI

Dottxt raises USD 11.9 million in seed funding to improve LLMs

Eruditus raises USD 150 million in Series F funding to expand global presence

One Click Metal partners with Dativic to expand metal 3D printing in Spain

NVIDIA releases Sana, an AI model for image generation

Avathon launches system-level industrial AI platform

Perplexity launches Internal Knowledge Search for enterprise users

Neuphonic raises GBP 3 million in pre-seed funding to expand capabilities

Bain & Company expands partnership with OpenAI to accelerate AI solutions for clients

OpenAI launches ChatGPT app for Windows in preview

Extended Reality

Meta turns back on first release of AR glasses

Contact us