New photonic architecture will power AI inference for models exceeding 20 trillion parameters at up to 10,000 tokens per second per user

SAN FRANCISCO, Oct. 1, 2026 — Volantis, a semiconductor company building a new category of AI inference system, today announced an $88 million Series A co-led by Lachy Groom and Abstract Ventures, with participation from John Doerr, VXI Capital, Triatomic and Susa Ventures. The round also includes angel investors Dwarkesh Patel, Naveen Rao and Sholto Douglas.

Volantis is building a new AI inference architecture that eliminates the tradeoff between memory capacity and bandwidth. Its first system, A-1, is being designed to run models exceeding 20 trillion parameters at up to 10,000 tokens per second per user, while reducing inference cost per token.

The company’s founding team includes semiconductor and photonics veterans from NVIDIA, AMD, Broadcom and Ayar Labs. Their previous work includes the first CoWoS product, the first high-volume tunable VCSELs and early silicon photonics co-packaged optics systems.

“As AI agents take on more work, how fast they complete that work will increasingly determine how fast companies can operate,” said Tapa Ghosh, CEO and co-founder of Volantis. “Today’s hardware forces a tradeoff between running the largest, most sophisticated models and running them fast. We started Volantis to eliminate that tradeoff.”

Breaking Through the AI Memory Wall

Running large AI models at high speeds requires enormous memory capacity to hold the model and enormous bandwidth to continuously feed data into the compute engine. Existing architectures force a compromise between the two.

On-chip SRAM provides high bandwidth but limited capacity, constraining model size and context windows while increasing cost and energy consumption. GPUs and other HBM-based systems provide greater capacity, but bandwidth limits how quickly they can serve increasingly large models. Even newer approaches such as 3D DRAM remain on the same tradeoff curve.

Volantis is designing A-1 to increase memory capacity and bandwidth simultaneously by nearly two orders of magnitude, allowing increasingly large models to run at much higher inference speeds.

That performance could significantly accelerate AI agent workflows. A coding agent that completes a task in two minutes instead of 30 gives developers more opportunities to test ideas, ship software and iterate. As agents take on longer and more complex work, inference speed increasingly determines how quickly they can complete it.

A New Approach to Photonic Interconnects

Volantis is building a new category of photonic interconnect designed specifically to connect compute chips to memory.

Its optical fabric connects large numbers of memory chips into a unified pool, aggregating their bandwidth as memory is added. This allows A-1 to increase memory capacity and bandwidth together while using lower-cost off-chip memory.

While photonics is already used to move data inside data centers, existing technologies have largely focused on chip-to-chip connections. Chip-to-memory connections require more than 100 times as much data to travel over much shorter distances, creating different requirements for energy and cost.

Volantis developed a photonic platform specifically for this environment. Its architecture uses custom micro-VCSELs rather than external lasers, drawing on the existing gallium arsenide VCSEL supply chain and avoiding indium phosphide supply constraints. The micro-VCSELs are small, temperature-stable and low power, enabling end-to-end links consuming less than one picojoule per bit. Volantis plans to unveil additional innovations behind the architecture as A-1 moves toward commercialization.

Investment and Commercialization

Volantis plans to deliver its first integrated inference engines to customers in 2027.

The financing will support the development and commercialization of A-1 and its photonic memory architecture, including expanding the company’s engineering team and moving the system toward customer deployments.

About Volantis

Volantis is a semiconductor company building a new architecture for AI inference. The company combines photonics and semiconductor design to increase memory capacity and bandwidth, enabling larger AI models to run at significantly higher speeds.

For more information, visit https://volantissemi.ai/

SOURCE Volantis

Built by experts from British Intelligence, cybersecurity and fin-tech, including leaders who have previously scaled multiple unicorn companies, Overmind makes it easier for organizations to build, train and deploy specialized AI models using their own data and code

SAN FRANCISCO, Oct. 1, 2026 — Overmind, a platform for building, fine-tuning and deploying specialized small language AI models, today announced the open source release of its training and deployment platform, giving developers and companies a new way to build AI systems tailored to their own data, code and business needs.

The company’s origins trace back to founder and CEO Tyler Edwards’ experience in British intelligence, including roles at MI5 and MI6, where he oversaw teams exploring how AI could automate intelligence analysis and other high-stakes processes. Working on these problems, Edwards saw how difficult and expensive it was to build AI models for highly specialized environments, and how many resources those efforts required. He also saw the potential to bring that capability to organizations far beyond intelligence and national security.

“Some of the hardest problems in AI aren’t about building a bigger model, they’re about building the right model for the right job,” said Edwards, co-founder and CEO of Overmind. “I saw that firsthand in intelligence, where generic solutions often aren’t enough. The same is true in business. Companies have unique data, code and expertise, yet most are still relying on general-purpose models built for everyone. We built Overmind to make specialized AI practical, affordable and accessible to organizations across industries.”

Edwards founded Overmind alongside Sam Brunt, a serial startup operator who has helped scale Funding Circle, Pipe and Vertice into high-growth companies. Together, they built Overmind to give teams the power to train and ship their own AI models without having to hire a team of AI specialists. The platform brings together the key stages of specialized model development, from agent observability and data preparation to model training and inference. Those traces can then be evaluated against benchmarks to measure relevance and performance, with the results informing fine-tuning and redeployment. This end-to-end workflow allows companies to turn their proprietary data and software into models built for their own products and workflows. 

“We’re entering a phase where companies are going to look beyond one-size-fits-all frontier models and build AI systems specifically for the problems they need to solve,” said Brunt. “Specialized models can deliver stronger performance on narrowly defined tasks at a fraction of the inference cost, while proprietary data gives companies the opportunity to build models their competitors simply don’t have. The challenge is that building and managing these systems is still a complex and highly specialized process. Overmind abstracts away that complexity, giving organizations a simpler, more accessible way to retrain, modify, roll back, download and deploy their own models wherever they need. Just as importantly, companies should be able to own the models they build, without being locked into a single provider.”

In early deployments, Overmind’s approach is producing significant results. In one fintech deployment, a specialized model trained by Overmind reduced invoice-processing costs by 96%. In a separate legal benchmark, Overmind improved accuracy by 86% compared with out-of-the-box frontier models, while also reducing false positives and operating costs. The company has recorded tens of thousands of SDK downloads and is already working with companies across fintech, legal, healthcare and cybersecurity, adapting specialized AI to a range of industry-specific use cases.

Security is foundational to Overmind, particularly as organizations use the platform with proprietary data and code. The team’s experience in British intelligence and cybersecurity has shaped the platform’s infrastructure and development processes around secure deployment and scaling from the outset. Overmind is SOC 2 and ISO compliant.

The open source release gives developers and organizations greater control over how they build, train and deploy specialized AI, while allowing them to adapt the platform to their own needs. It also advances Overmind’s broader mission to make powerful AI accessible to everyone, regardless of capital or technical expertise. Overmind believes that openness and accessibility should extend beyond the models themselves to the platforms used to build with them. By opening the platform, Overmind is expanding access to the tools and infrastructure needed to experiment, customize, and deploy AI on users’ own terms. The company is continuing to work with early enterprise adopters as it expands into new industries and use cases.

For more information about Overmind, please visit: https://www.overmindlab.ai/

About Overmind

Overmind is building the infrastructure for specialized AI. Founded in 2025, the platform enables organizations to use their own data and code to train, evaluate and deploy small language AI models built specifically for their products and workflows. Overmind combines expertise from intelligence, cybersecurity and fintech with deep experience building and scaling technology companies, making specialized model development faster, more affordable and accessible to organizations across industries.

Media Contact:

Rachel Rogers

3107704917

[email protected] 

SOURCE Overmind

Expands Health & Wellness-Focused Private Equity Firm’s Brand Building Capabilities

DENVER, Oct. 1, 2026 — Manna Tree, a global private equity firm investing in companies that empower consumers to live better, longer, today announced a majority investment in Zambezi, the award-winning independent creative agency known for building billion-dollar brands and breakthrough consumer campaigns. Terms were not disclosed.

Headquartered in Los Angeles and celebrating its 20th anniversary in 2026, Zambezi helps brands drive outsized growth through integrated creative, strategic storytelling, production marketing and sports partners. The agency has a proven track record of helping companies become category leaders, most recently propelling Liquid I.V. to billion-dollar brand status.

Brand-Building as a Core Investment Thesis

Since its founding in 2018, Manna Tree has focused on investing in consumer health and wellness companies with strong momentum and market share, applying a strategy centered on identifying and scaling category-defining brands. The Zambezi investment marks the firm’s latest step in expanding the services and resources it makes available to its portfolio companies to support meaningful expansion, adding marketing to its existing talent, supply chain, and finance capabilities. With Zambezi, Manna Tree adds creative, media and production marketing capabilities that enhance its ability to support companies across their growth cycle while extending its reach across the broader health and wellness ecosystem.

“We are not just investing in a growing creative agency – with Zambezi we are investing in the future of our firm and our ability to support the next generation of category-defining health and wellness brands,” said Ross Iverson, Co-Founder and Managing Partner of Manna Tree. “We know that great companies require more than capital to scale – they need to be the best brand in the marketplace. The Zambezi team brings world-class expertise in storytelling, consumer engagement and brand building that will strengthen how we help companies grow.”

Manna Tree portfolio companies, which include leading better-for-you brands such as Good Culture, Gotham Greens, Generous Brands, New Primal, Plant People, Verde Farms, and others will have access to Zambezi’s full suite of services including brand strategy, creative development, media, production through FIN Studios, and sports and culture partnerships.

Unlocking Growth at the Intersection of Wellness, Sports, and Culture

The investment positions Manna Tree at the growing convergence of health and wellness with sports, culture, and community. Consumer wellness brands are increasingly building direct relationships with consumers through live experiences, athlete partnerships, and culturally embedded campaigns rather than traditional advertising alone. “Zambezi’s deep expertise in this space, including work with the NFL, TaylorMade, and ESPN, provides a clear pathway into these high-impact consumer touchpoints for not only our portfolio companies but high-potential health and wellness brands across the broader market,” said Ellie Rubenstein, Co-Founder and Managing Partner of Manna Tree.

Zambezi will continue to operate independently, serving its existing client base and pursuing new business, while also working alongside Manna Tree’s investment and portfolio operations teams to identify and act on shared opportunities. The agency will remain headquartered in Los Angeles and be led by its existing management team, including CEO Jean Freeman and President Laura Stayt.

“Today’s consumers expect brands to participate in culture, not simply market to them,” said Jean Freeman, Chief Executive Officer of Zambezi. “Manna Tree understands where consumer behavior is heading and shares our belief that strong brands are built through authentic connections, compelling storytelling and dynamic engagement. We’re excited to partner together as we help companies navigate their next era of growth.”

DLA Piper served as legal counsel to Manna Tree. Jones Spross served as legal counsel to Zambezi, with Pharus acting as its financial advisor.

About Manna Tree
Founded by Gabrielle (Ellie) Rubenstein and Ross Iverson in 2018, Manna Tree is a global private equity firm committed to investing in companies that empower consumers to live better, longer. The firm strategically invests in the consumer sector, particularly within the health and wellness ecosystem, focusing on growth-stage and buyout opportunities. Manna Tree has made investments in 15 companies to date, including companies like: Health-Ade, Gotham Greens, Vital Farms, Good Culture, Plant People, Verde Farms, Urban Remedy, and the New Primal.

About Zambezi
2026 AdAge Small Agency of the Year winner Zambezi is a proudly independent, women-owned agency bringing enduring creative and media ideas to sports, fitness, health and wellness and lifestyle companies, among others. Zambezi builds brands that are as high-performance as their products and services, spurring growth to $1B and beyond. Headquartered in Los Angeles with offices in Minneapolis, the agency serves leaders in their sectors, including TaylorMade, LPL Financial, Club Pilates, Traeger Grills, UKG, NFL, ESPN, Under Armour, Atlantis Bahamas, and more. The rapidly growing firm has been named to the Inc. 5000 list of fastest-growing private companies in the U.S. eight times, most recently in 2026. Zambezi’s  momentum over the past year has included being named ADWEEK Breakthrough Media Agency of the Year and receiving multiple awards for its creative work at Cannes Lions, The One Show, Emmys, Webbys, Clios, and more.

Media Contact:

Kate Schneiderman
[email protected] 

SOURCE Manna Tree Partners

The Q-Lab will build the research, tools, and field infrastructure needed to make better questioning
a rigorous discipline and practical skill

NEW YORK, October 1, 2026 — Siegel Family Endowment today announced a multiyear grant to support the launch of The Questions Lab (Q-Lab), an initiative by The Governance Lab (The GovLab) dedicated solely to the science and practice of inquiry. Q-Lab will serve as a formal hub for advancing question literacy and building the emerging field of question science, helping organizations across sectors strengthen their ability to formulate rigorous, inclusive, and actionable questions.

The grant builds on Siegel’s long-standing commitment to inquiry-driven philanthropy and follows the release of the foundation’s white paper, Better Questions, Better Insights: How Philanthropy Can Harness the Science of Questions for Deeper Impact. Together, the white paper and Q-Lab investment signal a broader effort to move the science of questions from an emerging idea to a field with dedicated research, practice, and institutional infrastructure.

“We have spent years using inquiry as a grantmaking practice, but this work cannot live inside one foundation alone,” said Joshua Elder, Senior Vice President and Head of Grantmaking at Siegel. “The Questions Lab gives the emerging science of inquiry a durable home. It creates the conditions for researchers, practitioners, funders, and communities to better understand what makes a question powerful, whose questions shape the agenda, and how disciplined inquiry can lead to deeper impact.”

Siegel’s support will help Q-Lab build the research, practical methods, and global community needed to make question formulation a more systematic and accessible practice across philanthropy, public interest technology, education, civic institutions, and other fields navigating complex social challenges.

The John Templeton Foundation is proud to join Siegel Family Endowment in supporting Q-Lab as it builds the intellectual infrastructure for a deeper and more humble culture of inquiry. “Many of humanity’s most important advances begin not with an answer but with a question,” said Timothy Dalrymple, President of the John Templeton Foundation. “At a moment when technologies can generate answers with remarkable speed, we have an opportunity to pay renewed attention to the questions that guide discovery, shape our understanding of the world, and open new possibilities for exploration.”

As artificial intelligence and data systems make it easier to generate rapid responses, the ability to ask sharp, generative questions is becoming increasingly important. Q-Lab will focus on building the science of questions as both a field of study and a practical capability, helping leaders move beyond information gathering toward deeper insight, better decision-making, and more democratic agenda-setting.

“The future of problem-solving depends not only on better answers, but on better questions,” said Stefaan G. Verhulst, Co-Founder of The GovLab and Q-Lab’s Founder. “Q-Lab is designed to help individuals and institutions ask questions that are rigorous, inclusive, and actionable. We’re excited to partner with Siegel and the John Templeton Foundation to build the infrastructure needed to make question science a discipline that serves the public interest.”

To learn more, please visit questionslab.org.

About Siegel Family Endowment
Siegel Family Endowment employs an inquiry-driven approach to grantmaking, informed by the scientific method and grounded in the belief that philanthropy is uniquely positioned to address complex societal challenges. Rapid technological change has reshaped how we live, work, and learn, transforming the global economy and redefining access to opportunity – from schools and workplaces to our built environment. To meet these shifts, we support technology that serves the public interest, including the tools, skills, and systems people need to engage with and shape a rapidly evolving world. Siegel Family Endowment was founded in 2011 by David Siegel, co-founder and co-chairman of financial sciences company Two Sigma.

SOURCE Siegel Family Endowment

dearCC combines human mentorship, community and peer networks with real-time labor-market intelligence and AI-powered career tools to help young people navigate a rapidly changing world of work

OAKLAND, Calif., Oct. 1, 2026 — New Work Foundation, the nonprofit co-founded by former Meta and Salesforce AI executive Clara Shih and technology industry veteran Nick Mehta, today launched dearCC, a free platform designed to inspire confidence and restore personal agency for Gen Z jobseekers in an economy reshaped by artificial intelligence.

Amidst one of the toughest job markets for new grads in recent years, dearCC replaces job-search isolation with a clear roadmap forward. The platform combines human mentorship and peer cohorts with real-time labor-market intelligence and practical AI-skilling tools, giving young workers actionable insight on where opportunity exists, the capabilities to pursue it, and a community to back them.

“The bottom rung of the career ladder is disappearing,” said Clara Shih, co-founder of New Work Foundation. “I spent years building business AI systems and saw firsthand how they automate tasks that used to be entry-level jobs. AI is both driving this dislocation and an important part of the solution, but AI alone won’t solve the challenge of helping young people break through. dearCC gives young professionals the mindset, skillset, and community to succeed in the new work economy.”

The launch is supported by a multi-year grant and anchor partnership with the Schultz Family Foundation, which supports innovative efforts to increase economic mobility for America’s next generation. Through this collaboration, New Work Foundation will integrate labor-market insights from the Schultz Family Foundation and Burning Glass Institute’s Where You Work Matters data into dearCC, equipping young jobseekers with actionable guidance on employer quality and economic mobility.

“Young people are entering one of the most consequential transitions in the labor market in generations. They need more than another job board: They need trusted guidance, better information and tools that help them understand where real opportunity exists,” said Howard Schultz, co-founder of the Schultz Family Foundation. “New Work Foundation shares our belief that AI should expand opportunity rather than make an already difficult transition into the workforce even harder. We are excited to invest in this work and enhance the new platform with the labor-market intelligence and career-navigation work we have developed through Where You Work Matters.”

“In this market, graduates need more than a degree. They need the skills to contribute from day one, including the ability to use AI as a tool to make themselves more capable and more valuable. Our analysis of tens of millions of career histories shows how much those first few years matter. The jobs graduates land, the skills they build, and the companies in which they build them can shape their trajectory for years to come,” said Matt Sigelman, President of the Burning Glass Institute.

dearCC: A Career Platform Built for Gen Z Jobseekers

Designed to replace the isolation and guesswork of the modern job hunt, dearCC equips entry-level jobseekers with the AI-powered tools, peer community, and mentor feedback needed to develop their mindset, skillset, and network for the AI economy:

• dearCC Field Report: An AI-powered jobs explorer that shows which roles are growing, flatlining or changing. Jobseekers can search by their college degree and compare AI exposure, wage trends, high-growth adjacent occupations, and geographic opportunities, helping people make more informed decisions about where to focus their search.
• dearCC Game Plan: A step-by-step assessment of a jobseeker’s LinkedIn profile, resume, and current job-search strategy, paired with hands-on opportunities to build AI capabilities relevant to their target role. Users receive a personalized gap analysis, outreach recommendations, interview preparation and portfolio-project ideas designed to help them stand out.
• dearCC Community : An online community designed to counter the isolation of today’s increasingly automated hiring process with role-specific learning modules, guidance, and peer support. Jobseekers are invited to monthly master classes led by expert mentors across various fields, including Will Sentance (Founder, Codesmith), April Underwood (Former Chief Product Officer, Slack), Katie Stanton (Former Chief Marketing Officer, Color Health), Michael Matlock (Chief Accounting Officer, Benchling), and more. Users can also sign up to get matched to a Crew: six jobseekers who meet weekly with a field mentor to provide support, structure, and accountability. 
• dearCC Weekly Brief: A weekly newsletter personalized to each jobseeker’s role, level, and experience to help them understand how AI is being used in and reshaping the occupations and industries they hope to enter.

“dearCC has given me the community, inspiration, and AI skills to finally get a job after 12 months of searching,” said Jesus Garnica, a first-generation college graduate of San Mateo Community College and San Francisco State.

“When you’re staring down 500 unanswered job applications, you don’t need theoretical advice and platitudes,” said Samantha Wen, UC Berkeley class of 2023 and Gen Z co-founder of New Work Foundation. “You need a plan for a specific job and a group of people who will give you straight talk and empathy.”

Ryon Baldwin-Williams, a 2025 Michigan State graduate and Gen Z advisory board member, added: “dearCC unstuck my search process and has given me a lifelong community.”

All products are free and available beginning today at http://dearcc.org.

Strategic Partnership, Founding Supporters, and Leadership Team

In addition to support from the Schultz Family Foundation, New Work Foundation has secured generous founding support from Morfit Family Foundation, Brian Halligan, Sheryl Sandberg and Tom Bernthal, Chirantan “CJ” Desai and Meera Desai, Joe Lonsdale, Nikki Pechet, April Underwood, and Aaron Levie, as well as financial and technology grants from Salesforce (including Slack), Anthropic, Google, OpenAI and Vercel.

New Work Foundation also welcomes five seasoned industry leaders to its all-volunteer leadership team with experience from companies including AT&T, BEA Systems, Informix, Salesforce, Smartsheet, Vera Therapeutics, and VMware:

• Eric Loeb joins as Chief External Affairs Officer.
• Jody Kohner joins as Chief Learning Officer.
• Pete Godbole joins as Chief Financial Officer.
• Eric Stahl joins as Chief Marketing Officer.
• Julien Capers joins as General Counsel.

About New Work Foundation

New Work Foundation is an independent nonprofit organization dedicated to equipping young American workers with the AI tools, labor-market intelligence, and community to thrive in the new work economy. Founded by former Meta and Salesforce AI executive Clara Shih and technology industry veteran Nick Mehta, the foundation operates with an explicit commitment to provide free tools for the broad-based public interest.

Learn more at http://newworkfoundation.com.

Media Contact: [email protected]

SOURCE New Work Foundation

Proceeds from the facility will support the deployment of compute infrastructure dedicated to customer contracts

NEW YORK, Oct. 1, 2026 — SharonAI Holdings Inc. (NASDAQ: SHAZ) (“Sharon AI” or the “Company”), a leading Australian Neocloud delivering trusted AI infrastructure, today announced it has entered into its inaugural US$356m committed senior secured, GPU-backed SPV debt facility priced at a fixed rate of 9.95%, excluding fees.

The facility will be secured against the GPUs and associated cash flows, with the contract-backed funding structure reflecting the security and delivery discipline underpinning Sharon AI’s platform and marks a further step towards the Company’s ambition of delivering gigawatt-scale AI compute capacity across Australia, New Zealand and the broader Asia-Pacific.

The facility includes marquee Australian, Asian and global investors including Goldman Sachs and select large private credit funds. This facility is the first in an expected series of GPU financings supporting the scheduled build out of over 68,000 NVIDIA GPUs deployed by mid-2027.

With the closing of this transaction, Sharon AI will have secured over US$2.6bn of institutional debt and equity capital over the past 10 months, extending the Company’s capital markets program and underscoring Sharon AI’s ability to access diversified sources of capital as it scales its AI infrastructure platform for a growing customer base of hyperscale, AI natives, government, enterprise, and research organizations.

“As demand for sovereign and secure, trusted AI infrastructure continues to outpace available supply globally, and particularly across Australia, New Zealand and the broader Asia-Pacific, access to scalable debt capital is an important enabler of our growth,” said James Manning, Co-founder and Chief Executive Officer of Sharon AI.

“This facility demonstrates how we expect to access debt markets to fund our GPU deployments, leveraging our book of quality customer offtake now standing at a TCV of over US$8.8bn. This is designed to enhance return on equity and ultimately drive increased long-term shareholder value. With a strong balance sheet, growing contracted capacity pipeline and a disciplined approach to capital allocation, we believe we are well positioned to continue scaling our AI platform across the Asia-Pacific region.“

Jarden Australia acted as sole financial advisor and arranger.

ENDS

About Sharon AI

Sharon AI (NASDAQ: SHAZ) is a leading Australian Neocloud delivering trusted sovereign AI infrastructure. Through its AI Factory platform and world-class ecosystem of technology and co-location partners, Sharon AI expands access to the scalable capabilities organizations need to build, train and run AI, from model training through to inference and agentic AI. Serving customers globally, Sharon AI helps organizations move faster from AI potential to measurable value. For more information, visit www.sharonai.com.

Media
[email protected]

Investors
[email protected]

Disclosure Information

Sharon AI primarily uses its Investor Relations page (https://sharonai.com/investors/) to disclose material non-public information and to comply with its disclosure obligations under Regulation FD. The Company also notes that, at times, it discloses material non-public information through other communication mediums including, but not limited to, its X account (sharon__ai) and/or LinkedIn account (sharon-AI), press releases, and regulatory filings with the SEC, or through conference calls, webcasts, and investor days, etc. that the company may hold.

Forward-Looking Statements

This press release may contain, and our officers and representatives may from time to time make, “forward-looking statements” within the meaning of the safe harbor provisions of the U.S. Private Securities Litigation Reform Act of 1995, which are not historical facts, and which are not assurances of future performance. Forward-looking statements are based only on our current beliefs, expectations and assumptions regarding the future of our business, future plans and strategies, projections, anticipated events and trends, the economy and other future conditions. In some cases, you can identify these statements by forward-looking words such as “believe,” “may,” “will,” “estimate,” “continue,” “anticipate,” “intend,” “could,” “should,” “would,” “project,” “strategy,” “plan,” “expect,” “goal,” “seek,” “future,” “likely” or the negative or plural of these words or similar expressions or references to future periods. Examples of such forward-looking statements include but are not limited to express or implied statements regarding Sharon AI’s management team’s expectations, hopes, beliefs, intentions or strategies regarding the future including, without limitation, statements regarding:

• Service and product offerings;
• The deployment of assets and expansion of network procurement;
• Sharon AI’s ability to engage with additional potential customers;
• Expansion of Sharon AI’s data center footprint and capacity;
• The strengthening of Sharon AI’s partner network;
• Additional or future GPU financings;
• Complete fulfillment of all customer contracts;
• The impact and effect of debt structures designed to enhance return on equity and to drive stockholder value; and
• The Company’s position to continue scaling its AI platform across the Asia-Pacific region.

In addition, any statements that refer to projections, forecasts or other characterizations of future events or circumstances, including any underlying assumptions, are forward-looking statements. Because forward-looking statements relate to the future, they are subject to inherent uncertainties, risks and changes in circumstances that are difficult to predict and many of which are outside of our control. You are cautioned that such statements are not guarantees of future performance and that actual results or developments may differ materially from those set forth in these forward-looking statements. Therefore, you should not rely on any of these forward-looking statements. Important factors that could cause actual results to differ materially from these forward-looking statements include, among others, all of the risks described in the “Risk Factors” section of the Company’s most recent Annual Report on Form 10-K filed with the SEC and other reports subsequently filed with the SEC. Additional assumptions, risks and uncertainties are described in detail in our registration statements, reports and other filings with the SEC, which are available at www.sec.gov.

The forward-looking statements and other information contained in this press release are made as of the date hereof and Sharon AI does not undertake any obligation to update publicly or revise any forward-looking statements or information, whether as a result of new information, future events or otherwise, unless so required by applicable securities laws.

SOURCE SharonAI Holdings Inc.

The funding will scale production of OSCP’s ITAR-free photonic IMUs in Montréal as demand grows for navigation that works when GNSS is jammed.

MONTRÉAL, Oct. 1, 2026 — OSCPS Motion Sensing Inc., operating as OSCP, today announced that it has closed a Series A financing led by New Science Ventures, with participation from Emerging Ventures, and 2050 Capital. Somu Subramaniam, founder and managing partner of New Science Ventures, has joined OSCP’s board of directors.

OSCP’s photonic gyroscopes and inertial measurement units (IMUs) are built for platforms that must keep navigating when GNSS is jammed, spoofed or unavailable, a problem that now reaches well beyond declared conflict zones.

The company will use the funding to scale production in Montréal, grow its engineering and sales teams, and bring its next generation of photonic inertial sensors to market.

“Losing the satellite signal used to be a battlefield problem. Now it shows up over airports and shipping lanes,” said Kazem Zandi, founder and chief executive of OSCP. “Customers want navigation they can trust when GNSS drops out, and they want it ITAR-free. This round lets us build more of it here, and faster.”

“We are very excited to join the team at OSCP to build the world’s best IMUs in terms of navigation performance, size, and cost. These IMUs have the potential to completely change the autonomous navigation paradigm in a wide range of applications in the defence and transportation sectors,” said Somu Subramaniam, founder and managing partner of New Science Ventures.

Since the round closed in August, OSCP has introduced NavigationGate, an inertial navigation system for platforms that lose satellite signal, added ArduPilot support and an open-source ROS 2 driver for its MK2 IMU family, and qualified as a supplier on Canada’s Defence Drone Initiative Marketplace. For details, visit oscp.com.

ABOUT OSCP

OSCPS Motion Sensing Inc., operating as OSCP, designs and manufactures photonic gyroscopes and inertial measurement units. Founded in 2015 and based in Montréal, Québec, the company builds inertial sensors that deliver tactical-grade photonic accuracy at the size, weight, power and cost of MEMS, for navigation where GNSS is jammed, degraded or unavailable. Its customers work in space, defence, marine and subsea, rail, robotics and ground autonomy. OSCP products are made in Canada and are ITAR-free.

ABOUT NEW SCIENCE VENTURES

New Science Ventures is a venture capital firm that invests in companies taking novel scientific approaches in information technology and life sciences. Founded in 2004, with offices in Greenwich, Connecticut, and London, the firm has more than US$1 billion in assets under management, based on original committed capital. For more information, visit newscienceventures.com.

Media contact: [email protected]

Five accomplished executives spanning technology, finance, operations, consulting and transformation will advise the independent IT sourcing firm as it continues to grow and evolve

BOSTON, Sept. 30, 2026 — UpperEdge, an independent IT sourcing and vendor negotiation advisory firm, today announced the formation of its inaugural Advisory Board. The five members are a former biopharma CIO, two former senior leaders from Accenture and EY, a former consumer products President and CFO, and a procurement and supply chain executive. Together they bring decades of experience leading, funding and delivering large-scale technology transformation.

The board’s formation comes as large enterprises navigate a rapidly changing technology landscape, including major platform transformations, evolving cloud and AI strategies, and significant changes in how enterprise software and IT services are priced, packaged and contracted. Collectively, the five board members have experienced these decisions from multiple perspectives, including the buyer, the provider, and the business executives responsible for funding and delivering transformation.

The Advisory Board will work with UpperEdge leadership throughout the year on growth strategy, client relationships, new market opportunities and the continued evolution of the firm’s offerings. It will support the firm as it expands its advisory services and invests in new, technology-enabled ways to deliver its negotiation expertise to clients.

“UpperEdge has built a strong business, but I believe we have significant opportunity ahead of us,” said David Blake, CEO of UpperEdge. “I wanted to surround our leadership team with accomplished people who bring different experiences and perspectives, who understand the executives and organizations we serve, and who are willing to challenge our thinking. Their perspective will help us sharpen our strategy, identify new opportunities, and continue evolving UpperEdge as we enter our next phase of growth, while ultimately helping us deliver even greater value to our clients.”

Advisory Board Members

Vince Ambrosino is a consumer products executive with operational, financial and strategic leadership experience at McCain Foods, Suntory and PepsiCo. At McCain Foods, he served as President of the North America Potato Division and Chief Financial Officer for the North America Region, where he helped restore growth and profitability. As Chief Operating Officer of Suntory Holdings in Japan, he led global initiatives to strengthen the company’s operating model and expand across Asia. Earlier in his career, he held senior finance and operating roles at PepsiCo in the U.S. and Canada, contributing to turnaround efforts and sustained margin expansion.

David Davidson is an enterprise transformation executive with more than 35 years of experience in corporate strategy, large-scale ERP transformation and cost optimization. As a Senior Managing Director at Accenture, he served on the North America Leadership Team and led the firm’s CFO & Enterprise Value Strategy & Consulting practice. Since retiring from Accenture in 2022, he has advised organizations in the music, wealth management and consumer products industries through David A. Davidson Consulting.

Lori Foster is an accomplished business leader and entrepreneur with more than 30 years of experience leading supply chain organizations and business transformations across multiple industries. Throughout her career, she led complex, large-scale global SAP implementations at multinational organizations, driving improvements in processes, technology, operations, and organizational performance. Today, Lori is an entrepreneur and business owner, applying her decades of leadership experience to building her own ventures.

Todd Smith is a former CIO with more than 30 years of experience leading global IT organizations through growth, transformation, and M&A. As Chief Information Officer of Horizon Therapeutics, he built a cloud-first IT organization that supported the company’s growth from $300 million to nearly $4 billion in revenue. He previously led Takeda Pharmaceuticals’ global SAP Center of Excellence and IT integration efforts and began his career at Accenture. He brings the perspective of the enterprise technology leaders UpperEdge advises.

Michael Yadgar is a technology executive and transformation strategist with more than 30 years of experience helping organizations navigate growth and large-scale change. A former partner at Accenture and EY, he built and led global SAP and technology businesses, including growing EY’s SAP practice to $2 billion in revenue. He advises C-suite executives and boards on complex transformation and operational challenges across a wide variety of industries.

The Advisory Board marks the latest step in UpperEdge’s evolution as it expands its advisory services, invests in technology-enabled offerings and continues helping the world’s largest enterprises get the most value from their technology investments.

About UpperEdge

UpperEdge is an independent IT sourcing and vendor negotiation advisory firm serving Fortune 500 and Global 2000 companies. UpperEdge helps enterprises plan, negotiate and manage their most significant technology investments across enterprise software, cloud, AI and IT services. Because it has no vendor affiliations, UpperEdge represents only its clients’ interests. Learn more about the Advisory Board at https://upperedge.com/who-we-are/our-team/advisory-board/.

Media Contact:
Alyssa Meyer
UpperEdge
(810) 569-2605
[email protected]

SOURCE UpperEdge

The AI 3D multimodal model company also launches its official iOS and Android app, giving 3D printing enthusiasts a quicker, simpler way to create 3D models on their phones.

SILICON VALLEY, Calif., Sept. 30, 2026 — Meshy, the world’s leading AI 3D multimodal model company, today announced that its annual recurring revenue (ARR) has surpassed US$100 million, up 100x from US$1 million in under two years, marking one of the fastest revenue ramps in the AI industry and making Meshy the first AI 3D company to reach the milestone.

The milestone caps a period of rapid scaling across every dimension of the business. Meshy is now used by more than 15 million registered users and over 3,000 companies and educational institutions worldwide, who together have generated more than 100 million 3D models on the platform. Half of the world’s ten most valuable companies, by market capitalization or private valuation, are Meshy customers, and teams at 200 of the Fortune Global 500 build with Meshy.

Meshy launched the world’s first publicly accessible generative AI product for 3D in 2023, creating the category, and has led it ever since in model capability, user scale and revenue. This month, the company released Meshy 7.1, raising geometry generation to Ultra 4K resolution, and open-sourced its full alignment benchmark suite for image-to-3D generation: on both the geometry and texture benchmarks, Meshy 7 leads the industry, scoring 5.6 percentage points above the average of the latest comparable systems in end-to-end texture alignment. The $100 million milestone shows that the hardest frontier in multimodal AI now runs as a scaled, sustainable business. In July 2026, the company closed a nearly $400 million Series B at a $1.5 billion valuation, the largest funding round to date in AI 3D.

“Surpassing $100 million in ARR in under two years reflects the growing demand for AI-powered 3D creation,” said Faye Pan, VP of Growth at Meshy. “From game studios and film teams to educators and home 3D printing enthusiasts, our customers are showing just how broadly this technology can be used—and how much opportunity lies ahead. We’re grateful to the customers and creators who have trusted Meshy to bring their ideas to life. Their creativity and support have made this milestone possible.”

Understanding and generation: the division of labor behind the growth

Meshy’s climb from US$1 million to US$100 million in ARR coincided with the fastest two years of progress in general-purpose models, culminating in releases such as GPT-6 Astra. The two curves are connected: frontier models excel at understanding, planning and coding, while Meshy’s diffusion-based foundation models excel at generation, reconstructing every detail of a text prompt or an image into assets ready for a game engine or a 3D printer, typically in under a minute. A general model knows a face has two eyes; Meshy knows exactly how large those eyes should be. The two already work together in production: the spatial understanding of frontier models can drive Meshy’s API to turn a single image into a complete 3D scene, a workflow that was impossible a model generation ago.

The official Meshy mobile app is now live

Alongside the milestone, Meshy announced that its official mobile app is now available on iOS and Android. Powered by Meshy 7, the app turns photos, sketches and text into fully textured 3D models directly on a phone, with printability checks, AI repair, AR preview and export to industry-standard formats including STL, GLB, FBX, OBJ, USDZ and 3MF. Every creation syncs across web and mobile through a single Meshy account, and users can explore and remix models from the Meshy community.

The app makes 3D creation more accessible than ever: anyone can photograph an object, generate a model in about a minute, preview it in AR at real-world scale, and send it to a 3D printer, with the entire workflow running on a phone.

From AI for Work to AI for Fun

Meshy’s mission is an unlimited supply of fun powered by AI. Mora, the research architecture (Multimodal Open-world Real-time Architecture) it recently introduced, generates explorable, interactive worlds in real time; a publicly playable early demonstration, Mora 1, is live at mora.fun. The company’s first AI-native game, Black Box: Infinite Arsenal, generates in-game gear and combat mechanics from player prompts and makes its official public debut at Steam Next Fest at the end of 2026. Nearly all of the industry’s attention and resources today go to AI for work. Meshy is one of the earliest companies building for the phase after it, and among the best prepared.

“Much of AI’s early progress has focused on helping people work more efficiently. We see an equally exciting opportunity to help people create, express themselves and play,” Faye Pan added. “Across games, film and interactive worlds, AI is opening up new possibilities for both creators and audiences. Mora offers a glimpse of that future: worlds people can generate, explore and play in, all in real time. Our ambition is to make the possibilities for creativity and play limitless, and we’re building toward that vision step by step.”

About Meshy

Meshy builds multimodal foundation models for AI-powered 3D generation, turning text and images into production-ready 3D assets for games, film and 3D printing. Founded by Ethan Hu, who holds a Ph.D. from MIT, Meshy serves more than 15 million registered users worldwide. In July 2026, the company raised nearly $400 million in a Series B round at a $1.5 billion valuation, the largest round to date in AI 3D. Learn more at meshy.ai.

SOURCE Meshy