Centific, the Market-Leading Enabler of Advanced AI, Closes Transformative $60M Series A Round USA - English USA - English USA
SEATTLE, June 25, 2025 /CNW/ -- Centific, the AI Data Foundry trusted by the world's top model builders, AI labs, and enterprise innovators, today announced the close of its $60 million Series A funding round, led by long-time Midas Lister Jenny Lee of Granite Asia, an investment firm behind 56 unicorns and 31 IPOs globally.
Founded in 2020, Centific has operated behind the scenes to power many of the world's most transformative AI breakthroughs. Today, the company equips innovators with the infrastructure, oversight and deep expertise required to develop and safely deploy foundational models, multimodal systems and next-generation agentic AI; intelligent platforms that perceive, reason and act at scale.
Centific is a recognized innovation partner to NVIDIA, selected for its leadership in real-world Vision and Language AI inferencing. The company was recently featured by Jensen Huang at both CES and GTC as a leading innovation partner, underscoring its role in helping scale AI from prototype to production.
"Enterprises globally are moving from AI experimentation to enterprise-wide deployment, but the journey requires scale, trust, and deep integration with legacy systems. Centific is built for this moment." Said Jenny Lee, Senior Managing Partner at Granite Asia." With its global delivery infrastructure, strong existing client base, and commitment to safe, responsible AI, it is uniquely positioned to become a foundational partner in the enterprise AI stack. At Granite Asia, we back the builders of enduring infrastructure for the future of industry, and Centific reflects that vision."
Powering the Frontier of AI with safety, speed, and scale
As organizations move beyond model training into large-scale, real-world AI applications, Centific delivers the technical sophistication, governance frameworks, and expert oversight required to succeed.
"AI is evolving from isolated models to fully agentic systems that perceive, reason, and act at scale," said Venkat Rangapuram, CEO of Centific. "Our full-stack Data Foundry meets this moment by combining deep domain expertise, human-in-the-loop assurance, and true multimodal orchestration—all delivered with unmatched speed and scale."
A future-ready platform for real-world AI
Centific's AI Data Foundry brings together a global network of experts across 1000's of domains in STEM, vertical and professional, and consumer and lifestyle, human-in-the-loop assurance with built-in QA checkpoints and escalation paths to eliminate bias, hallucinations, and failures, true multimodal orchestration supporting complex prompts across text, image, video, audio, 3D, and spatial data, and unmatched speed and scalability to deploy, localize, and contextualize AI at the enterprise, citywide, and national levels.
Centific's AI Data Foundry integrates every component organizations need to power safe, scalable agentic AI deployments:
Infrastructure-Agnostic Deployment: Scale models seamlessly across cloud, core, edge, and far-edge environments to deliver truly transformational experiences
High-Precision Model Training: Workflow-orchestrated datasets fine-tune LLMs and domain-specific agents with unparalleled accuracy
Optimized AI Inferencing: Low-latency, high-assurance pipelines for perception-based Vision and Language AI in production
Agentic AI Experiences: Digital twins, avatars, and multi-agent coordination that bring Physical AI to life
Governance & Risk Mitigation: Human-in-the-loop validation, synthetic-data controls, and audit-ready compliance workflows integrated at every stage
Trusted by the Magnificent Seven, top model labs, and Fortune 500 organizations, Centific's Data Foundry provides a secure, repeatable infrastructure that keeps pace with rapid innovation while minimizing operational and regulatory risk.
Deploying capital to drive next-gen AI innovation
With this foundation in place, Centific is channeling this new capital into four key areas:
Expand Functionality & Capabilities: Enhance our platform to accelerate the journey from prototype to production; delivering resilient, agile systems finely tuned for real-world impact.
Supercharge R&D Frontiers: Continue research and innovation programs fueling breakthroughs and pushing the frontiers of model architectures, Vision AI, and specialized AI; driving cutting-edge advancements.
Scale the Enterprise AI Backbone: Cement Centific's position as the foundational partner; powering mission-critical AI deployments at the speed of innovation.
Amplify Strategic Ecosystem Alliances: Deepen collaborations with NVIDIA, Microsoft, AWS, Dell, Lenovo, and GPU-as-a-Service leaders; driving global adoption.
"This funding round isn't about necessity, it's about ambition. Having trained most the world's leading AI models, our Zero Distance Innovation™ ethos now pivots to enterprise impact, unlocking industries' ability to deploy safe, scalable AI at speed and scale." said Venkat Rangapuram, CEO of Centific. "With significant interest from top-tier firms, we selected Granite Asia for their unrivaled track record in scaling early-stage tech innovators. Built profitably on a foundation of trust and global reach, this investment turbocharges our mission: to empower enterprises with AI systems that are resilient, agile, and meticulously fine-tuned."
Looking ahead
As demand for sovereign, secure, and scalable agentic AI infrastructure surges, Centific is perfectly poised to lead the charge. The company envisions AI evolving into an unseen yet indispensable layer of everyday life. Fueled by this vision, Centific is building a unified ecosystem, bringing together world-class talent, robust platforms, and cutting-edge technology, to power the next generation of AI-driven innovation.
Centific powers the world's most advanced AI. As the trusted data and infrastructure partner for thousands of production-grade systems, Centific empowers innovators to scale models and agents across the full lifecycle—faster, safer, and with expert oversight. From perception-based inferencing to HITL-evaluated datasets, Centific is the backbone for the future of AI.
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Calgary Herald
3 days ago
- Calgary Herald
What is ‘sovereign AI' and why is it a priority for Canada in the age of Trump?
Article content Countries with the resources and know-how to forge ahead on AI development and deployment are building AI factories, which are hubs that unite computing power (known as compute), data, researchers and companies to advance AI models and applications. Article content 'The AI factory will become the bedrock of modern economies across the world,' Jensen Huang, founder and chief executive of AI chip giant Nvidia Corp., said during a media Q&A in 2024. Article content The European Union's InvestAI initiative launched in February will mobilize a 200 billion euro AI investment, with 20 billion euros of that reserved for AI gigafactories. At least 15 AI factories are expected to be operational in the EU in 2025 and 2026, including sites in Germany, Finland, France, Italy and Spain. Article content Stargate LLC, the AI joint venture between OpenAI Inc., SoftBank Group Corp., Oracle Corp. and Metagenomi Inc., said it will invest US$500 billion in the coming years to boost the U.S.'s AI infrastructure and to 'secure American leadership in AI.' Article content Article content Canada, meanwhile, announced the Canadian Sovereign AI Compute Strategy in December 2024, which will dole out more than $2 billion over five years to support the domestic AI industry. About half of those funds are earmarked to build out the country's lagging computing infrastructure. Article content Industry leaders and policymakers have for years lamented that while the foundations of AI have their roots in Canada, the country has largely failed to transform its early research lead into tangible economic benefits. Article content The Mark Carney government said it aims to change that. It remains committed to Ottawa's sovereign AI strategy and has also made new promises, such as tax credits to help small businesses adopt AI and to integrate the technology into the public sector. Article content So far, the federal government has only allotted up to $240 million to Toronto's Cohere Inc. to help the company invest in and buy AI compute at a new multi-billion-dollar Ontario data centre to be opened this year and operated by New Jersey-based cloud provider CoreWeave Inc. Article content Article content Evan Solomon, Ottawa's recently minted minister of AI and digital innovation, has said that building Canada's sovereign AI does not require every piece of the tech stack to be Canadian. Article content CIGI's Araya argued the government should fund homegrown tech firms first. Article content 'I wouldn't go looking outside Canada for procurement, (as) it undermines our economic development over time,' he said. 'It also limits our capacity to know exactly how we're managing our data; a huge portion of our data has to go through U.S. infrastructure.' Article content Solomon has also said Canada requires foreign partners, companies and capital to fulfil its vision of sovereign AI. He cited the need to work with the U.S., the United Kingdom, France and Germany, and has signalled that Saudi Arabia could be a potential investor in Canadian AI. Article content Forging strategic alliances is key for Canada's long-term AI strategy, especially if the country can leverage its AI intellectual property (IP) to play from a position of power, said Daniel Wigdor, a University of Toronto computer science professor and the cofounder and chief executive of Axl, a venture studio that aims to build 50 AI companies over the next five years. Article content He said the bulk of AI patents filed in Canada leave the country, with 75 per cent of them going to the tech giants, while only seven per cent of the remaining quarter are assigned to Canadian companies. Article content 'If we can start to recapture that … this strategy can pan out,' he said. 'But if we're just going to sell our stuff to the highest bidder and be dependent on them, there's a big danger in that.' Article content What might sovereign AI look like in the future? Article content Much of the current discussion on sovereign AI revolves around building AI data centre infrastructure and obtaining the energy and cutting-edge chips required to power them. Article content It's a highly costly endeavour, with McKinsey & Co. Inc. estimating in an April report that companies in the compute value chain will need to invest US$5.2 trillion in data centres by 2030 to satiate global AI demand. Article content Araya said Canada is right to invest in data centres, but needs to think beyond that. Article content Article content 'We can't just say we'll take all our public tax dollars and put them into data centres,' he said. 'We've got to think about how we make more money. It's one thing to have the infrastructure in place, but we need the service economy that sits on top of that.' Article content As GenAI models advance, some scholars and industry voices are advocating for an alternative vision of AI that is less power-hungry and more equitable over the bigger-is-better mentality espoused by Silicon Valley's AI giants. Article content Sovereign AI, Wigdor said, doesn't just mean owning the data layer or developing large language models (LLMs). Article content 'If we're going to be spending Canadians' money to establish a sovereign AI play, it should be to leverage our lead in applied computing in AI — applying the technology to solve real-world problems — valuable to people and businesses, and let private industry pay for the rest,' he said. 'That's where we can leap ahead.'
Winnipeg Free Press
4 days ago
- Winnipeg Free Press
Imagining that the machine is human
Opinion When I taught computer science, often on the first day of class, once my excited nerdlings had sat themselves down in front of a computer to begin their quest to become the next Bill Gates and conquer the world, I would flick the classroom lights off and on several times. This was to contain the visible excitement of soon-to-be programmers in their happy place and focus their attention on me, but was also to put a perspective on how computers work and to start my first lesson as such. Computers are stupid. They are no smarter than the light switch that I had been flicking on and off to get their attention. At a very basic level, computers operate in binary, ones and zeros, on and off, like a light switch. In fact, mathematically, computers can only count to one. Not very smart. This would usually cause some furrowed brows among my students who couldn't believe I would be disrespectful to their beloved computers. Of course I would explain further. The simplest storage unit in a computer is the bit (one or zero), and eight bits make a byte. A byte can represent a number, which is more useful to humans. We can do calculations using bytes, and since math was their first use, we call them 'computers.' In class I would introduce my charges to binary and base-2 calculations. We divert from there. Bytes can represent numbers but the same byte could represent a character on a keyboard. Other bytes can represent pixels on a screen and many more can represent an entire picture. An important thing to remember is that the computers still only work with numbers in the background, it is humans that decide what those numbers mean. Computer use took off with the GUI (graphical user interface) that allowed users to use a keyboard and interact with pictures of objects on a screen using a mouse (think Windows 95). Back then this was called the user illusion because it gave everyone the illusion that they were interacting with a computer in a way that was more human-like. Before long, advanced chips, many more leaps in computing power, and access to virtually unlimited data, and the computer can execute very complicated algorithms to produce something even more familiar to us. Maybe eerily like us. But it's still based on layers of illusion. From bit to byte to number to pixel to picture to GUI to touch screen, to VR, to AI, these layers of illusion are still built with number crunching. And so we go from computers that can barely add two numbers together with the help of dozens of mathematicians to machines that think like us. But they don't really think, and that's the illusion. Even the CEO of Nvidia, the most valuable publicly traded company in the world (as of July 16) with a dominant role in powering AI infrastructure wouldn't use AI to think. In a recent interview, founder Jensen Huang said 'I am not asking it (AI) to think for me, I'm asking it to teach me things I don't know.' That's because anyone that understands how computers work know that AI is just another layer of illusion created by advancing technology. There are some out there that will spout platitudes about what AI is doing, both evangelizing and anthropomorphizing AI to be something it is not, to be revered and perhaps to be feared. Because the danger in using AI is not what it can do but what we think it can do. This may lead to humans giving machines agency and power that it doesn't deserve, especially when some of them can't even self-regulate their own 'thoughts.' Just ask Elon Musk, who had to quickly shut down his Grok AI when it spit out hate-filled screeds. And that is mainly the point, AI is a machine that is created by us as a tool. But unlike many tools, this one is a digital doppelganger and we now have a human illusion in AI. But remember that's what we are seeing, and shouldn't see it for what it's not. We have always imitated things we know in new technologies, but AI is no more human than a plane is a bird or a submarine is a fish. For all the bandwagon hopping by everyone quickly trying get into the 'inevitable' adoption of AI, even as the Canadian government created a new minister of artificial intelligence and digital innovation, there needs to be a clear understanding of the purpose of the technology as a tool for human advancement, just like any machine that humans have created. Because, like I might say to my students about their beloved computers, it is just a machine and we can shut it down by turning off the switch. David Nutbean was a technology instructor at RRC Polytech, a digital literacy administrator for a school division, and a computer science teacher. He writes from his home in Oakville, Manitoba.
Canada News.Net
01-08-2025
- Canada News.Net
China demand sparks fresh Nvidia H20 orders despite AI trade risks
BEIJING/SHANGHAI/NEW YORK: Nvidia has placed an order for 300,000 more H20 chips with contract manufacturer TSMC, two sources told Reuters, as strong demand from Chinese clients prompted the U.S. chipmaker to shift course and supplement its existing stockpile. One source said robust interest from Chinese firms drove the decision, just weeks after the Trump administration reversed an effective ban on the sale of H20 GPUs to China. That ban, introduced in April over national security concerns, had threatened Nvidia's foothold in the world's second-largest AI market. "It depends on the level of orders," Nvidia CEO Jensen Huang said during a recent trip to Beijing. He noted that any resumption in production would require about nine months to ramp up. The H20 chip was specifically developed to comply with U.S. export controls after restrictions were imposed in 2023 on more advanced GPUs like the H100 and Blackwell series. Though less powerful, the H20 remains compatible with Nvidia's widely used AI software tools, a key advantage in the Chinese market. Nvidia had previously indicated it would rely on its inventory of 600,000–700,000 H20 chips, but the new orders point to renewed momentum. For comparison, according to SemiAnalysis, Nvidia sold around 1 million H20 units in 2024. The Information reported after Huang's China visit was that Nvidia had limited H20 supplies and no immediate plans to resume wafer production. The recent shift suggests that it has changed, though the orders are still contingent on U.S. export licenses. Nvidia has said it expects to receive approvals soon, but the U.S. Department of Commerce has yet to grant them, according to one of the sources and a third individual familiar with the matter. In the meantime, Nvidia has asked prospective Chinese buyers to submit updated order forecasts and supporting documentation, sources said. Neither Nvidia nor TSMC commented on the orders or license status, and the U.S. Commerce Department also did not respond to requests for comment. The H20's reinstated availability comes amid broader trade negotiations between Washington and Beijing over rare earth magnets—critical components in industries ranging from EVs to defense. Still, the Biden administration's decision to allow resumed sales has sparked bipartisan criticism in Congress, where lawmakers fear it could weaken U.S. efforts to lead in AI innovation. Nvidia and industry analysts argue that continuing to serve Chinese customers is critical. Pulling back, they say, risks driving developers toward domestic alternatives such as Huawei, whose less powerful chips have gained traction in the wake of U.S. bans. Before April's restrictions, Chinese tech giants like Alibaba, ByteDance, and Tencent were already increasing orders of H20 chips, especially for running cost-efficient AI models such as DeepSeek. Even with U.S. sanctions in place, the demand for Nvidia hardware hasn't waned. Smuggling and repairs of older banned chips have surged, further underscoring Nvidia's popularity in China. The company previously warned that the April ban could force it to write off US$5.5 billion in unsold inventory and cost it $15 billion in lost sales.
