The round was co-led by major South Korean venture capital firms Atinum Investment and IMM Investment. It was well supported by a broad syndicate of new and existing institutional investors across Asia, including SBI Investment, Mirae Asset Capital, Corstone Asia, Kiwoom Investment and the Korea Development Bank.

Challenging the Von Neumann Bottleneck

Generally, when dealing with a computer, the architecture in question is that of having data stored in one place, the memory, while it is worked on elsewhere, the CPU or the GPU. For most applications, this physical separation is not a problem, because the distance is just insignificant. For large language models, however, data needs to be moved all the time to increase context window sizes, perform data analyses and generate tokens. This bottleneck of constant data circulating comes at a cost. Latency increases, energy expenditure will become very high, and overall efficiency can only be scaled to a certain degree; this is known as the von Neumann bottleneck or memory wall.

Source: XCENA

XCENA addresses it by means of memory-centric computing, an engineering field that integrates large memory arrays much bigger than typical processor caches right with the computational resources. Instead of moving data back and forth across long, slow links, as conventional processor-accelerator systems do, XCENA performs computation where the data sits.

More News: Pax AI Cuts Crime 27% and Raises $40M Seed Funding

The MX1 Architecture and CXL Standard

The cornerstone of XCENA’s strategy is its flagship MX1 computational memory solution. Using the open standard Compute Express Link (CXL 3.x standard), MX1 can be directly linked to host processors via high-speed cache-coherent lanes. Besides, the device embeds thousands of power-efficient, proprietary RISC-V processing cores and vector engines directly with the pooled DDR5 memory.

In operation, the MX1 takes many of the most intensive, data-centric tasks in the system, such as preprocessing, key-value (KV) cache management, and data caching inside the memory module itself. By removing the data design from the primary server chips, XCENA offers an incredible reduction in infrastructure footprint. Preliminary metrics show that under certain computationally intensive workloads, hardware with no more than a single line of MX1s can produce results that would otherwise have used many expensive real servers.

To simplify adoption, the startup is also packaging its hardware with a full-stack SDK named XFLARE. This SDK enables enterprise customers, telecommunication networks, and research institutions to migrate workflows onto memory-centric hardware with reduced software rewrite.

Roadmap to Production

Operating from two main R&D centers based in Sunnyvale, California, and Banqiao, South Korea, XCENA employs more than 90 engineers. The corporate senior management team has its background firmly in the chips arena. Founded in 2022 by semiconductor veterans from Samsung Electronics and SK Hynix, the world’s biggest and second-biggest memory chip suppliers.

XCENA will be leveraging the Series B runway to build a closer presence in Northern California for working with hyperscale cloud providers and infrastructure players. MX1 architecture is in the prototype stage and starting evaluation with early partners.

This long-term commercialisation approach avoids the high capital costs of establishing internal manufacturing capability. XCENA will instead take a fabless route, sourcing from current manufacturing ecospheres, with mass production of commercial-grade chips set to begin by the end of 2026 through Samsung’s foundry lines.

The post XCENA Raises $135M Series B for Memory-Centric Computing appeared first on Ventureburn.

The big funding round followed an unparalleled proof of concept deployment. As the company has reported, Pax’s live intelligence platform reduced violent crime by 27% after just six months of its largest municipal rollout.

A Data-Driven Cure for an Epidemic of Violence

The protection of public safety is, to this day, an extremely high-stakes problem in Latin America, where violence is one of the most pressing social and economic problems. According to calculations by the Inter-American Development Bank, crime and violence represented a drain of some 3.5% of regional gross domestic product, causing a regional loss of about $241 billion. In Brazil, where Pax is now testing its technology, an average of 40,000 murders are registered annually; only 39% are theoretically ever investigated to a conclusion.

Source: Pax

Pax’s founding team assert that this systemic failure is not due to a lack of effort by officers but simply amounts to a massive fragmentation of data.

David Peixoto, CEO and founder of Pax AI, has experienced scaling enormous businesses in Brazil before, from edtech titans like Arco, which he grew to become the first Brazilian company to list on the Nasdaq. To grow Pax, Peixoto recruited a world-class engineering team, drawing engineers from Stanford, Harvard, MIT, and Brazil’s best schools, ITA and USP.

More News: Canals Secures $35M to Expand AI Automation

How the Pax Intelligence Graph Works

Pax is used as a single intelligence layer plugged into the current physical infrastructure. It does not demand any new, high-cost hardware; instead, it uses thousands of municipal security and state security cameras already on the street.

Once the data enters the system, Pax ML machine learning systems convert these separate points – a vehicle model, a partial plate, a location, and an incident report – into a dynamic intelligence graph. Then, if said car hits a camera associated with an armed robbery, an instant alert would be sent to the following officers, showing the most probable path.

It’s changing the way local police work. In addition to reducing violent crime by 27%, police forces working with Pax have already cracked over 2,000 difficult criminal cases in 30 cities, covering everything from homicide and carjacking to theft rings. Most importantly, Pax more than doubled police productivity while increasing perceived neighbourhood safety by 59%.

Addressing the Ethics of AI Policing

Supporters of civil liberty across the world have raised many concerns about the dangers of AI in policing, including structural concerns over the development of a biased and disproportionate algorithmic system, as well as the development of a kind of surveillance state that is invasive and permanent in nature. Pax has explicitly shown how the friction points can be mitigated by accommodating strong safeguards directly into the software platform.

The platform provides an absolute audit trail. Every database query, video search, and alert interaction gets individually written to an indisputable, authorised, declared operator tied to it. Peixoto also insisted the tool is a silent analytical copilot, not an automaton. ‘The tool doesn’t execute; it researches. ‘The officer researches; the officer executes,’ Peixoto said. ‘The platform merely multiplies their awesomeness.’

Given its infusion of $40 million of new capital, the company is planning an aggressive expansion beyond its original 30 Brazilian cities in order to take its real-time intelligence platform to the wider Latin American region.

The post Pax AI Cuts Crime 27% and Raises $40M Seed Funding appeared first on Ventureburn.

The size of the raise, which was completed right after the company went public, demonstrates the increasing demand from venture capitalists for fundamental infrastructure that can scale the next frontier of physical AI and generative video.

From Prompts to Living Environments

Until now, the world of generative video has been one of prompts and waits. Existing tools such as OpenAI’s Sora or Runway enable simple inputs of text, followed by a long pause and a static, non-editable video file. These outputs, while artful, are one-shot clips, static and cannot be interacted with.

Reactor is based on a radically different principle: world models. Instead of producing a single line of pixels, a world model creates a long-lasting whole world with its own rules (physics, space, logic, etc.) and not simply a movie to be observed but a place where the user and developer can literally go in and interact with the environment.

Source: Reactor

“World models are redefining what AI can do, moving from systems that generate content in isolation to ones that perceive and respond in real time,” said Alberto Taiuti, co-founder and CEO of Reactor, in the company’s launch statement. “We are building the critical layer between the model labs and the developers who want to create with them. This is about enabling a new form of media, one where experiences that weren’t previously possible are generated live, and anyone can build and distribute them.”

More News: Canals Secures $35M to Expand AI Automation

Pedigree from Apple Vision Pro and Luma AI

The technical mountain Reactor is trying to climb requires rare expertise in spatial computing, real-time rendering, and low-latency systems. Fortunately, the startup’s founding duo possesses exactly that pedigree.

CEO Alberto Taiuti is a prominent figure in the 3D generative space, having previously served as the co-founder and Chief Technology Officer of Luma AI, where he designed the core infrastructure behind some of the market’s most popular 3D and video generation tools. Joining him as Chief Technology Officer is Bryce Schmidtchen. Both Taiuti and Schmidtchen previously worked alongside each other at Apple, where they served as core technical leads developing the Apple Vision Pro spatial computing headset.

Over the past several months, the co-founders have leveraged their network to quietly assemble an elite engineering roster. Reactor’s broader team pulls together researchers and systems experts from top-tier tech and entertainment giants, including Meta, Google, Netflix, Adobe, Microsoft, and Replicate.

The Strategic Ecosystem: Hollywood and AWS

The commercial implications of real-time world generation have already attracted heavy hitters from both entertainment and cloud computing. Veteran media executive Jeffrey Katzenberg, Founding Partner at WndrCo, has joined Reactor as a board observer.

“Every major shift in media has been driven by new tools that expand what creators can do,” Katzenberg noted. “AI is a transformative moment, but the real opportunity lies in making these technologies usable at scale. Reactor is building the platform that can enable a new generation of storytelling and interactive experiences.” The platform has already begun live deployments. Gaming and media studio Overworld is currently utilising Reactor infrastructure to create entirely interactive live entertainment environments.

To run the compute-heavy workloads, Reactor has also announced Amazon Web Services (AWS) as its preferred cloud vendor. The pact will try to heavily reduce the expense of live inference by utilising custom chips and optimised AI services so that running live AI worlds becomes affordable to most application programmers.

The post Reactor Raises $59M to Build Real-Time AI Worlds appeared first on Ventureburn.

The funding round was led exclusively by Base10 Partners, a prominent venture capital firm known for backing technological innovations that automate the real economy. The firm’s portfolio includes industry giants like Stripe, Figma, and Notion. 

The new capital will be used to aggressively expand Canals’ engineering footprint and rapidly build out additional end-to-end transactional tools tailored for global suppliers, distributors, and contractors.

Automating the Complexities of the Real Economy

While high-profile generative AI startups frequently target consumer tech or creative industries, Canals has taken a highly specialised approach by tackling the foundational backend friction points of global commerce. Wholesale distribution represents a massive $8.2 trillion global ecosystem. However, it remains heavily reliant on manual processes due to the inherent complexity of its supply chains.

In a typical distribution workflow, transaction inputs are messy and fragmented. Invoices, purchase orders and product lists arrive from thousands of distinct vendors in mismatched formats ranging from unformatted spreadsheets, handwritten physical notes, and unstructured PDFs to voice messages and complex email chains. Historically, processing these inputs required manual review by experienced staff who relied on years of institutional product knowledge to map data accurately into enterprise resource planning (ERP) systems.

Source: canals ai

Canals bridges this gap by offering specialised enterprise AI models that interpret unstructured data with immense precision. Instead of forcing distributors or their clients to conform to rigid, pre-defined templates, the Canals platform instantly reads, extracts, and translates incoming client asks into ERP-ready, structured documents within minutes.

“Our customers operate in the real world, with messy inputs, imperfect data, and constant change, which made automation difficult to impossible before AI,” said Michael Delgado, CEO and co-founder of Canals, in a statement accompanying the announcement. “We’ve built the most accurate AI for wholesale distribution to finally free teams from manual data entry and enable stronger relationships across the supply chain.”

More News: Orbital Industries Raises $50M to Scale AI Hardware

Capital Efficiency and Major Enterprise Trust

One of the most remarkable aspects of Canals’ trajectory is its capital efficiency. Before securing its Series A round from Base10 Partners, the company achieved massive enterprise scale entirely through organic growth and customer revenue, operating without initial institutional venture backing.

To date, the Canals platform has processed more than 8 million sales orders and eclipsed $5 billion in payables across its user base. The company’s software is currently deployed at more than 100 wholesale distributors globally. This roster includes prominent industry market leaders such as DSG, Locke Supply, and The Kendall Group.

According to customer metrics cited by the company, distributors utilising Canals have successfully doubled their quote conversion rates due to drastically faster response times. Furthermore, backend operations have achieved up to 96% touchless invoice processing, allowing finance teams to pivot from manually fixing individual line-item invoice errors to supervising high-level corporate payment strategies.

“From conversations with many Canals customers as well as our own contacts in wholesale distribution, we kept hearing the same thing over and over again: Canals is customer-obsessed and delivers outstanding ROI,” noted Jason Kong, General Partner at Base10 Partners. “They’re outperforming others marketing similar solutions by orders of magnitude, so it was an easy decision to partner with Michael and this incredible team.”

With $35 million in fresh capital and an entrenched market position, Canals is poised to lead the digital transformation of industrial logistics, turning manual data friction into an absolute competitive advantage for global supply chains.

The post Canals Secures $35M to Expand AI Automation appeared first on Ventureburn.

Global venture capital firm Headline led the growth capital round. It also drew backing from new institutional investors Eurazeo and Forestay Capital alongside a powerful contingent of returning investors, including Index Ventures, OMERS, Coatue Management, and Dawn Capital. 

This massive cash injection and simultaneous asset acquisition position Fonoa to rewrite the rules of international indirect tax, transitioning corporate finance from manual, periodic reporting to completely real-time automation.

Solving the Fragile, Segmented Tax Stack

Founded in 2019 by three former Uber executives, Davor Tremac, Filip Sturman, and Ivan Ivanković, Fonoa was born directly out of the operational frustrations of managing hyper-growth, borderless digital platforms. While ride-sharing, streaming services, and digital marketplaces scale effortlessly across geographical boundaries, their underlying financial teams are often forced to confront an incredibly complex, fragmented, and antiquated global tax system.

Source: Unsplash

“While technology has transformed much of finance, tax systems have remained neglected, leaving accounting teams to manage the same fragmented stack for decades,” explained Davor Tremac, CEO and co-founder of Fonoa. “Companies have historically used one vendor for tax ID determination, another for digital e-invoicing, and a third for filing regional returns, with manual spreadsheets desperately holding the pieces together.”

This patchwork workflow has become completely unsustainable as governments worldwide aggressively clamp down on digital tax evasion. Regulatory bodies are rapidly shifting away from historical, retroactive monthly or quarterly tax reporting. Instead, countries are mandating real-time transaction reporting and automated e-invoicing, meaning tax compliance must now happen at the exact millisecond a consumer hits a “buy” or “book” button.

More News: Orbital Industries Raises $50M to Scale AI Hardware

Fusing Infrastructure with PwC’s Edge

The strategic crown jewel of Fonoa’s dual announcement is its acquisition of PwC’s Indirect Tax Edge software platform. Used by some of the largest multinationals in the world, Edge is an enterprise-grade compliance application specialising in Value Added Tax (VAT) and Goods and Services Tax (GST) management, automated e-filing, and macro-level tax data analytics.

The deal perfectly bridges a historic product gap in the TaxTech landscape. By combining Fonoa’s native API infrastructure with the advanced filing capabilities of Edge, Fonoa now hosts a unified, single data model that manages the entire lifecycle of indirect tax.

Within this integrated loop, AI agents autonomously track changing tax rules across jurisdictions, actively monitor transaction streams, flag reporting anomalies, and assemble audit packs in seconds. Under the structured terms of the deal, PwC will continue to deliver deep indirect tax consulting and managed services to its global clients using Fonoa’s upgraded infrastructure.

Explosive Transaction Volume and Global Reach

Fonoa’s platform is fundamentally optimised for digital-first enterprises. Unlike traditional legacy tax software providers built around slow enterprise resource planning (ERP) batch updates, Fonoa is entirely cloud-native. The efficiency gains are stark, with some enterprise clients reporting up to a 90% reduction in tax calculation latency.

Today, Fonoa boasts an established operational footprint that includes the following:

• 190+ Jurisdictions: Real-time automated tax determination.
• 100+ Countries: Instant, programmatic corporate tax ID validation.
• 1 Billion+ Transactions: Safely processed and logged through its architecture annually.

This massive scale has turned the startup into the primary compliance engine for a roster of the world’s most recognisable digital brands, including Netflix, Uber, Canva, Booking.com, and Nebius.

With $110 million in fresh capital, Fonoa plans to aggressively expand its global engineering capacity and integrate more predictive AI models into its core modules. As international tax bodies lean into real-time digital scrutiny, Fonoa is transforming compliance from an ongoing corporate liability into an invisible, fully autonomous utility.

The post Fonoa Raises $110M to Expand AI Tax Platform appeared first on Ventureburn.

This jaw-dropping round was led by industry giants Altimeter Capital, Dragoneer, Greenoaks, and Sequoia Capital. Also included in this round are the co-leads: Capital Group, Coatue, D1 Capital Partners, GIC, ICONIQ, and XN.

Including spending commitments from tech hyperscalers, who have already invested more than $15 billion in this area, with Amazon contributing an additional $5 billion.

Explosive Commercial Traction Shakes the Market

The scale of the fundraiser alone is amazing, but it is also heavily overshadowed by Anthropic’s astronomical revenue growth. The firm announced it has an estimated yearly run rate of revenue exceeding $47 billion, an incredible jump from the $14 billion run rate just a few months ago in its Series G.

Source: Unsplash

One of the largest driving forces behind this exponential growth has been the organisation’s wide adoption of the Claude ecosystem, especially tools aimed towards developers. Initiatives like Claude Code and the Model Context Protocol (MCP) have transformed the startup from a small research lab into a critical component of core infrastructure for the largest organisations in the world.

“Claude is becoming more and more vital for our expanding worldwide customer base, and we’re doing everything we can to make tools such as Claude Code and Cowork as useful as possible,” said Anthropic CFO Krishna Rao. “This investment allows us to meet the unprecedented demand we are currently seeing, continue to lead in research, and have Claude in more of the workspaces.”

More News: Orbital Industries Raises $50M to Scale AI Hardware

Securing the “Physical” Supply Chain of AI

In addition to investing cash, the Series H is an aggressive play to acquire the physical compute and hardware that support training-frontier models.

Anthropic explicitly brought in all the key strategic infrastructure collaborators at once, including Samsung, Micron, and SK hynix, the pillars of worldwide memory, storage, and logic chip supply chains. By cementing strong collaborations with these hardware developers, the company hopes to insulate itself from the current global semiconductor crisis.

At the same time, Anthropic has radically expanded its computational footprint through huge energy deals and infrastructure, such as:

• Amazon: New capacity of up to 5GW.
• Google & Broadcom: 5-gigawatt next-generation tensor processing unit (TPU).
• SpaceX: Direct connection to cluster capacity on the Colossus 1 and Colossus 2 networks.

An Eye on the Public Markets

Alongside the new funding, Anthropic launched its latest model upgrade, Claude Opus 4.8, which provides enhanced performance for agentic work, complex coding, and professional execution.

The timing of this great round is no accident. Wall Street players and venture capital investors see this Series H as the step before the much-anticipated public filing. Anthropic and OpenAI are said to be working on private draft documents for their first IPOs, heating up the race from Silicon Valley darlings to public tech giants.

The post Anthropic Raises $65B in Historic AI Funding Round appeared first on Ventureburn.

The round was oversubscribed and was led by leading European venture fund Balderton Capital, with new investor Crosspoint Capital, with original pre-seed and seed investors General Catalyst and Ten Eleven Ventures also providing substantial support. The new fundraiser lifts total funding to date for Geordie AI to $36.5 million and valuation to $155 million.

Managing the Shadow Sprawl of Autonomous Agents

Founded by seasoned execs from Darktrace, a heavily funded all-in-one SaaS security company, and Snyka, a developer-focused cybersecurity startup, Geordie AI tackles one interesting operational risk that is more often lurking at the back: an unseen proliferation of non-human, autonomous digital workers.

As enterprises aggressively transition from passive generative AI chatbots to fully autonomous AI agents which can execute multi-step tasks, access private files, write code, and call external APIs, security teams are losing visibility over what these digital entities are actually doing. While an employee might manually query an app, an enterprise AI agent might autonomously interact with dozens of backend corporate databases, software integrations, and third-party cloud services in seconds.

Source: Unsplash

“The core issue isn’t just building these agents; it’s understanding what they have access to and what they could potentially expose,” explained Henry Comfort, CEO and co-founder of Geordie AI. “Enterprises are rapidly deploying an entire workforce of digital employees that operate autonomously, meaning they make autonomous, real-time decisions. If left unchecked, over-privileged agents can trigger data leaks, silent operational failures, or cascading code errors.”

The market demand for specialised oversight is reflected in the company’s early metrics. Geordie AI reported an astounding 1,300% growth in annual recurring revenue (ARR) over the first five months of the year. The platform has already been integrated across approximately 30 live customer environments, including institutional financial data giant AlphaSense and Franco-American biotech researcher Owkin.

More News: Transition Ventures Raises $150M for AI Startups

Real-Time Posture Management and the Beam Mitigation Engine

Geordie AI distinguishes itself from standard log-auditing software by operating as a runtime governance layer. The platform connects via cloud infrastructure, codebase, or endpoints to achieve full visibility across an entire corporate tech stack in under 10 minutes. 

Once active, it continuously maps an enterprise’s agentic ecosystem, inventorying every internal and third-party AI agent, analysing their configuration postures, and identifying over-permissioned access paths. Beyond merely identifying risks, Geordie AI enables active intervention through its proprietary mitigation module, named Beam.

Beam serves as a proactive, real-time context-engineering engine. Instead of overloading an AI agent’s narrow context window with endless security rules, Beam dynamically monitors the agent’s behavioural traces and workflow decisions as they happen. If an agent attempts an off-goal manoeuvre, such as pulling restricted client records or misinterpreting a corrupted prompt, Beam intercepts and steers the agent back into compliance with strict corporate safety policies. This defence-in-depth approach prevents system exploits without forcing security teams to shut down the underlying AI tool entirely.

Capitalizing on the RSAC Innovation Win

The Series A investment concludes a string of marquee industry wins for the young London company: In March, Geordie AI was awarded first place at the 2026 RSA Conference (RSAC) Innovation Sandbox, long regarded as a stamp of approval for the market, defining cybersecurity firms prior to global scaling.

“The organisations today that can safely approve and deploy autonomous agents are the ones capturing a massive competitive advantage,” added Comfort. “Our goal is to act as the primary, immutable source of truth that turns AI agent governance from an IT roadblock into a safe growth lever.”

Management intends to deploy the $30 million to expand its global engineering core and aggressively scale its commercial footprint in the United States, cementing its role as the foundational security infrastructure for the era of autonomous enterprise software.

The post Geordie Raises $30M to Secure Enterprise AI Agents appeared first on Ventureburn.

The round was led by European deep-tech growth fund Plural with strategic participation from Nvidia Ventures. Other institutions to increase their stake in the company include Radical Ventures, Compound and Fly Ventures. 

In addition to the cash injection, the company published its newly revamped image from Orbital Materials to Orbital Industries to indicate its widened commercial emphasis to scale up industrial tech from the atom to fully scaled computing infrastructure.

The Thermal and Spatial Bottleneck of Frontier AI

As tech giants like OpenAI, Anthropic, and Meta race to deploy massive frontier models, they are running into severe physical walls. Next-generation AI chips generate extreme levels of heat inside increasingly dense environments, pushing standard water-based cooling systems past their physical limits. Simultaneously, traditional data centre permitting and construction timelines can drag out for three to seven years, creating a massive supply lag.

Source: orbital industries

Originally founded in 2022 by Chief Executive Officer Jonathan Godwin (formerly a machine learning researcher at Google DeepMind), Chief Technology Officer James Gin-Pollock, and Chief Operating Officer Daniel Miodovnik, the startup initially focused on using AI to discover materials for carbon capture and sustainable aviation fuels. However, intense market pull from the infrastructure sector led the founders to pivot toward solving the hardware constraints of advanced AI data centres.

“When people imagine a better future, they think about physical things: technologies that give them more freedom, more time, more life. AI will get us there faster,” said CEO Jonathan Godwin in a statement. “Frontier AI gives us PhD-level expertise across every discipline, meaning small, agile teams can move from materials discovery to commercial hardware in a way that simply wasn’t possible before. What used to take a decade, we can now do in months.”

More News: Capchase Raises $200M+ to Expand Financing Platform

Molecular AI Meets Modular Engineering

Orbital Industries separates itself from standard hardware manufacturers through its underlying technical asset: Orb, the company’s open-source AI model designed to simulate the quantum-mechanical behaviour of atoms. Instead of relying on a decade of trial-and-error laboratory experiments, Orbital uses its proprietary AI loop to design advanced chemical structures from scratch.

The capital from the Series B will scale two core product lines:

1. PFAS-Free Cooling Fluid: An AI-formulated, non-toxic, and non-hazardous liquid designed specifically for immersion cooling setups. This fluid protects high-density GPU architectures (such as Nvidia’s Blackwell chips) from overheating without relying on harmful forever chemicals. This line is supported by a multi-year partnership with Amazon Web Services (AWS) to optimise data centre decarbonisation and water utilisation.
2. Modular Compute Infrastructure: Prefabricated, off-site manufactured data centre units designed by AI algorithms to house ultra-dense computing clusters. Orbital claims this approach compresses traditional three-year construction pipelines down to just six months, allowing operators to deploy plug-and-play high-density facilities exactly when and where energy grids have available capacity.

Institutional Validation

The participation of Nvidia’s NVentures stands out as a strong industry validation. The chipmaker rarely backs physical infrastructure startups directly, signalling that Orbital’s molecularly designed cooling fluid is a highly plausible answer to the immense thermal output expected from upcoming chip generations.

“The capacity obstacle in AI is entirely real,” noted Plural partner Ian Hogarth, who led the investment. “Orbital is addressing the fundamental constraints of power, cooling, and speed of deployment. By bypassing traditional chemical development cycles using their Orb platform, they are establishing an entirely new category of automated industrial hardware design.”

With engineering hubs split between London and San Francisco, Orbital Industries plans to use the $50 million to rapidly expand its deployment team and continue upgrading its atomic simulation models, with an eye toward eventually applying its AI materials platform to semiconductors, critical minerals, and aerospace.

The post Orbital Industries Raises $50M to Scale AI Hardware appeared first on Ventureburn.

The company raised a seed round, which was led by Transformation Capital, a leading growth equity venture firm focused on digital health. The round included participation by the company’s existing seed investors, Twelve Below and Virtue Ventures, as well as multiple strategic angel investors. The new round increased Solstice’s funding to roughly $25 million.

The capital will be used to fund the company’s go-to-market expansion, accelerate product engineering, and substantially grow the company’s own product and customer success teams.

The Multi-Billion Dollar Bottleneck

The global market for biopharmaceutical companies is forecasted to go above $100 billion in drug sales this year alone. The huge budget, however, has little impact; launching and promoting a therapy is still restrained by old workflows.

The most common offender is the Medical, Legal and Regulatory (MLR) review. Since the advertising of medicines is heavily controlled for reasons of patient safety, each piece of material – whether it is a complicated clinical take, an out-of-pamphlet aimed at health professionals or an introductory electronic message aimed at the wider world – has to be reviewed thoroughly for balance, medical content and legal validity.

Source: Solstice Health

The average time it takes to move a single marketing asset through prescribed internal MLR processes is three months and at least three cycles. With the time period for drug patents becoming more rigid, the consequence of an MLR review taking longer than expected is a delay in access to the drug for patients and a regression of profits. “Few business challenges are as high-stakes and time-sensitive as commercializing a new drug,” said Aris Saxena, co-founder and CEO of Solstice. “We built Solstice to ensure that we can substantiate all of a product’s marketing claims using clinical data before they even reach the MLR desk. While it used to take us three to five review cycles, we now merge that down to one or two, compressing months into just 10 days.”

More News: Capchase Raises $200M+ to Expand Financing Platform

AI Built for Strict Compliance

Set up in 2023 by Li Yiwen and Aris Saxena, Solstice sidesteps the trap of being a one-size-fits-all generative AI littered with ‘hallucinations’ that aren’t appropriate in heavily regulated industries. Instead, Solstice is a kind of AI-native appliance that integrates its three most basic pillars of commercialization:

• Clinical Grounding: The system automatically queries drafted medical claims against dense clinical literature and trial data, fact-checking for non-compliant language or early regulatory risk.
• Unified Content Creation: It can help marketing teams create on-brand, very precise educational content created specifically for physicians or consumers.
• Performance Metrics: The platform tracks which compliant content is successfully driving physician adoption and patient engagement.

Through automating the otherwise time-consuming claim mapping when an ad marketer must manually match each sentence with a particular page of a clinical trial report, Solstice, for the first several corporate users, has eliminated up to 84% of operational inefficiencies.

The Shift to “Speed-to-Market” Advantage

With the pharmaceutical playing field increasingly crowded with niche biologics and rare, disease-specific companies, commercial agility has become a differentiator. For investors, Solstice’s specialist has become less a luxury and more a necessity.

With its new $21M balance sheet, Solstice is the one company that can really turn from a promising software vendor to the de facto OS for biopharma commercialization. By allowing brands to cut their regulatory turnaround times to less than 48 hours in some cases, Solstice has quietly defined the way forward for the delivery of modern medicine to the market.

The post Solstice Secures $21M Series A for Pharma Innovation appeared first on Ventureburn.

Transition Ventures was established in 2021 by a team of well-known operators and investors. David Helgason, once the co-founder and former long-time CEO of gaming software giant Unity Technologies, led the firm. Transition Ventures initially launched with an early-stage climate technology mandate, but the exponential growth of generative AI has sent pressures into a whole new dimension on electrical grids that serve the world, vital material supplies, and hardware pipelines.

As a reflection, Transition Ventures has since changed its central thesis to “Physical AI”, which is the use of next-generation machine intelligence to rebuild, perfect, and safeguard essential real-world infrastructure. 

The Pivot to Physical Bottlenecks

The company’s strategy is a literal solution to the “gigawatt-scale” data center bottleneck. With AI training models requiring exponential leaps in compute, the energy grid and computer chip supply chains have become the primary problems for tech hyperscalers.

Source: Unsplash

“The data center industry was already a huge market, but now the AI revolution is driving demand exponentially higher,” said David Helgason, Partner at Transition Ventures. “The fundamental challenge of our era has moved on to physical limitations. How to put more and more computing into the materials we have and the energy we can produce. The traditional venture model of funding continuous iterative software progress has reached a dead end.”

Transitions Fund II will make checks from pre-seed all the way through Series A and will focus on deep-tech startups that marry innovative software with physical engineering. This U.K./U.S. cross-Atlantic institution works out of major European and U.S. centers (namely London and New York), bringing deep tech research across the Atlantic into institutional deployments in the U.S.

More News: Capchase Raises $200M+ to Expand Financing Platform

An Elite Operator Bench

The funds raised reveal the strength of Transition’s core team. Together with the Helgason brothers (David Helgason & Ari Helgason), Transition has strong institutional and venture experience:

• Kristian Branaes was previously at Atomico and the Canada Pension Plan (CPP) Investment Board.
• Clara Ricard (Investment Alumni from Balderton Capital).
• Mona Alsubaei was a former investor at Union Square Ventures’ climate sleeve NY anchor.
• David Pacák already has hands-on experience with early-stage execution from Earlybird and Picus Capital.

All the team members have been either founding or scaling companies worth more than $15 billion, within software and deeptech ecosystems.

A Battle-Tested Portfolio

Transition’s pivot is also supported by a handful of superstar winners in its active portfolio. Transition backed Olix Computing early on, a photonics computing startup that replaces electricity with light to perform AI calculations. Olix has seen a valuation spike to an approaching $1 billion.

Other notable portfolios illustrating the physical AI thesis include:

• Applied Atomics: A business set up by ex-SpaceX engineers that is working on manufacturing a small modular reactor (SMR) that can productively supply colocation data centers with their own dedicated zero-carbon nuclear power.
• Seneca: A comprehensive ground and air autonomous drone network for ultra-fast AI-based forest fire abatement.
• Invisix: A dedicated hardware specialist in the field of semiconductor metrology (measurement & inspection of microchips during manufacture).

Transition’s initial climate reports were high-profile, with the bankruptcy of ocean carbon-removal startup Running Tide in 2024 amid a collapse of voluntary carbon markets. However, Transition’s pivot into physical AI infrastructure has hit home with institutional LPs. Recasting climate resilience as power generation, data centre efficiency, and resource use efficiency has landed Transition as one of the most resilient late 2020s investment mandates.

The post Transition Ventures Raises $150M for AI Startups appeared first on Ventureburn.