Five People

LoopX Raises $4.05M to Build Physical AI for Mining

Mining has a visibility problem that becomes a safety problem fast. Dust blinds cameras. Darkness erases detail. Massive equipment moves where perception cannot afford an off day. LoopX Innovation Inc. just closed a $4.05M Seed round, putting fresh capital behind Physical AI built for those conditions.

Announced August 4, 2026, the financing was led by BDC Seed Venture Fund and Orion Industrial Ventures, with returning investor Hatch participating. Founded in January 2022, the Toronto- and Sudbury-based company was started by Dr. Chao Yu, CEO; Dr. Lucas (Yaodong) Cui, CTO/CIO; James Zhang, Head of Product; and original co-founder Bruce Huang, whose current status is unconfirmed. LoopX began with autonomous last-mile delivery robots before moving its AI and robotics into mining, where software answers to rock, steel, people, and physics.

LoopX combines Vision AI, LiDAR, RF sensor fusion, and edge computing to create situational awareness through dust, glare, and total darkness. CAS provides 360-degree perception across line-of-sight and non-line-of-sight conditions. LoadSight uses 3D LiDAR and edge AI to monitor payloads, cycle times, and load-haul-dump performance, while BoomGuard addresses safety around drills, bolters, and other boom-equipped machines. Seeing a problem earlier matters more underground than explaining it afterward.

The traction gives this round its weight. With MICA support, LoopX moved from prototype to deployments across six mines in Canada, the United States, and South Africa, creating six full-time jobs. The company has filed five U.S. patents. Cementation Americas has installed and tested CAS across multiple pieces of mining equipment. In July 2026, LoopX signed a three-year R&D partnership with Vale Base Metals to develop and validate Physical AI for underground autonomous mining fleets in live production environments.

LoopX also received $480K in FedNor support for commercialization. Generative AI is being integrated into LoadSight so operators can query payload trends and loading inefficiencies through natural language, while the company develops its next-generation Vision-Language-Action Physical AI model. The Seed capital will expand applied AI engineering, customer success, and go-to-market capabilities.

BDC Seed Venture Fund, Orion Industrial Ventures, and Hatch are backing a company that reached Seed with deployments, patents, industrial relationships, and field validation taking shape. Dr. Chao Yu brought autonomous driving and vehicle electrification experience from General Motors, while Dr. Lucas (Yaodong) Cui brought deep learning, autonomous systems, and perception expertise. Technical pedigree gets attention. Evidence gets capital.

For mine operators pursuing safer equipment interaction, better productivity intelligence, and autonomous fleets, LoopX is making its case where conditions are uncontrolled and results have nowhere to hide. What comes next is how much responsibility operators will trust machines to carry underground.

Opal Therapeutics Lands $1.25M NSF Award for Uterine Organoids

Opal Therapeutics just secured a $1.25M National Science Foundation NSF SBIR Phase II award to tackle a stubborn problem in women’s health drug discovery: researchers studying human reproductive biology have often relied on models that cannot reproduce the human menstrual cycle. Opal is betting that better answers start with better biology, and the National Science Foundation is putting capital behind that bet.

Founded in 2023 in San Francisco by Dr. Morgan Stanton, PhD, Opal develops patient-derived uterine organoids, including endometrial and myometrial models, for drug discovery across endometriosis, fibroids, and other gynecological conditions. The platform offers a human-derived alternative to rodent models that cannot replicate key elements of menstrual-cycle biology. Change the model, and researchers can change the questions they are capable of asking.

The $1.25M in non-dilutive funding will support expanded lab infrastructure, organoid manufacturing capacity, and imaging and inflammation assay throughput. Opal previously received approximately $275K in NSF funding. The company did not disclose a valuation, and no venture investors are confirmed. For an early-stage biotech, using grant capital to expand technical capability without making dilution the price of scientific progress is a meaningful strategy.

Opal has reported producing organoids at a scale of tens of thousands, securing Papaya as its first paid customer, and signing a Material Transfer Agreement with Baylor College of Medicine to expand its tissue biobank. The sequence tells the business story: build difficult infrastructure, demonstrate commercial demand, expand the biological asset, then secure capital aimed at the next scaling challenge. Fundraising gets considerably more persuasive when the science has receipts.

Dr. Morgan Stanton, CEO, brings a PhD in Chemistry from Worcester Polytechnic Institute, postdoctoral research at the Max Planck Institute for Intelligent Systems, and previous discovery biology leadership at Herophilus. Opal’s advisors include Dr. Aabid Shariff, PhD; Dr. Matteo A. Molè, PhD; Carolyn Kreuzkamp, MBA; Alexia Krispin, MBA; Dr. Scott Lundeen, PhD; Dr. Jean-Charles JC Neel, PhD; and Dr. Hugh Taylor.

For pharma and biotech teams developing treatments in women’s health, more relevant models can influence which compounds survive scrutiny, how quickly weak assumptions encounter human biology, and which programs become practical to pursue. Patient-derived models do not guarantee better drugs, but they can give developers a more relevant place to start.

The $1.25M NSF award now gives Dr. Morgan Stanton and Opal Therapeutics more capacity to bring that biology into view at scale. What drug discovery uncovers when women’s health gets models designed around the biology researchers were trying to understand is the part worth watching.

LifeMine Therapeutics Raises $263M to Advance LIFE-001

LifeMine Therapeutics just pulled $263M from the capital markets by going somewhere drug discovery has spent decades underestimating: fungi. Machine learning, human genetics, genomics, bioinformatics, and synthetic biology are the tools. LIFE-001 is now the clinical test of whether that hunt can deliver something transplant medicine badly needs.

The financing combines an oversubscribed $188M Series E led by Milky Way Investments with a previously completed $75M Series D, bringing total funding to $558M. Bezos Expeditions, Gates Frontier, and RA Capital Management joined as new investors, while GV, LoLa Capital Partners, GSK, Invus, and ARCH Venture Partners continued their backing. The thesis is straightforward: nature has already manufactured extraordinary chemistry. The opportunity is finding it, engineering it, and turning it into medicine.

The immediate focus is LIFE-001, a precision-engineered, long-acting calcineurin activation inhibitor designed for organ transplantation. Unlike legacy calcineurin inhibitors such as cyclosporine and tacrolimus, LIFE-001 was designed de novo to avoid immunophilin-dependent toxicity. More than 1.6M people worldwide live with transplanted organs, making the equation consequential: prevent rejection without forcing patients to trade one medical problem for another.

LifeMine's Phase 1 SAD/MAD study has enrolled more than 120 adults, with no clinically meaningful renal, metabolic, or cardiovascular safety signals reported to date. The $263M provides runway for a Phase 2 kidney transplant study and a Phase 1b islet cell transplant study planned for early 2027. This is where financing stops being a headline and becomes clinical execution.

LifeMine consolidated operations in Watertown in 2025, prioritized LIFE-001, and built evidence before raising the next round. Its Top-Down Drug Discovery platform integrates human genetics, genomics, bioinformatics, machine learning, and synthetic biology to mine small molecules from fungi. Its GSK collaboration, launched in 2021-2022 with $70M upfront, targets up to three undisclosed drug candidates. Focus gets more persuasive when the platform has external validation and the lead asset has human data.

Driving that execution are Gregory Verdine, Ph.D., CEO; WeiQing Zhou, MBA, COO; and Richard Rick Klausner, M.D., Chairman. The leadership bench also includes Simon Cooper, M.B.B.S.; Yves Zinggeler, MBA; Christine Chris K. Ward, Ph.D.; Jaya Gautam, Ph.D.; Nitender Goyal, MD, MBA; Tarek Zeidan, Ph.D.; Karen Guarco, MBA; and Patrick Davidshofer. The organization now has capital, clinical milestones, and commercialization preparation converging around the same asset.

LifeMine is digging through biology's chemical inventory instead of starting with an empty sheet of paper. $558M in total funding says investors see enough beneath the surface to keep digging. The more important number comes next, when LIFE-001 enters its 2027 studies and the science has to show exactly what LifeMine found.

Naïve Raises $28.5M Series A for AI Agent Infrastructure

$28.5M says Naïve is betting the next AI infrastructure war won't be won by the agent that talks best. It'll be won by the infrastructure that lets agents actually do business.

The Series A, announced August 5, 2026, was led by Nexus Venture Partners, with Y Combinator, Zetta Venture Partners, Liquid 2 Ventures, and DEEPCORE participating. Add Naïve's previously disclosed $2M seed, and total funding reaches $30.5M. No disclosed valuation. What matters more is what the capital is chasing and why investors are chasing it now.

Sean Dorje, CEO, and Dennis Zax, CTO, dropped out of UC Berkeley and built Naïve around a dangerous question: what infrastructure does an AI agent need when the assignment graduates from writing an email to running the business? Naïve is assembling compute, identity, payments, memory, orchestration, model routing, governance, and access to 10,000+ tools. KYC/KYB, LLC formation, virtual cards, cloud infrastructure, human approval for sensitive actions, and 300+ AI models are part of the stack. Agents aren't getting another chatbot costume. They're getting operating infrastructure.

Developers appear interested. Naïve says 30,000+ developer customers have signed up since its public launch. Sean Dorje says annualized run-rate revenue grew roughly 10x in six months into the low double-digit millions. Use cases include AI automation agencies, faceless TikTok and YouTube channels, and an autonomously operated rental-car business. That last one would have sounded like a fever dream not long ago. Now it's being tested in the wild.

Naïve didn't merely build another agent. It attacked the expensive, fragmented plumbing agents need to become economically useful. The company reports agent startup at 2.3 milliseconds using 1.2MB per instance, plus memory recall requiring 11x fewer tokens per query. When AI economics depend on inference, memory, permissions, and execution cost, infrastructure starts determining what businesses can actually deploy.

The new capital will fund research across serverless agents, inference optimization, persistent memory, and agent-workforce orchestration, alongside engineering hiring. Sean Dorje and Dennis Zax are targeting the hard engineering problems: making agents cheaper to operate, persistent enough to remember context, governed enough for sensitive actions, and capable enough to execute real work.

Nexus Venture Partners and Abhishek Sharma led the round, joined by Y Combinator, Zetta Venture Partners, Liquid 2 Ventures, DEEPCORE, and other operators backing Naïve. Capital gets more interesting when usage, revenue growth, technical ambition, and a painful infrastructure problem arrive at the same table.

For Sean Dorje, Dennis Zax, and the Naïve team, $28.5M provides more fuel to build those rails. Infrastructure for autonomous companies can sound naïve right up until 30,000+ developers start testing the premise. If AI agents become economic actors, who owns the infrastructure underneath them?

Intelligence Raises $7.9M for Design Arena, Led by Index Ventures

Intelligence just raised $7.9M in seed funding for Design Arena, but the bigger story is what Grace Li and Kamryn Ohly figured out before most people finished arguing about AI benchmarks: machines can measure plenty of things, but taste remains stubbornly human. That distinction is becoming valuable.

Grace Li, CEO, and Kamryn Ohly, CTO, were Harvard classmates who studied computer science, with Grace Li also studying neuroscience and Kamryn Ohly studying education. Both worked at Apple. In 2025, weeks before graduation, they started building an AI game-generation engine. The games worked. They just were not particularly appealing. Technically correct and I actually want this live in different neighborhoods. So they built a simple this-or-that system for humans to judge competing outputs. That internal tool became Design Arena.

Today, users submit prompts, receive anonymized results from competing AI models, and vote head-to-head. No logo worship. No brand halo. Just output versus output, with human preference deciding what survives. Design Arena now covers more than 25 categories across websites, apps, games, images, video, audio, slides, 3D design, and more, with models including Claude, Grok, GLM, Gemini, Kimi, and GPT families entering the Arena. The company reports roughly 5.3M-5.5M users across more than 190 countries, up from approximately 40K-50K users during its first weeks in 2025.

Intelligence also reports ARR climbing from $5M to $60M in six months, with a team of 10. Frontier AI labs are paying for private access to the human-preference data generated through these evaluations. The business sits inside the benchmark. Consumer participation creates preference data. Preference data helps labs understand what humans actually favor. That data becomes commercially valuable infrastructure. The crowd gets the Arena. AI companies get signal. Intelligence sits between increasingly capable models and the inconvenient reality that capability alone does not guarantee somebody likes the result.

Index Ventures led the $7.9M seed, with Conviction and A* participating. The company also identifies Y Combinator, whose Summer 2025 batch included Intelligence, as a participant. Other coverage names Valkyrie instead, so the fourth investor remains unconfirmed. The investment points toward a larger opportunity: human judgment is becoming part of the infrastructure required to make increasingly capable AI useful, desirable, and commercially relevant.

For Grace Li, Kamryn Ohly, and the Intelligence team, the founder lesson may be as valuable as the funding news itself. Their first product did not become the company. The mechanism they built to understand why that product missed became the company. Users supplied judgment, labs supplied demand, and an evaluation tool found itself sitting on a valuable stream of preference data.

Harvey Adds Goldman Sachs, J.P. Morgan to Legal AI Bet

Harvey just added Goldman Sachs Alternatives and J.P. Morgan Growth Equity Partners to the cap table, and the most interesting number in the announcement is the one nobody disclosed. No investment amount. No new valuation. Just two institutional heavyweights putting strategic capital behind a company that, four years ago, began with two roommates asking whether GPT-3 could do something useful with legal questions.

Winston Weinberg, CEO, brought the lawyer's view after practicing antitrust and securities litigation. Gabe Pereyra, President, brought the technical muscle from Google Brain and DeepMind. Together, they built a domain-specific AI platform now serving 2,400+ customers across 70+ countries, including 75+ AmLaw 100 firms. Harvey isn't exactly approaching growth like a junior associate waiting to be called into the room.

Before the Jul. 28, 2026, investment from Growth Equity at Goldman Sachs Alternatives and J.P. Morgan Growth Equity Partners, Harvey recorded its first quarter with more than $100M in net-new ARR. Its valuation moved from $3B in Feb. 2025 to $5B in Jun. 2025, $8B in Dec. 2025, and $11B by Mar. 2026. That's a market sending increasingly expensive signals about where professional AI may be headed.

More than $1B in disclosed funding preceded this strategic investment, with existing backing including OpenAI Startup Fund, Sequoia Capital, Kleiner Perkins, Andreessen Horowitz, EQT, and GIC. The investment amount and corresponding valuation remain undisclosed. Goldman Sachs Alternatives and J.P. Morgan Growth Equity Partners arriving now matters because they're backing a business already demonstrating enterprise adoption and serious revenue acceleration.

Under CTO Siva Gurumurthy, Harvey supports contract analysis, due diligence, compliance, and litigation workflows while advancing agentic, multi-step AI. The company acquired Hexus and the Lume AI team in Mar. 2026, followed by Benchmark in Jul. 2026, adding infrastructure for asset managers. Its leadership bench includes Alan Ghelberg, CFO; Katie Burke, COO; Anique Drumright, CPO; John Haddock, Chief Business Officer; Keith Enright, Chief Strategy Officer; and Steve Zad, CRO.

Harvey stacked the evidence before these institutional heavyweights arrived: customer adoption, ARR expansion, enterprise penetration, acquisitions, executive depth, and technology aimed directly at expensive professional workflows. Winston Weinberg, Gabe Pereyra, Siva Gurumurthy, and the Harvey team now have Goldman Sachs Alternatives and J.P. Morgan Growth Equity Partners joining an already formidable investor base. Legal AI once had to prove it could do the work. Harvey's trajectory is pushing the conversation toward what professional-services firms decide AI should be doing inside their own workflows.

Fluxco Raises $26M for AI Transformer Procurement Platform

An 80,000-pound piece of electrical equipment can make a multitrillion-dollar infrastructure boom wait for permission to turn on. That is the transformer problem Fluxco wants to solve, and investors just put $26M behind the effort.

The Austin-based AI procurement platform, founded by Brian Tochman, Benji Miller, and Casey Wu, raised $26M in seed funding as data centers, factories, renewable projects, EV charging networks, and utilities compete for access to power. Congruent Ventures and 8VC co-led the round, joined by Trust Ventures, Koch Disruptive Technologies, Overture Ventures, New System Ventures, White Star Capital, and Cubit Capital.

Brian Tochman identified the constraint through years of investing across energy, including Oklo, Antora Energy, and Base Power. Demand accelerated while U.S. manufacturing capacity moved offshore and transformer lead times stretched from months toward years. Brian Tochman has cited grid-equipment spending increasing from roughly $20B to $60B over four years. Infrastructure can have capital, land, engineering talent, and ambitious timelines, but electrons remain stubbornly unimpressed by PowerPoint.

Fluxco attacks the procurement layer. Its AI-native platform converts plain-English engineering requirements into biddable specifications, connects them with a global OEM network, compares bids across price, lead time, and compliance, and tracks orders through production, testing, logistics, delivery, service, and warranty.

The early numbers explain the investor interest. Fluxco says it booked $30M in GMV within its first 60 days, entirely through overseas suppliers because domestic OEMs could not keep pace. Antora Energy received 34 transformer quotes instead of the typical five, reducing price and timeline by roughly 40%. More competition, more visibility, less waiting. Procurement starts looking less like paperwork and more like infrastructure strategy.

Fluxco also combines procurement with Transformer as a Service leasing, commissioning, servicing, and warranty management while working toward an automated U.S. factory and broader access to power-distribution equipment.

The $26M round puts Brian Tochman, Benji Miller, Casey Wu, and the Fluxco team into a larger conversation about what the AI and electrification buildout actually requires. Congruent Ventures, 8VC, Trust Ventures, Koch Disruptive Technologies, Overture Ventures, New System Ventures, White Star Capital, and Cubit Capital are backing infrastructure several layers beneath the applications collecting most of the attention.

AI may command the headlines, but electricity still gets the final vote. Between the model and the megawatt sits a transformer somebody must specify, source, buy, build, ship, commission, and maintain. Fluxco is making a serious bid to become the place where that entire transaction happens.

Bending Spoons to Acquire Airtable in $1.285B Deal

Bending Spoons just agreed to buy Airtable, and the numbers tell two stories at once. The all-cash deal values Airtable at $1.285B on an enterprise basis, while its net cash pushes the implied equity value to approximately $2.25B.

That is a serious exit by any normal standard. It is also a long walk down from Airtable's $11B pre-money Series F valuation in 2021, the kind that makes a cap table stare at the wall.

The transaction is not closed. Both boards approved it, but regulatory approvals and customary conditions remain. Bending Spoons expects the deal to close later in 2026, and the companies will operate independently until then.

Here is what the valuation headline can miss: Airtable is not a museum piece from the zero-rate era. Annual recurring revenue grew more than 20% year over year to approximately $480M as of June 2026, with more than 500,000 organizations using the platform, including 80% of the Fortune 100.

On those figures, enterprise value works out to roughly 2.7x ARR. That is arithmetic, not a company-reported multiple, but it shows how the market now prices scale, growth, and the work required to turn admired software into durable cash flow.

For Luca Ferrari and Bending Spoons, Airtable fits the thesis: recurring revenue, deep workflow dependence, and room to apply its product, technology, monetization, and AI playbook. Bending Spoons had completed more than 50 acquisitions by its IPO and says it has never sold a material business.

For Howie Liu, Andrew Ofstad, Emmett Nicholas, and the Airtable team, the asset is bigger than a prettier spreadsheet. Airtable sits where structured data, custom apps, automations, interfaces, and AI agents meet. DeepSky and Superagent aim to turn a system of record into one that can reason and act.

That makes the customer question practical. Can Bending Spoons improve Airtable's economics and product speed without damaging the trust of teams that run real operations on it? The announcement promises long-term investment, but pricing, staffing, integration, and roadmap details remain undisclosed.

The market signal is uncomfortable and useful. Strong adoption does not preserve a 2021 valuation forever, and AI ambition does not excuse weak economics. Airtable reached a definitive deal with real revenue, real customers, and a buyer that believes operational discipline can unlock more value.

Bending Spoons and Airtable have reached the agreement. The signature is done; the operating proof comes next. Enterprise customers, software founders, and public-market investors will be watching.

Nielsen’s $2.15B DoubleVerify Deal Reshapes Ad Measurement

Nielsen just put $2.15B behind a simple idea: measuring an audience is only half the job if nobody can verify the media that reached it. The company agreed to acquire DoubleVerify for $13.60 per share in cash, a 30% premium to DoubleVerify's 60-trading-day average as of August 5. Both boards approved the transaction, which is expected to close by Q1 2027, pending shareholder approval, regulatory clearance, and customary conditions.

Nielsen measures who is watching across television, streaming, audio, sports, and digital media. DoubleVerify checks whether digital impressions are real, viewable, brand-suitable, and capable of producing a useful outcome.

Put those capabilities together, and the pitch becomes bigger than measurement. Nielsen wants to connect audience intelligence with media quality, verification, optimization, and campaign outcomes across linear TV, CTV, social, mobile, and AI-enabled advertising workflows.

Automated advertising makes decisions faster, while accountability still arrives through separate vendors and conflicting definitions. A unified independent stack could give advertisers, agencies, publishers, and platforms a cleaner line from audience to environment to outcome.

Karthik Rao, CEO of Nielsen, has pushed the company beyond the old ratings-box caricature. Mark Zagorski, CEO of DoubleVerify, has moved DV from verification into optimization, attention, attribution, and AI-powered measurement. The deal puts those strategies under one roof, assuming regulators and shareholders sign off.

The financial shape is telling. Nielsen says the combined company would generate more than $4B in pro forma revenue and extend solutions to companies responsible for more than $300B in advertising spend. DoubleVerify entered the deal after reporting Q2 revenue of $193.8M, $65.3M in adjusted EBITDA, a 34% margin, and $210M in cash with no debt.

DoubleVerify will become privately held and retain its name and brand. Providence-affiliated funds, which own about 11.8% of its shares, agreed to support the transaction. Financing combines committed debt, incremental equity, and Nielsen cash.

The real question is whether integration can preserve the independence both companies sell while removing the fragmentation customers hate. Combining audience truth with media-quality truth makes sense. Making the systems and client experience feel like one product is the work that starts after the press release.

If Nielsen gets that part right, this deal will not just measure more of advertising. It could make the measurement finally add up.

VisionWave Plans D-Fence Acquisition for AI Defense Push

VisionWave is trying to make the fence part of the brain. The Nasdaq-listed defense technology company signed a term sheet to acquire at least 51% of D-Fence, an Israeli perimeter security business founded by Uriel Bin more than 30 years ago. The proposed deal would be paid in VisionWave stock, with an implied valuation of approximately $5M for the initial stake and a two-year option on the remaining 49% at an implied valuation of approximately $20M.

That is the financial frame. The operating thesis is more interesting. D-Fence builds perimeter intrusion detection systems, pressure sensors, autonomous surveillance tools, counter-drone technology, and command-and-control software for military, airport, border, energy, and other critical-infrastructure environments.

VisionWave is assembling the layers above that moment. Its portfolio spans RF sensing, AI autonomy, computational acceleration, autonomous platforms, and mission control. Adding D-Fence would push the stack outward, from interpreting signals and coordinating machines to detecting what is physically happening at the boundary.

CEO and Executive Chairman Douglas Davis is betting that D-Fence's field experience can strengthen that chain. CTO Dr. Danny Rittman leads VisionWave's technology side, while D-Fence founder and CEO Uriel Bin brings decades of perimeter-security experience.

There is real caution behind the ambition. This is a mostly non-binding term sheet, not a closed acquisition. A definitive agreement, due diligence, VisionWave shareholder approval under Nasdaq rules, regulatory approvals, and customary closing conditions still stand between the announcement and ownership.

The structure also deserves daylight. No cash would go to D-Fence shareholders, but VisionWave may provide up to $1M per year as a repayable loan for contract execution and approved operating expenses. A price-protection mechanism could require additional VisionWave shares after closing, putting potential dilution beside the strategic fit.

D-Fence says its technology has appeared in more than 100 projects across 12 countries and six continents. VisionWave says those claims have not yet been independently verified as diligence continues. That qualification separates a compelling platform story from a proven result.

The parties are targeting a definitive agreement by September 30 and an October 2026 closing. Until then, the signal is clear, but the outcome is not: VisionWave wants to own more of the security stack, and D-Fence could give it the layer where digital intelligence meets the physical world.