The Founder You Were Becoming

OSCP Closes Series A for GNSS-Denied Navigation

Jamming does not have to stop an autonomous vehicle to ruin its mission. It only has to make the vehicle confident about the wrong location.

That is the commercial pressure behind OSCP's undisclosed Series A. New Science Ventures led the round, with Emerging Ventures and 2050 Capital participating. Somu Subramaniam, founder and managing partner of New Science Ventures, joined the board.

Founder and CEO Kazem Zandi is building for the minutes after the satellite fix becomes unreliable. OSCP makes photonic gyroscopes and navigation hardware in Montréal for drones, aircraft, vessels, robots and ground vehicles operating through jamming, spoofing or signal loss.

An IMU can keep measuring motion without GNSS, but tiny errors compound while the outside reference is gone. Customers are buying more trustworthy time before uncertainty swallows the mission.

OSCP's approach uses light moving through a photonic integrated circuit to measure rotation. The company pairs that optical axis with MEMS sensors, aiming to bring the stability of larger optical systems into packages that fit smaller platforms.

The company reports that its photonic MK2E2 finished a 21-minute GNSS-denied road test 5.7 m from ground truth, versus 20.5 m for its MEMS MK2M2 under the same aiding setup. It is a company test, not independent certification, but it shows what drift feels like after the vehicle keeps moving.

The product is also becoming a system. NavigationGate combines an OSCP IMU with on-board sensor fusion, dual-antenna RTK-compatible GNSS and customer aiding inputs. ArduPilot, ROS 2, RS-422 and CAN-FD support address the integration gap between sensor and platform.

Then there is the supply chain. OSCP says its products are made in Canada without US-origin ITAR-controlled content. Canadian export rules still apply, but allied buyers can evaluate the hardware without automatically importing a US re-export dependency into the platform.

The round closed in August and will scale Montréal production, engineering, sales and the next sensor generation. OSCP has also qualified in three streams of Canada's Defence Drone Initiative Marketplace. Qualification is a route to compete, not a contract.

That leaves the Series A with an honest job. The photonics must survive calibration, packaging, yield, environmental testing, customer integration and repeat manufacturing. OSCP now has more room to make resilient navigation procurable, supportable and ordinary enough to trust when the sky stops cooperating.

Deeper Analysis 👇 https://devcuration.com/articles/oscp-closes-series-a-for-gnss-denied-navigation

doxx.net Raises $38M to Build Agentic Defined Networking

The next firewall rule may be written by the same AI agent whose traffic it is supposed to control.

That is a strange amount of authority to hand software through a network designed around human users, public naming systems, centralized services, and infrastructure that quietly records where everything goes. An agent can browse, call APIs, move files, change settings, and act with somebody else's permission. The network underneath it still has to decide what that agent may reach and who gets to watch.

doxx.net raised a $38M Series A to build an answer. Andreessen Horowitz led, with Animo Ventures and focal.vc participating. a16z partner Joel de la Garza is joining the board as founder and CEO Barrett Lyon opens Agentic Defined Networking to public beta.

Barrett Lyon has spent more than 20 years building internet infrastructure. Barrett Lyon founded Prolexic, acquired by Akamai, then built BitGravity, XDN, and Defense.net. That history matters because doxx.net is starting below the settings screen.

The company says it owns the IP space, ASN, DNS root, certificate authority, bare-metal servers, and global backbone spanning 6 continents. People and agents can browse, communicate, transfer files, connect devices, create private names, and configure policy through an API with native MCP support.

Users can begin without an email address or phone number. The company says accounts use proof-of-work, communications travel directly between peers, and DNS-level protection blocks malicious destinations before a person or agent connects.

That architecture asks a useful question: how much personal information should a privacy service need before it can protect you?

doxx.net says it blocked more than 38M potential threats during closed beta since December 2025. That is a company-reported metric, not an audited count of attacks prevented. The round arrived without a disclosed valuation, customer count, revenue, prior-funding total, or detailed use-of-proceeds plan.

The open beta now has to make an ambitious infrastructure thesis feel ordinary enough to use. Consumers and prosumers are first, while Barrett Lyon told Axios that an enterprise offering is planned by year-end.

If agents are going to inherit more authority, the network cannot remain passive plumbing. Identity, routing, policy, communications, and threat protection become part of the agent product itself. doxx.net is financing the layer where those permissions stop being abstract and start deciding what software can actually touch.

Deeper Analysis 👇

https://devcuration.com/articles/doxxnet-raises-38m-for-agentic-networking

Volantis Raises $88M for Photonic AI Inference

Tapa Ghosh and Roy Meade are betting that the next leap in AI inference begins with a component already manufactured by the hundreds of millions for smartphone face recognition.

Volantis Semiconductor has raised an $88M Series A co-led by Lachy Groom and Abstract, with John Doerr, VXI Capital, Triatomic and Susa Ventures participating. Dwarkesh Patel, Naveen Rao and Sholto Douglas joined as angels. The round brings company-reported funding to $97M.

The check is large because the physics problem is larger than a normal chip upgrade. AI inference needs enough memory to hold increasingly large models and enough bandwidth to keep the compute engine fed. Electrical connections only travel so far inside a package, which forces architects to trade memory capacity against speed.

Volantis wants to move that traffic with light.

Its A-1 system uses integrated micro-VCSEL lasers and optical waveguides to connect compute with a much larger memory pool. The company says the design can support models above 20T parameters at up to 10,000 tokens per second per user while fitting into existing data-center racks. It also targets 10 TB of memory and 240 TB/s of bandwidth inside a 15U, 20 kW system.

Those are design targets, not independent benchmarks. The first integrated inference engines are planned for customer delivery in 2027. That gap matters because advanced packaging is where beautiful diagrams meet heat, yield, supply chains and manufacturing reality.

But Volantis did not assemble a team unfamiliar with that collision. Tapa Ghosh is CEO and co-founder. Roy Meade, the CTO and co-founder, previously led Micron's HBM program and served as VP at Ayar Labs. The broader team cites work across NVIDIA, AMD, Broadcom, Marvell, Ayar Labs and Micron, including CoWoS, HBM, co-packaged optics and volume VCSEL production.

The capital will support A-1 development, commercialization and engineering expansion. Volantis currently lists 30 roles across photonics, packaging, architecture, ASIC design, systems and validation.

That hiring plan reveals the actual wager. The AI market has spent years rewarding more compute. Volantis is building around the idea that compute's value increasingly depends on how quickly and economically memory can feed it.

If A-1 reaches customers on the terms Volantis is targeting, the company will have changed which models can be practical at interactive speed and which teams can afford to run them.

Deeper Analysis 👇 https://devcuration.com/articles/volantis-raises-88m-photonic-ai-inference

Charter Space Raises $5M for Space Insurance

A spacecraft company can spend years proving that its hardware works, then discover that an insurer still cannot price what the engineers already understand.

That gap is where Charter Space is building its business.

The El Segundo company has raised an oversubscribed $5M Seed led by Crystal Venture Partners, with QED, Blank Ventures, Gaingels and Hustle Fund participating. Charter says the round brings total funding to $8M after an earlier $3M pre-Seed.

Co-founders Yuk Chi Chan and Yukun Yin started Charter in 2021 with a software problem. Space programs were still managing requirements, manufacturing records, test results and mission decisions across tools that did not give the whole organization one reliable view.

Ubik became that system of record. The more interesting move came when Charter carried the same technical evidence into underwriting.

Insurance has always translated uncertainty into a price. Space makes that translation expensive. A satellite does not arrive with decades of standardized loss data, and novel missions can force an underwriter to understand hardware, launch, operations and liability before anybody can discuss terms.

Charter launched the nationally licensed Charter Interplanetary Risk Corporation in 2026 to connect those worlds. Its brokerage offers commercial coverage for space and defense companies alongside specialty protection from pre-launch transport and testing through launch, orbital transfer and on-orbit operations.

Charter reports serving more than 50 companies across the U.S. space and defense industrial base and more than $35M in gross written premium backlog. Those are company-reported figures, not audited revenue.

Crystal Venture Partners invests where technology changes how risk gets understood and transferred. Charter is trying to make the technical record behind a mission useful before a loss, while the operator is still looking for coverage and capital.

The financing will expand distribution, sales and products for in-space servicing, space-based nuclear power and lunar resource utilization. No valuation or round terms were disclosed.

When a spacecraft becomes easier to insure, lenders and other capital providers have a better-defined risk to evaluate. Equity stops carrying every dollar of the industry's expansion alone.

Charter now has to turn technical fluency into repeatable underwriting evidence, then turn that evidence into policies that work when the mission leaves the slide deck and meets physics.

Deeper Analysis 👇 https://devcuration.com/articles/charter-space-raises-5m-space-insurance

Notey Raises $3M to Turn Music Practice Into Play

The money behind Notey Inc is doing 2 jobs at once: building a better reason for a child to pick up an instrument tomorrow, and building faster routes into the homes and schools where that decision gets made.

Notey has closed a $3M Seed package led by Tivoli Partners and co-led by Malta Venture Capital. The total includes a $1M non-dilutive development grant from Malta Enterprise, an important distinction in a funding market that often throws equity and public support into one shiny headline.

The product starts with a demanding idea. A real guitar becomes the controller. Notey listens through the device microphone and turns practice into progress through songs, worlds, mini-games and immediate feedback. A child advances by playing the instrument, not by tapping around a lesson while the guitar gathers dust.

Mak Grgic, CEO and co-founder, brings the performer's view of why practice loses people. Adithya Bellathur, CTO and co-founder, brings the engineering needed to make live audio drive the game. Chief Strategy Officer Yishai Pinchover adds experience building game-based education into a distribution business.

The company reports more than 100,000 consumer users and activity in 60 schools across 4 districts. Those are company figures, but they explain why this round is pointed beyond another feature release.

Notey has opened a Maltese subsidiary as its European product, engineering and distribution base. Telemach is becoming a channel into Slovenia and the wider Balkans. EdMusica is part of the Latin American expansion plan. WorldBuilder is preparing for public launch, while piano, voice and recorder support is targeted for Q1 2027.

Consumer education products can spend heavily to win a download, then discover that attention, trust and distribution are 3 different businesses. Notey is trying to connect them: make practice engaging enough to repeat, give teachers visibility and use partners that already know the families.

The Malta base can shorten the distance to European schools and partners. Telemach can put the product in front of households without Notey acquiring each family one at a time. New instruments can widen the practice room, but only if the game keeps the instrument at the center.

The useful question is no longer whether music education can be gamified. Plenty of apps can add points and badges. Notey's harder assignment is proving that gameplay can keep a child practicing long enough for parents, teachers and distribution partners to hear the difference.

Deeper Analysis 👇 https://devcuration.com/articles/notey-raises-3m-to-turn-music-practice-into-play

Hop Aero Raises $11M for Hypersonic Rocket Cargo System

A 15-minute delivery promise gets very quiet when the cargo reaches reentry.

Engine fire and payload deployment are useful milestones. They do not help the person waiting in a contested environment unless the vehicle can survive reentry, find unprepared ground, and deliver where the mission needs it.

That is the trust problem Hop Aero is financing with $11M.

Co-founder and CEO Matija Milenovic and co-founder and CTO Jacob Balaj are building Rook, an autonomous cargo rocket designed to carry 250 kg up to 750 km in about 15 minutes. The concept launches from a standard 40-foot shipping container, follows a suborbital path, and is intended to land vertically without a runway.

Initialized Capital led the financing. Y Combinator, Lunar Ventures, System.One, and Outsized Ventures participated, with Paul Graham and Jason Miller joining as strategic angels. Payload says the total includes capital raised before and during YC, all through SAFEs. This is seed financing, not a priced $11M equity round.

Hop Aero says it designed and 3D-metal-printed an engine with its hopOS tooling, completed a hot-fire, ran a static fire of a roughly half-scale Rook, tested a small tethered prototype, and demonstrated 4 drones sliding from the payload bay.

The U.S. Air Force also awarded Hop Aero a $1.25M Phase II contract for an engineering design unit integrating propulsion, avionics, structures, and autonomous landing logic. The military has spent years studying how rockets might move cargo when roads, runways, and ordinary delivery timelines stop cooperating.

The remaining work is less cinematic and more important. Hop Aero has not flown untethered or demonstrated the hypersonic reentry and precision landing sequence. Full-scale Rook 2 will use a different engine and propulsion architecture, with its first flight targeted for mid-2027.

That is where the round earns its meaning. Matija Milenovic and Jacob Balaj are not being funded to repeat a clean subsystem test. They are being funded to make several unforgiving systems arrive at the same answer, at the same time, after a flight profile that leaves little room for a polite recovery plan.

If Rook works, the customer is buying more than speed. The customer is buying the confidence to make a logistics decision later, launch closer to the moment of need, and reach a place conventional infrastructure cannot serve. Hop Aero now has to turn that confidence from a rendering into an operating habit.

Deeper Analysis 👇 https://devcuration.com/articles/hop-aero-raises-11m-hypersonic-rocket-cargo

ArkeaBio Raises $7M for Methane Vaccine Field Trials

The cattle industry already knows how to move a needle through a herd. ArkeaBio is trying to make that ordinary farm routine carry a climate result.

The Charlestown agricultural-bioscience company closed a $7M Series A1 co-led by Breakthrough Energy and AgriZeroNZ. The round supports R&D and farmer pilots as ArkeaBio moves its methane-reducing cattle vaccine from animal studies toward broader field programs.

Feed additives can reduce methane, but they have to reach the animal through a controlled feeding system, often every day. That works differently when cattle spend their lives on pasture. A vaccine could travel through animal-health manufacturing, veterinary distribution and herd-care routines that farmers already use.

Founder and current CBO Colin South built ArkeaBio around that operating advantage. The company says its vaccine stimulates antibodies that reach the rumen and interfere with methane-producing microbes. If the biology holds up, the delivery model could reach cattle that sit outside the daily-feed infrastructure.

The word "if" is carrying real weight. ArkeaBio has reported early methane reductions in the 10-15% range. An independent Global Methane Hub review cited a reported 12.9% reduction in 10 cattle over 105 days, then made the important point: the experimental conditions and results had not been published in a scientific journal. Larger, longer trials still have to establish efficacy, duration and reproducibility.

The leadership change matters in that context. Frank Wooten became CEO with the financing after co-founding and leading livestock virtual-fencing company Vence through its acquisition by Merck Animal Health. CTO Zach Serber brings industrial-biology experience from co-founding Zymergen and working at Amyris. ArkeaBio now has a scientific program and a leadership group shaped for the less glamorous work of trials, regulators, partners and adoption.

There is movement beyond the lab. Emissions Reduction Alberta lists an active $4.9M demonstration in Kinsella, supported by $2.35M in public funding, to evaluate a novel inoculant in beef cattle. Active is the honest word. The product is not publicly established as approved, commercial or proven at herd scale.

That is the job attached to this $7M. ArkeaBio must turn a familiar farm action into repeatable climate evidence without asking pasture producers to rebuild the way cattle are managed. The needle is ordinary. The proof has to survive the field.

Deeper Analysis 👇

https://devcuration.com/articles/arkeabio-raises-7m-methane-vaccine-field-trials

Walapay Raises $4.6M for Global Payments Infrastructure

The payment that looks finished in New York can still spend hours becoming usable money in Nairobi, Manila or São Paulo.

The screen says sent. Local licensing, bank access, FX, liquidity, compliance and the recipient's account still have work to do. Cross-border payments often celebrate the middle of the journey.

Walapay has raised a $4.6M Seed round led by Generative Ventures to take responsibility for more of the part that comes after.

Brothers Tom Borgers and Dimitri Borgers built Walapay around a decision: stablecoins are one rail, not the entire product. The team connects them with bank accounts, local rails, FX, collections and payouts through one API.

That distinction matters. A stablecoin can move value quickly across a network. A business still needs the money to arrive in the right currency, under the right controls, through infrastructure its bank, regulator and finance team can live with.

Walapay says a cross-border payment may touch 4 or 5 banks and payment providers before settlement. Every handoff can add fees, delay and another place for the transaction to become somebody else's support ticket.

The company's answer is to own more of the licensing, banking and liquidity relationships behind its routes. Its website describes 60+ currencies and 180+ countries, with accounts, collections, FX, payouts and stablecoin orchestration in one stack.

The financing also includes Commerce Ventures, Polygon, Verda Ventures, NGC Ventures, FGV Capital, AAF, Jsquare, Knollwood, Big Brain Holdings and others.

Walapay plans to expand its licensing footprint, deepen banking partnerships and hire. None of that makes a flashy demo. All of it determines whether the product survives a real corridor, a compliance review and a customer who expects the money today.

The company reports $2.5B in annualized payment volume and names Kast, Nuvei and Bastion as customers. Tom Borgers says the infrastructure serves more than 100,000 customers. Those figures are company-reported, not audited proof, but the operating responsibility is already real.

Walapay's next chapter will be written market by market. Every new corridor adds utility, but it also adds another set of banking relationships, local rules, liquidity demands and failure modes the team has chosen to own.

That is where the $4.6M goes to work: inside the licenses, partnerships and operational muscle that turn global settlement into local money people can actually use.

Deeper Analysis 👇 https://devcuration.com/articles/walapay-raises-4-6m-global-payments-infrastructure

SedMed Raises $4M for Non-Electric Toilet Lift

Bathroom mobility is where dignity, patient safety, and worker safety collide in a room nobody wants to discuss.

SedMed began with a stroke and a daily task that suddenly became difficult. Co-founders Jeremy Bronen and Timothy Krupski turned that problem into a non-electric toilet lift that attaches to most standard toilets and helps users sit and stand without batteries, motors, or bathroom remodeling.

Now 108 BioCapital is leading a $4M financing for the Milford, Connecticut medical-device company.

The amount matters. The handoff matters more.

SedMed says its lift is already used across more than 40 hospital systems. That is company- and investor-reported adoption, not independent proof of clinical outcomes, but it does show the business has moved beyond a university design project and into institutions that understand the cost of a difficult transfer.

A bathroom transfer can ask 2 people to solve 1 problem at the same time. The user wants independence and privacy. The caregiver wants to avoid becoming the lifting mechanism. OSHA treats toilet-to-chair transfers as a high-risk patient-handling task because repeated manual lifting can injure the staff member helping.

SedMed's answer is intentionally mechanical. An adjustable gas spring can provide up to 80% assistance across 5 resistance settings, with no electricity to install, charge, or lose during an outage. That simplicity is the product strategy: fewer dependencies between the person who needs help and the moment help is needed.

The new capital is expected to expand commercial reach, grow distribution and channel partnerships, advance the IP portfolio, and fund additional hospital-based studies. That last item deserves attention. Adoption can open doors. Evidence determines whether health systems keep opening them, standardize the device, and explain the economics to procurement teams.

108 BioCapital is also looking beyond U.S. hospitals. The investor pointed to home use, India, and other international markets where a non-electric design may travel more easily than equipment built around power, renovation, or complex installation.

The market is getting older. The operating question is getting more personal: how much assistance can be added without taking independence away?

SedMed is building around that narrow, unglamorous, unavoidable moment. If the studies, distribution work, and home adoption move together, the toilet lift becomes a smaller handoff between a person and a caregiver at exactly the point where both need more control.

Deeper Analysis 👇 https://devcuration.com/articles/sedmed-raises-4m-in-funding

Parakeet Health Raises $10M to Scale Patient-Access AI

A missed call at a medical practice can become more than voicemail. It can become an unfilled appointment, a patient who gives up, and clinical capacity that never reaches the person waiting for care.

Parakeet Health has raised an oversubscribed $10M Series A to work inside that gap. Canvas Ventures led the round, with Blank Space Ventures, StoryHouse Ventures, and HMC INQ participating. Total funding now stands at $13M.

Healthcare already has software for one channel or one task. The operating mess begins when a patient moves between calls, texts, email, web scheduling, and the fax machine healthcare refuses to bury.

Parakeet is building one patient-access platform across those handoffs. Its agents manage calls, outreach, fax processing, scheduling, referrals, cancellations, and follow-up while integrating with existing EHR and practice-management systems. A performance model ties Parakeet's economics to appointments booked, interactions resolved, and revenue generated.

That model reflects the people behind it. Co-founder and CEO Jung Park spent more than 25 years in healthcare operations, with leadership work at One Medical, Epocrates, and California Skin Institute. Co-founder and CTO Aaron Lee and co-founder and CPO Eric Mao bring the engineering and product side. Enterprise healthcare AI is rarely defeated by the demo. It gets defeated by the scheduling rule, insurance nuance, or workflow exception the demo never met.

Parakeet says ARR grew 10x over the past year. It now works with six of the 10 largest U.S. dermatology groups, supporting more than 2,800 providers across 1,100+ locations. Those are company-reported figures, but the footprint is becoming visible: its Qualderm Partners collaboration spans nearly 160 practices across 17 states.

The capital will expand the technology and team, deepen personalization, and move the platform into more specialty and outpatient markets.

Dermatology gave Parakeet a dense collection of scheduling rules, repeat visits, specialty workflows, and measurable economics. The next stage asks whether that operating discipline can travel across specialties without losing the customization that made it valuable.

Healthcare organizations do not need another cheerful robot at the front door. They need patient demand, staff capacity, scheduling logic, and follow-up to survive the trip through the building. Parakeet now has $10M more to make that trip feel like one system instead of five departments passing the same problem around.

Deeper Analysis 👇 https://devcuration.com/articles/parakeet-health-raises-10m-in-series-a-funding

NexBank Capital Raises $150M in Non-Voting Common Equity

A private bank can look flush with cash and still run into the harder limit: how much common equity stands behind the next dollar of growth.

NexBank Capital has raised $150M through the sale of non-voting common stock. It closed September 29 and announced October 1, bringing aggregate equity offerings since 2021 to $540M.

That security type matters. Deposits and debt help fund a bank, but common equity absorbs losses before creditors do. It gives management room to add assets and take calculated risk without pretending every dollar on the balance sheet carries the same obligation.

NexBank entered this raise with $17.872B in assets, $10.844B in deposits, $9.3B in net loans, $7.419B in cash and securities, and $1.102B in stockholders' equity at June 30, 2026. Its bank subsidiary reported a 9.07% Tier 1 leverage ratio, a 14.92% Tier 1 risk-based capital ratio, and a 15.39% total risk-based capital ratio.

The new capital does not arrive with a detailed shopping list. NexBank Capital says it will use the proceeds as growth capital and for general corporate purposes, giving the company flexibility to be opportunistic. It did not name the investors, disclose pricing or valuation, or identify the loans, acquisitions, securities, technology, or business lines that will receive the money.

That makes the raise less about a promised destination and more about who now holds the map.

Matt Siekielski, President and CEO, oversees the strategy and operating performance of NexBank Capital and NexBank. Stacy Hodges, EVP and CFO, oversees financial management and corporate strategic planning. Their platform spans institutional, commercial, mortgage and personal banking nationwide.

The capital history is also worth reading correctly. NexBank Capital disclosed $390M of equity placements through the end of 2022, including preferred, voting common, and non-voting common stock. This new $150M is the latest component of the $540M total, not a $540M round wearing a smaller headline.

Cantor Fitzgerald advised on $75M of the raise, and Hunton Andrews Kurth served as company counsel. Neither role identifies an investor.

The useful tell will be where the balance sheet moves next. More loans, a different asset mix, an acquisition, or another line of business will reveal what management considered worth backing with permanent capital.

NexBank Capital bought flexibility. The next set of quarterly numbers will begin showing which choice earned it.

Deeper Analysis 👇 https://devcuration.com/articles/nexbank-capital-raises-150m-common-equity

Photon Raises $4.5M for Messaging-Based AI Agents

The hardest part of building a consumer AI agent may be persuading anyone to install it.

Daniel Tian and Ryan Zhu learned that after building consumer apps as students. The code could work. The launch could happen. Then came the old distribution tax: discovery, download, signup, a new interface, and the hope that a user remembered to return.

Photon takes a different route. Put the agent inside iMessage, WhatsApp, and other channels where the conversation is already happening.

That thesis now has a $4.5M Seed round behind it. Gradient and A* co-led the financing, with Vercel, HongShan, Z Fellows, Llama Ventures, Karman, and angel investors participating. Tian is Photon's co-founder and CEO. Zhu is co-founder and CTO.

Their product, Spectrum, is an open-source SDK and hosted platform for messaging-based agents. Developers get a unified API, channel framework, CLI, and observability instead of rebuilding transport for every product.

The behavior is familiar before the agent says a word. Users know how to open a thread, reply, react, share media, and follow a notification. That can turn distribution from a storefront fight into a contact-level relationship, but it also raises the standard. A bad app waits behind an icon. A bad messaging agent interrupts your day in the same place your family, friends, and work already live.

Photon says more than 40,000 developers had signed up when TechCrunch reported the round, while Tian's October 2 update put the figure above 50,000 with 450,000 monthly npm downloads. The company also reported 10x revenue growth in four months, less than 3% churn, and 5x messaging-volume growth in one month. Those are company-reported metrics, not audited results, but they explain the investor interest.

The hosted platform adds the less photogenic work: 99.95% uptime, SOC 2 Type II compliance, HIPAA-compliant infrastructure, routing, and failover. That machinery matters when an agent moves from a demo into a workflow where a missed message becomes an operating problem.

Photon even held a funeral for mobile apps in a San Francisco church. The coffin made good theater. The business is being built in the handoff between an agent's intelligence and a channel people trust enough to keep open all day.

The $4.5M gives Photon room to develop that layer. Developers will decide how much of the app economy fits naturally inside a conversation, one production deployment and one very personal notification at a time.

Deeper Analysis 👇 https://devcuration.com/articles/photon-raises-4-5m-seed-messaging-ai-agents

Metaview Raises $60M Series C for Agentic Recruiting

A recruiting team can add AI to every stage of hiring and still lose the one thing each stage needs: memory.

The hiring manager explains the role in an intake call. A sourcer turns that conversation into a search. A screening tool evaluates applications. Interviewers add evidence. Between those handoffs, the definition of a good hire can become disconnected prompts, notes and scorecards.

Metaview has raised a $60M Series C to make that context travel.

Insight Partners led the round, with GV, Intrepid Growth Partners, Seedcamp, Vertex Ventures US, Plural and Garuda Ventures participating. The financing brings Metaview's company-reported total to $110M.

Siadhal Magos and Shahriar Tajbakhsh founded Metaview in 2018 after building teams at Uber and Palantir. It grew from hiring conversations into agents for sourcing, outreach, application review, screening, notes and reporting.

That expansion is landing in a recruiting market where volume stopped being a polite problem. Greenhouse found annual applications per recruiter rose 412% from 2022 through 2025 while average recruiter headcount per organization fell 56%. Candidates can apply faster than teams can create attention, leaving strong applicants waiting inside the queue.

Metaview is betting the answer is not another assistant that makes one task quicker. The platform needs to remember the role, carry feedback forward and give each agent the same operating context. People still set the criteria and make the final decision.

The company reports 7,000+ customers and 6M+ interviews captured. Insight says some customers cut time to hire by more than 75%. Those reported outcomes show the scale of the workflow Metaview wants to own.

The capital will bring fillmore, its autonomous outbound recruiting coworker, into general availability, fund new specialist agents, grow the team from 80 to 250 by the end of 2027 and open New York alongside London and San Francisco. It will also support 10x Recruiting, a program for people learning to run AI-powered hiring systems.

There is a serious handoff inside that growth plan. The more work agents perform, the more clearly the platform must show what they saw, why they acted and where human judgment takes over. Hiring consequences do not disappear inside a speed metric.

Metaview now has the capital to make shared context an operating layer for recruiting. The next chapter will be written inside the decisions its agents accelerate and the judgment its customers refuse to outsource.

Deeper Analysis 👇 https://devcuration.com/articles/metaview-raises-60m-series-c-agentic-recruiting

Micro1 Reportedly Raises $100M-Plus at $4B Valuation

A model can read every contract it can find and still need a lawyer to explain why one clause creates a problem on Friday afternoon.

That gap between information and judgment has become a very expensive piece of AI infrastructure. Forbes reports that micro1 raised more than $100M at a $4B valuation, citing 2 people familiar with the deal. The company declined to comment, so the round, valuation and participants remain reported rather than company-confirmed.

The speed is hard to ignore. Ali Ansari started micro1 as an AI recruiting business. A data customer used it to recruit engineers for annotation work, exposing the larger opportunity: frontier models needed doctors, lawyers, engineers and other experts to create training data, evaluate outputs and teach systems how work gets judged.

micro1 turned that recruiting machinery into a supply chain for human judgment. Zara vets experts while the data operation supports model training and evaluation. The company has expanded into reinforcement-learning environments, enterprise operational data and robotics data, where useful material hides in decisions and workflows that never reached the public internet.

The reported economics explain why investors care. micro1 announced a $35M Series A at a $500M valuation in September 2025, led by 01 Advisors with Antler Elevate participating. Forbes now says the valuation climbed 8x in roughly a year. TechCrunch reported a $500M gross annual run rate in August and estimated $150M to $200M net after expert payments.

That distinction matters. Gross volume shows how much work is moving through the system. Retained revenue shows what kind of business is being built underneath it.

Forbes says 2 unnamed frontier AI labs and 2 unnamed xAI cofounders joined the new round. Customers putting capital behind a supplier is a useful signal: human data is moving closer to the strategic core of model development, especially as labs compete on evaluations, agent performance, robotics and access to real enterprise workflows.

The open questions are now commercial ones. micro1 has not disclosed the round type, lead investor, full syndicate, primary-versus-secondary split or use of proceeds. The company still has to prove that extraordinary data demand can become durable margins without letting quality, rights or trust decay as the operation scales.

The models keep learning. micro1 is building the market that decides whose judgment reaches them, how it is tested and what that judgment is worth.

Deeper Analysis 👇 https://devcuration.com/articles/micro1-reportedly-raises-100m-plus-at-4b-valuation

Augmeta Raises $3M for Agentic KPI Operations

The dashboard did its job days ago. The team is still deciding who owns the number.

A campaign paused during an email migration. Checkout slowed after a release. Conversion slipped while the evidence scattered across analytics, logs, product calendars, and six other systems. By the time the weekly review asks what happened, the expensive part is no longer detection. It is the handoff.

Augmeta has raised a $3M Seed round led by Depth Ventures, with Next Wave NYC, investments through the Index Ventures and Sequoia Capital scout programs, and angels participating. Salman Jamali, CEO and co-founder, is building the company with co-founders Nitin Bhaskaran and Arslan Jamali.

Their premise is operational: give every KPI an AI Operator that stays with the metric from change to explanation to action. The software monitors continuously, investigates the systems around a movement, estimates the dollar impact, brings evidence to the right team, and checks whether the response worked. Customers decide what the system may change, so autonomy follows permission.

Dashboards show that a number moved. Enterprises still depend on people to notice the signal, find the context, negotiate ownership, and return later to see whether the fix held. Augmeta is trying to make that loop persistent while people keep control of consequential decisions.

At Tractor Supply, Augmeta says its operators work across more than 100 e-commerce KPIs and produce material used in business reviews. The retailer renewed after its first year. It is company-reported traction inside a demanding environment: many metrics, many systems, long deployments, and consequences measured in revenue rather than demo applause.

Salman Jamali previously led engineering for Opendoor's home operations. Nitin Bhaskaran led product teams at Home Depot and Opendoor. Arslan Jamali worked as a senior product manager at Amazon Business and in finance at AWS and Prime Video.

Depth Ventures' Paul Jun and Jeff Arnold backed a team that can speak engineering, product, and money in the same meeting. The round will support faster deployments, broader KPI coverage, and more capable operators. Augmeta sells annual enterprise contracts today and has discussed value-based pricing over time, which would make attribution part of the business model.

The capital buys deployment capacity and product depth. The opportunity depends on whether enterprises trust Augmeta to carry context across systems, respect permissions, and stay accountable after the number moves.

Deeper Analysis 👇 https://devcuration.com/articles/augmeta-raises-3m-for-agentic-kpi-operations

FieldAI Reportedly Raising $700M at $10B Valuation

A construction site can invalidate a robot's map before lunch. A wall moves, equipment arrives, crews reroute, and the autonomy that looked fluent in a demo discovers that concrete has opinions.

That operating mess is the market FieldAI chose. Business Insider's Rya Jetha reports the Irvine physical-AI company has signed a term sheet to raise $700M at a $10B valuation, 5 times the $2B post-money valuation reported after its 2025 financing.

This is a reported term sheet, not a formally closed round. The lead investor and participants have not been named, and FieldAI did not comment to Business Insider. The size still says plenty about the wager.

Founder and CEO Ali Agha and President and CSO Shayegan Omidshafiei are building an intelligence layer across robot types. FieldAI says its Field Foundation Models reason about physics, uncertainty, and risk without prior maps, GPS, predefined routes, or constant cloud access.

That is a different business from building one impressive robot. Hardware manufacturers can keep changing the body. FieldAI wants its software to become the brain customers reuse across machines, sites, and jobs.

The company has put that idea into the field with Boston Dynamics, NVIDIA, and industrial customers. Business Insider reports more than $135M in revenue and customer contracts across 30-plus customers. Those figures come from a person familiar with the company, not audited public financials, but they help explain why prospective investors may see more than a laboratory story.

Physical AI has a nasty commercial loop. Customers want proof that robots can handle the real world before deploying them broadly. The models improve by collecting more data from real deployments. FieldAI argues that risk-aware autonomy can earn enough trust to put more machines into service, creating more data, stronger models, and a wider operating footprint.

The reported $10B valuation prices that loop as a platform, not a feature.

FieldAI raised $405M across 2 rounds in 2025 from investors including Bezos Expeditions, Canaan Partners, Khosla Ventures, NVentures, Prysm, and Temasek. None has been identified as part of this reported financing. If the new round closes, the unanswered questions move toward repeatability: contract conversion, deployment economics, margins, and whether one autonomy layer can keep learning across industries without becoming a pile of expensive exceptions.

The robot gets the photograph. The business gets judged on what happens the next morning, when the site has changed again and the customer still expects the system to know what to do.

Pulse Fund Closes $63M Inaugural Climate Venture Fund

Venture categories are tidy because databases need columns. The physical economy never agreed to stay inside them.

A battery breakthrough can change electric mobility. Agricultural efficiency can redirect infrastructure spending. Grid capacity can decide whether a factory, fleet, or climate technology gets built or waits for a connection that exists mainly in a planning document.

That is the territory Tenzin Seldon built Pulse Fund to underwrite. The Los Angeles firm has closed its inaugural venture fund with $63M in commitments for early-stage climate companies across the energy transition, infrastructure, food and agriculture, and mobility.

Backers include Beneficient, C6 Partners, Pivotal Foundation, members of the Pritzker family, Veronica Chou, David Osborn, and an unnamed Asia-based sovereign wealth fund. They are backing a solo managing partner's wager that valuable climate companies will be found where one market changes another's economics.

The portfolio already makes that argument in operating form. Endera, Floodbase, InventWood, and it's electric span commercial vehicles, flood intelligence, engineered wood, and curbside charging. Mast Reforestation, Plantible, Twelve, and Unravel Carbon carry the thesis into forests, food, carbon transformation, and software.

Pulse says it frequently leads rounds and takes board positions. That places the firm inside the governance and commercialization choices that decide whether a climate technology crosses from pilot to operating system.

That range can be an edge, but every broad thesis meets a burden. These markets run on different buyers, regulations, capital cycles, and technical risks. Seeing a connection is useful. Pricing both sides deeply enough is where the fund earns its name.

Pulse added a science layer through a 2024 Space Act Agreement with NASA's Goddard Space Flight Center. It connects companies with Earth-observation data, modeling, and climate technology while Pulse develops impact-measurement methods.

That will not erase commercial risk. It can improve the evidence when founders and investors decide whether a climate claim, physical constraint, and business model belong in the same investment case.

The $63M close gives Seldon a vehicle for following those connections into early-stage companies. Founders between categories now have an investor designed to understand the handoff. Pulse now has to show that systems thinking can survive selection, pricing, governance, follow-on capital, and time.

Deeper Analysis 👇 https://devcuration.com/articles/pulse-fund-closes-63m-inaugural-climate-venture-fund