The Room Has to Be Ready

Mistral

Independence gets expensive the moment AI leaves a policy speech and enters a data center. Models need chips, power, training runs, inference capacity, service levels, and enterprise teams willing to trust the stack with work that actually matters.

Mistral has raised €3B, about $3.5B, in a Series D at a post-money valuation above €21B, roughly $24B. Samsung Electronics led the round, with the EQT Group-managed Scaleup Europe Fund and existing investor PSG as co-leads.

Arthur Mensch, Guillaume Lample, and Timothée Lacroix launched Mistral in 2023. Arthur Mensch is CEO, Guillaume Lample is Chief Science Officer, and Timothée Lacroix is CTO. In three years, their research lab has become a larger operating proposition: open-weight models, enterprise software, custom-model training, regional inference, and the compute underneath it.

Mistral says the capital will deepen frontier research, increase training capacity, expand infrastructure, and support commercial growth across a business that now operates in 20 countries and serves more than 125 global enterprises, including Airbus, ASML, and HSBC. Those operating figures are company-reported, but the direction is plain.

The investor lineup fits that industrial ambition. Samsung now leads the Series D after ASML led Mistral's €1.7B Series C in 2025. The Scaleup Europe Fund brings an EU-backed growth-capital instrument into the syndicate, while PSG, Advent, BlackRock-managed accounts, the Grand Duchy of Luxembourg, and existing shareholders add financial reach.

Mistral also raised $830M of debt in March to finance Nvidia infrastructure for a Paris-area data center. The €3B announced today is equity. The transactions show a company financing research and physical capacity through different pools of capital because the stack demands both.

The check buys room rather than inevitability. Mistral still has to prove that open weights, regional control, infrastructure ownership, and enterprise deployment can behave like one dependable business. The founders now have to coordinate science, construction, reliability, and sales without letting one layer turn the others into an expensive promise.

Europe has spent years debating whether it can own more of the AI value chain. Mistral now has Samsung, ASML, EQT, PSG, and public European capital helping pay for the answer. The next part will be written in capacity delivered, models improved, and enterprise workloads that stay after the symbolism wears off.

DivergeIT

The spreadsheet behind an MSP roll-up counts revenue, technicians, locations, contracts, and tools. The column for customer trust is usually blank.

DivergeIT has spent more than 27 years asking businesses to hand over systems they cannot afford to lose. Now M|C Partners is making an undisclosed strategic investment and turning that history into a platform for organic growth and future MSP acquisitions.

CEO Jarrod Koch will keep leading DivergeIT and work with M|C Partners to identify and integrate targets. Founding Partner Jim Lambert remains part of the company story behind that mandate. No amount, valuation, ownership stake, or fund vehicle was disclosed.

That missing number is less useful than the disclosed job.

DivergeIT plans to invest in talent, technology, AI capabilities, and complementary MSP acquisitions. It already spans managed IT, security, compliance, cloud, consulting, and AI services across more than 25 states.

M|C Partners is not learning the MSP roll-up on this deal. The firm helped build Thrive by combining regional providers and ultimately rolling up 9 MSPs around a shared platform. Principal Arvind Viswanathan and Managing Partner Gillis Cashman are backing a familiar thesis: technology gets harder to manage, customers outsource more of it, and the market remains fragmented.

The elegant version of that thesis lives in a model. The difficult version arrives on Monday morning after an acquisition.

Two service desks have different escalation habits. Engineers use different tools. Security controls carry different exceptions. The buyer inherits every undocumented promise made across years of customer relationships.

That is why DivergeIT's company-reported 98% CSAT and 96% client retention matter. They do not prove future integration performance, but they identify the asset M|C Partners wants to scale: clients believing someone still owns the outcome when infrastructure gets messy.

AI raises the stakes. DivergeIT is expanding from managed infrastructure and security into AI discovery, governance, custom agents, and ongoing support. Acquisitions can add specialists and reach, but every new tool and team expands the surface that has to be secured, documented, and made coherent for the customer.

The platform will be judged in those handoffs. If DivergeIT can add MSPs without making clients feel the seams, it can make operational trust portable across the expansion without hollowing out the service.

ChipMango

The semiconductor industry can announce a new fab in one afternoon. Training the engineers who make its designs usable takes longer, and most of that work happens far from the ribbon cutting.

ChipMango is building in that gap. The company closed a $1.9M funding round led by Atlantica Ventures VC Fund, with DFS, Kaleo Ventures, Madica, Trilinear Technologies, Malta Ventures, and other strategic investors participating.

Founder and CEO Ola Fadiran and co-founder and COO Jovan Andjelich paired 2 businesses that usually live apart. One gives learners industry-aligned curricula, professional design tools, and practical work in Arm technologies. The other delivers semiconductor design, verification, edge-AI, and intelligent-hardware services to customers.

The connection matters more than the course catalog. Chip design is learned through tools, judgment, and repeated exposure to work that can fail in expensive ways. A classroom can teach the method. A verification program teaches what happens when a schedule and product depend on the answer.

ChipMango says its engineers support Trilinear Technologies on verification for DisplayPort and multimedia IP used in SoCs, FPGAs, and ASICs. Arm separately lists ChipMango as an Approved Training Partner across North America, EMEA, and Africa. The relationships turn a talent thesis into operating work.

The capital will expand engineering and product capacity, support new design engagements, scale the learning platform, and advance edge-AI and intelligent-sensor work. ChipMango also plans to grow across the US, Africa, and Europe, including a European Design Center in Malta.

Deloitte estimates the industry needs more than 1M additional skilled workers by 2030. That figure is usually presented as a hiring problem. ChipMango is treating it as a chain of access, instruction, professional tools, commercial practice, and customers that have to move together.

ChipMango has not disclosed valuation, revenue, learner placement, project retention, or how quickly trained engineers move into paid delivery. Those measurements will show whether education and services reinforce each other at scale or simply compete for the same management attention.

That handoff is where the $1.9M matters. More engineers can enter the pipeline, more customer work can test what they know, and evidence from each engagement can return to the next cohort. ChipMango is financing that loop while the semiconductor industry is still counting the people it does not have.

Ollie

A family assistant has to earn access in reverse. It needs the school calendar, the inbox, the group chat, and eventually the payment flow before it can save meaningful time. Every new permission makes the product more useful and the trust decision heavier.

Ollie AI raised a $7.5M Seed round led by Khosla Ventures, with AI House participating, to build inside that tension. The San Diego company works through text, connecting calendars and email, joining shared conversations, and coordinating the small handoffs that turn a household into a part-time operations team.

Bill Lennon, Co-Founder & CEO, Christy Shannon, Co-Founder & CMO, and Rushabh Doshi, Co-Founder & CPTO, are selling relief from invisible work. The product only earns that promise if it can remember enough to help without making the family regret what it shared.

That is why the less glamorous details carry weight. Ollie says its subscription model keeps the household, rather than advertisers, as the customer. It says it does not sell user data or use customer content to train public LLMs. Its recent SOC 2 compliance adds audited controls to a category that often asks for trust before earning much operating history.

Ollie's login design makes the trade-off visible. When the assistant needs to enter a website or complete a purchase, TechCrunch reports that it opens a cloud browser and sends the user a remote-session link instead of collecting credentials in chat. That adds friction. It also keeps the most sensitive step in the user's hands.

Reliability remains the expensive part of the promise. TechCrunch encountered an outage with Ollie's text provider, and Bill Lennon was direct: probabilistic models require defensive systems that catch a long list of small failures. In family life, a bad answer often becomes a broken handoff. A missed camp deadline, school notice, appointment, or payment lands on the person the software claimed it would relieve.

Khosla Ventures and AI House are backing more than another consumer chatbot. They are backing a wager that family context can become a durable product category if the company treats privacy, consent, and reliability as product mechanics rather than legal footnotes.

The $7.5M gives Ollie room to keep earning permissions one household at a time. Each permission expands what the assistant can do and raises the standard the team has to meet the next morning when the family expects the right message to arrive.

Crusoe

Chase Lochmiller and Cully Cavness started Crusoe by taking computers to energy the market was wasting. Eight years later, they are asking investors to back a larger choice: put power, data-center construction, accelerators, networking, and cloud software on one delivery schedule.

Bloomberg reports Crusoe finalized a funding round of more than $3B at an approximately $30B post-money valuation. Atreides Management, LP and Valor Equity Partners reportedly co-led, with Mubadala Capital participating. Crusoe has not published its own announcement, so the formal round label, ownership terms, and use of funds remain private.

Crusoe announced an initial $1.375B Series E close at an expected valuation above $10B in October 2025. The latest reported valuation roughly triples that level in less than a year. The number deserves attention; the obligation behind it deserves more.

Chase Lochmiller, Co-Founder and CEO, and Cully Cavness, Co-Founder, President, and Chief Strategy Officer, built Crusoe around vertical integration. Energy sourcing feeds campus development. Campus development supports high-density compute. Crusoe Cloud turns the physical stack into training and inference capacity. Nitin Perumbeti, CTO, owns the technology side of making that capacity usable.

Crusoe said in June that it had 4.9 GW of contracted AI infrastructure across data-center projects and cloud capacity, with a development pipeline above 40 GW. Contracted capacity is not operating capacity. The distance between them includes permits, interconnections, equipment orders, cooling, commissioning, and work that refuses to become easier because somebody called it AI.

The company has moved major projects through that sequence. The first phase of its planned 1.2 GW Abilene campus went live on Oracle Cloud Infrastructure in September 2025, roughly a year after construction began. Crusoe says the eight-building campus could support hundreds of thousands of GPUs at completion.

Bloomberg also reported a roughly $13B, five-year Crusoe Cloud contract with Jane Street. That agreement remains private and attributed, but it explains the pressure behind this raise. Demand is arriving in commitments large enough to turn a cloud forecast into a construction and delivery calendar.

Crusoe's next advantage will come from keeping those clocks aligned. Lochmiller, Cavness, Perumbeti, and their team now have more capital to do it, and a valuation that will make every delayed handoff easier to see.

BrightPlan

Two employees can share a logo, an org chart, and the same HR portal while living inside completely different financial systems.

One is sorting a 401(k), another is working through a pension. Tax rules move. Benefits change. Privacy expectations shift. Even the meaning of "financial advice" can cross a border and come back wearing a different regulatory uniform.

That is the real scale problem behind BrightPlan's first Series C close.

Founder and CEO Marthin De Beer built BrightPlan to deliver financial wellness through employers, combining an AI experience with planning tools, benefit navigation, and human advisors. The harder job is making the answer relevant to the country, employer, benefits package, and person asking it.

ABS Capital Partners is leading the first close. The amount and valuation were not disclosed, and BrightPlan expects additional investors over the next 90 days before a final close. Capital has arrived, but the complete Series C has not been announced.

BrightPlan says it supports more than 10 million employees across 50+ countries, after 230%+ revenue growth over three years and with 0% enterprise customer churn. Those company-reported metrics show why a growth-equity investor would look past the wellness label and see infrastructure.

The distribution is already inside the enterprise. Retirement platforms, benefits systems, HR technology, consultants, brokers, and advisory relationships have the trust and access. BrightPlan's job is to connect those rails so an employee does not need another disconnected app to understand the choices already attached to work.

The capital is expected to support product development, enterprise distribution, partner expansion, global advisory capability, and entry into additional countries. More reach also brings more regulation, cultural context, privacy responsibility, and opportunities for a generic answer to become a bad one.

ABS Partner Jennifer Krusius is joining BrightPlan's board. ABS is putting a partner inside the decisions that will shape how BrightPlan scales the AI, advisor network, and routes into large employers.

Financial wellness used to sound like educational content parked next to open enrollment. BrightPlan is betting it can become a year-round decision layer for benefits, debt, retirement, and the life changes that rearrange the plan.

The next phase will be decided far from the funding headline. It will happen when one platform meets millions of private decisions and still knows which parts should scale, which parts must stay local, and when a human needs to enter the conversation.

Poseidon Aerospace

An empty cockpit changes more than payroll. It changes how much of the aircraft can carry freight, how long it can stay in motion, and which routes still make sense when demand moves faster than a crew schedule.

That is the operating argument behind Poseidon Aerospace. David Zagaynov, CEO and co-founder, and Parker Tenney, CTO and co-founder, started the company in 2024 after working at Amazon and Lockheed Martin, respectively. Their objective is plain: lower cost per flight-ton-mile without waiting for aviation to become less complicated.

Poseidon announced an $11M Seed in November 2025 led by Tamarack Global, with Draper Associates, Starship Ventures, Drover Ventures, Cade Ventures, GoAhead Ventures, Fortitude VC, and angel investors participating. The round was intended to move Poseidon from its quarter-scale Seagull demonstrator toward 2 full-size platforms, Egret and Heron.

Egret uses paved or unpaved runways. Heron is a seaplane for water and coastlines. Poseidon lists 4,000 lb payloads, with 1,650+ nautical miles for Egret and 1,550+ for Heron. These are company specifications, not operational results.

Removing the cockpit and onboard life-support burden could leave more of the aircraft for cargo. Removing the need to position a pilot for every leg could add flexibility when freight demand moves between regions. Poseidon has said it wants to operate regional cargo service, putting utilization, reliability, regulation, and customer service inside the product.

The Seed is now part of a larger financing story. On September 8, 2026, TechCrunch reported a $60M Series A led by TQ Ventures, with Hanwha Asset Management, G Squared, and JAWS joining and Starship Ventures, Draper Associates, and Drover Ventures returning.

The update sharpens the evidence boundary. Poseidon’s November announcement expected flight testing in mid-2026. The current report places Egret’s first uncrewed test flight by year-end. The schedule moved, the capital increased, and the obligation stayed put: a full-size aircraft has to turn attractive weight, range, and utilization math into a repeatable flight operation.

Zagaynov and Tenney chose conventional fixed wings and internal-combustion powerplants instead of making exotic propulsion the story. The $11M Seed helped Poseidon leave the quarter-scale aircraft behind. The $60M Series A now carries the pressure to the hangar door, where Egret has to fly and the empty cockpit has to start earning its place in the economics.

ParlayX

A trading desk stops being one person long before the venue account notices.

One strategy becomes 4. One trader becomes a team. Bots arrive. Capital needs boundaries. Risk wants to know who can place an order, operations wants every fill in one report, and nobody sensible wants withdrawal authority attached to a trading credential.

Prediction markets have grown through interfaces built for individuals. ParlayX is financing the layer required when institutions show up with an org chart.

The New York company closed a $1.25M Pre-Seed led by Dreamcraft Ventures. Pet Berisha of Sporting Crypto, Jeffrey Haas of ID8, Jack Corddry, formerly of Frax Finance, and Devin Ardalan of EVG joined the round.

Co-founder and CEO Andrew Gonzalez and co-founder and CTO Antonio Cosio built ParlayX around questions that become expensive at scale. Which trader can use which capital? Which bot can touch which strategy? Who holds the keys? Can a desk see orders, fills, positions, and capital at risk across venues without rebuilding the day from several dashboards?

ParlayX answers with segregated books called pods, scoped credentials, non-custodial key management, order and execution management, deterministic market matching, and a unified portfolio view. Its documentation says customer keys remain in the customer's environment and regulated-venue access is trading-only.

That design makes the company's commercial wager unusually clear. The more prediction markets fragment across venues, the more valuable the layer above them can become. A desk may want the price and liquidity available across Kalshi, Polymarket, and other exchanges without giving every strategy the same balance, permissions, or operational risk.

ParlayX says it has onboarded 10 market makers. The customers are unnamed and the number is company-reported, so the real evidence will come from production volume, retention, and whether institutions trust the platform with increasingly complex books. Still, early demand helped move the company from a retail concept toward infrastructure for market makers, hedge funds, sportsbooks, exchanges, and systematic traders.

The capital is expected to expand venue integrations, improve smart routing, and scale go-to-market. Dreamcraft invests early across gaming, betting, sports, media, and the infrastructure underneath them.

Prediction markets make uncertainty tradable. ParlayX now has to make the responsibility around that trading visible enough for a team to share.

Flamingo

An MSP can negotiate a cheaper software contract and still lose the margin inside the handoffs. One dashboard opens the ticket, another controls the device, a third tracks security, and a technician becomes the translation layer holding the stack together.

That is the expensive gap Flamingo is trying to close. Founder and CEO Michael Assraf built the thesis after working with managed service providers through Vicarius: open source can lower the vendor bill, but the larger gain arrives when routine work stops bouncing between disconnected tools and people.

Flamingo has raised a $4.5M Seed led by Vertex Ventures Israel, bringing reported total funding to $6.7M after a $2.2M round in 2025. Existing investors also participated, according to Assraf's CRN interview, though the new announcement did not name them or disclose a valuation.

The product is OpenFrame, an open-source-based infrastructure layer for IT and security operations. OpenFrame Core puts multiple categories on one API and data model, while Fae handles lower-risk end-user requests and Mingo works across technician and fleet operations. Sensitive actions can still require human approval.

The first generation includes RMM, remote access, patching, PSA, SIEM, MDM, and documentation. Flamingo plans to extend that foundation across 19 product categories, but the larger product surface remains a roadmap.

The demand signal is visible, with an early-stage asterisk attached. Flamingo reports 400 MSPs using the beta across 10,000 endpoints and another 2,500 providers on the waitlist. Those are company-generated figures, and the round now has to turn interest into paying deployments.

MSPs live between customers who expect technology to work and vendors that charge for every layer required to keep it working. Consolidating software can lower one bill. Letting agents resolve repetitive requests, coordinate tools, and preserve approval boundaries could change how much customer work the same technical team can safely carry.

That is also where Flamingo inherits the harder obligation. A single operating layer concentrates trust. Automation has to behave predictably across many customer environments, while technicians keep control over the work that should never disappear behind an agent.

Flamingo's beta supplied the demand. The Seed gives Michael Assraf and the team room to turn it into a working commercial system, one customer environment and one safely closed ticket at a time.

Cloverleaf Bio

Transfer RNA normally helps cells turn genetic instructions into proteins. Cloverleaf Bio is engineering the same molecular courier to carry a more hostile assignment: trap several RNA-modifying enzymes that fast-growing cancer cells depend on.

That idea has moved from a Yale lab into a $33M Seed. 4BIO Capital led the upsized, oversubscribed financing, with AbbVie Ventures, Eli Lilly and Company, Boehringer Ingelheim Venture Fund, Draper Associates, Mission BioCapital, and American Cancer Society BrightEdge participating.

Three pharmaceutical venture arms are backing Cloverleaf before human data. That breadth marks interest in the science, not human proof.

Co-founder and CEO Austin Draycott, PhD, co-founder and CSO Cole Lewis, PhD, and co-founder and Scientific Advisory Board chair Wendy Gilbert, PhD, built Cloverleaf around an awkward feature of cancer biology. Tumor cells can become unusually dependent on enzymes that modify RNA as they proliferate. Interfering with several dependencies at once could create another route around resistance.

Cloverleaf Bio's lead compound, CLB-001, is an engineered tRNA payload. The company reports greater potency and selectivity for cancer cells than healthy tissue in preclinical studies, activity in models resistant to frontline chemotherapy and current-generation ADC payloads, and no systemic-toxicity signal across tested animal doses. These are laboratory and animal findings, not human efficacy or safety data.

The financing will push an inhibitory tRNA program for hepatocellular carcinoma toward early clinical testing and move an antibody-tRNA conjugate for colorectal cancer through preclinical development. The National Cancer Institute helped fund the foundational work through its small-business research programs, giving the round a public-to-private handoff worth more attention than the usual venture-capital victory lap.

The science is clever. The work ahead is brutally physical: reproducible manufacturing, delivery, pharmacology, toxicology, regulatory evidence, and eventually a human body that has no obligation to behave like a model. Cloverleaf has not disclosed an IND timeline, trial design, or first-patient date.

That is what a $33M Seed buys in therapeutics: distance between a mechanism that makes sense and a medicine that can survive every handoff required to reach a patient. Cloverleaf Bio now gets to carry its engineered tRNA out of the lab, and the next record will be written by the data that arrives with it.

BrainChild Bio

Pediatric oncology has a market-sizing problem that can arrive before the science gets a fair hearing.

Diffuse intrinsic pontine glioma affects about 300 children in the U.S. each year. Standard care remains largely limited to radiation, and median survival is about 11 months. The need is brutal. The market is small enough to make conventional biotech investors hesitate anyway.

That tension sits inside BrainChild Bio's $116M Series A. An undisclosed family fund and foundation led, with Seattle Children's and Washington Research Foundation's WRF Capital participating. CEO Steven Brugger said institutional investors liked the data, but the pediatric numbers did not fit their models.

Founder and CSO Michael Jensen has spent years building the other side of that equation. BrainChild Bio emerged from Seattle Children's in 2023 around his team's CAR T research, clinical work, and manufacturing capability.

The lead program, BCB-276, is an investigational autologous CAR T therapy targeting B7-H3. It is delivered into cerebrospinal fluid, closer to the tumor, and supports repeat dosing without routine lymphodepleting chemotherapy.

A Phase 1 study treated 21 patients with DIPG and reported that repeated dosing was feasible and tolerable at the studied regimens. Median survival from diagnosis was 19.8 months, with 3 patients alive beyond 44 months at data cutoff. That was a small, single-center study without a randomized control group, so the signal earned a larger test, not a victory speech.

The capital will fund ILLUMINATE, a recruiting pivotal Phase 2 study expected to enroll 75 participants at 6 U.S. centers. It will also support BCB-214 for glioblastoma and other CNS tumors.

This is where the round becomes more than biotech arithmetic. Seattle Children's carried the work from a hospital research program into an independently managed company. WRF Capital joined. A family fund and foundation stepped into the lead position that conventional institutional capital would not take. That choice kept a rare-disease program moving toward the study that can settle the question.

Mission-aligned money cannot make an experimental therapy work. It can finance the manufacturing, regulatory discipline, and multisite clinical evidence required to find out.

BrainChild Bio now has the runway to move a pediatric-first program through the most expensive handoff in drug development. The children entering ILLUMINATE will supply the evidence that determines where this science can go next.

Stoke Space

Stoke Space is financing rocket number two before rocket number one has flown.

That sounds backward until you look at what a launch company builds around the vehicle. Reaching orbit is one test. Producing rockets, operating a pad, recovering both stages, and finding room for customers is an industrial system.

Andy Lapsa and Tom Feldman founded Stoke Space in 2019 around a stubborn idea: both stages should return and fly again. The company has now completed the initial close of a $1B Series E co-led by Point72 Ventures and Spark Capital, bringing total capital raised to a company-reported $2.3B.

General Innovation, Glade Brook Capital, US Innovation Technology, Washington Harbour Partners, Woven Capital, Y Combinator, and others joined the round. They are backing two development clocks at once.

Nova Pathfinder is the first clock. The fully reusable orbital configuration is moving through qualification toward an early 2027 first-flight target. Its job is to carry customer payloads, test Stoke's reuse architecture in orbit, and give the team operating experience no ground test can manufacture.

Nova Block 2 is the second. Stoke says the larger vehicle will deliver 15 metric tons to low Earth orbit in a fully reusable configuration, using the propulsion, manufacturing, reentry, and operating lessons built through Pathfinder. Its first launch is targeted for 2029.

The technical handoff matters. Pathfinder flies with 7 Zenith full-flow staged-combustion engines; Block 2 is planned with 14. Both use stainless-steel structures and the same production logic, so Pathfinder's tooling, tests, and operating history can keep working instead of becoming a museum exhibit.

The Series E funds flight preparation, production, test, launch, and recovery infrastructure, plus Block 2. Reuters reports a valuation of roughly $10B, but Stoke did not disclose one. Launch dates, recovery, rapid turnaround, and cost remain claims that flight has to settle.

That is the wager Point72 Ventures and Spark Capital are making before the first orbital attempt. Waiting for perfect proof could leave the factory, pad, recovery system, and larger vehicle years late. Scaling now puts $1B behind a schedule that physics can still edit.

Stoke is choosing overlap. Pathfinder has to turn the architecture into evidence while Block 2 turns the same architecture toward the heavier constellation missions customers say they need. The capital buys time on both clocks, but only flight can make them agree.

Dawraty

Institutional education has an awkward unit of sale. The learner watches the lesson, but the buyer must prove the right person completed the right training in the right language.

That distance between learning and accountability is where Dawraty Education Group is building. The Kuwaiti edtech company has secured Qatar Development Bank backing as part of a $2M Seed round. Dawraty reports a $10M valuation, and the round remains open to additional investors through November 2026.

Ryan Dougherty, M.D., M.S., co-founder and CEO, Hamad AlThunayan, co-founder, and Naser AlQimlas, co-founder and CTO, started with a problem they knew close to the ground. Students in demanding fields were stitching together PDFs, videos, tutoring, and notes that rarely matched their local curriculum or pace.

Dawraty turned that problem into a bilingual platform for healthcare education, professional certification, and exam preparation. Its institutional customers can deliver and track accredited learning without asking every learner to assemble a private system from scattered material.

QDB is entering after Dawraty began proving a route into Qatar. A recent partnership and reseller agreement with Qatar Finance and Business Academy created the company's first contracted institutional revenue in the country. The investor is supporting market entry that already has a customer relationship attached to it.

Across Kuwait, Qatar, Bahrain, and Jordan, Dawraty reports more than 15,000 active learners, 207% year-over-year revenue growth, profitability at the Seed stage, a 47x LTV-to-CAC ratio, and 76.8% cohort retention. These are company-reported figures. Institutional education software has to earn trust twice, first from the organization signing the agreement and then from the learner returning to the platform.

QDB's Startup Qatar program is designed to attract technology companies that will establish or expand operations in the country. Dawraty arrives with a bilingual product, an institutional buyer, and a regional distribution model that fits the mandate.

The capital is expected to expand Dawraty's offerings across Gulf and Arab markets. Underneath that plan sit the harder jobs: mapping curricula, integrating partner requirements, proving learner outcomes, and turning reseller agreements into durable revenue.

Every new market will test whether Dawraty can preserve the local detail that made its platform valuable while making the operating model repeatable enough to travel.

Aukera

A credit accordion is a promise with an asterisk. The lender may allow a facility to grow, but the borrower still has to earn the commitment that turns optional capacity into capital it can plan around.

Aukera Energy just crossed that line. Its EIG-led structured credit facility now holds €460M of committed capital after a €260M Series 2 tranche replaced an accordion that previously offered up to €250M of additional capacity.

That accounting matters. The original structure contained a firm €200M commitment, so this is a €260M increase in committed capital, not a fresh €460M equity round. The headline got larger because uncertainty got smaller.

Pascal Emsens and Catalin Breaban founded Aukera in 2021 after nearly a decade working together in renewable-energy investing. The company develops, finances, constructs, and operates battery storage and renewable infrastructure across Belgium, the UK, Germany, Romania, and Italy.

Those verbs sound tidy in a sentence. On the ground, each one has a different clock, contract, regulator, grid connection, construction risk, and revenue model. A 14 GW pipeline can look enormous in a deck while every individual project still has to survive those handoffs.

Aukera says close to 1 GW is already in construction or operation. The evidence includes the first 150 MW/300 MWh phase of the Gura Ialomitei battery project operating in Romania and the 170 MW/340 MWh Project Volt in Belgium reaching financial close with €97.5M of debt.

That record helps explain EIG's larger commitment. Rob Johnson, President and CIO of EIG Credit Management, pointed to Aukera's multi-market portfolio and delivery capabilities. AtlasInvest, Reggeborgh, and SFPIM, Belgium's sovereign wealth fund, remain its equity backers.

The European Commission estimates that EU storage capacity needs to rise from about 55 GW at the beginning of 2026 to 200 GW by 2030. Policy can publish that number in one line; developers still have to permit, finance, build, connect, and operate the assets behind it.

The facility's pricing, maturity, covenants, security, other participants, and project allocations were not disclosed. Neither was a valuation, because this is credit rather than equity. Those omissions keep the story honest without making it smaller.

Aukera has replaced part of its financing optionality with institutional commitment. The next movement belongs to the teams carrying 14 GW through five markets, one grid connection and commercial-operation date at a time.

Moonwalk Biosciences

Eli Lilly and Company just backed an obesity program designed to work somewhere its blockbuster incretins do not start: inside adipose tissue.

Moonwalk Biosciences closed an oversubscribed $70M Series B co-led by Alpha Wave Global and YK Bioventures. Lilly and Gaorong Ventures joined as new investors, while ARCH Venture Partners, Khosla Ventures, and Future Ventures returned. Moonwalk has now disclosed $127M across its 2024 and 2026 financings.

CEO and co-founder Alex Aravanis, President and CTO Arash Jamshidi, and Chief Platform Officer Justin Valley launched Moonwalk publicly to find disease-driving targets in the epigenome.

Then the company made a very biotech choice. Moonwalk used human genetics, epigenomics, and multi-omics to identify targets in fat cells, but chose siRNA as the faster route toward the clinic. Its adipose-targeting chemistry is exclusively licensed from Suzhou Siran Biotechnology.

That decision makes the platform story useful. Discovery companies can discuss possibility for years. Patients and regulators eventually require one target, delivery route, candidate, and clinical plan.

MW101 is that first narrowing. The candidate is designed to deliver siRNA directly to adipose tissue and modulate a novel non-incretin pathway without suppressing appetite. Follow-on programs address other adipose mechanisms.

Moonwalk reports lower body weight and fat mass in preclinical studies while preserving lean mass and without reducing food intake. It also reports durable adipose target engagement after one dose in non-human primates, supporting infrequent dosing.

Potential is doing important work there. MW101 has no human safety or efficacy data, disclosed gene target, or registered clinical study. The Series B will fund IND-enabling work and a first-in-human study targeted for late 2027.

Lilly's participation sharpens the context without inventing a partnership. A company with blockbuster incretin medicines is investing in non-incretin, fat-cell biology. It is not clinical validation, but it shows where obesity R&D is widening.

Moonwalk also appointed Stephen Djedjos as SVP and head of clinical development, moving clinical execution closer to the center of a company founded by scientists and platform builders.

The $70M gives Moonwalk time to turn an epigenetic discovery engine, licensed delivery chemistry, and preclinical evidence into one candidate ready for people. The platform chose a direction. The clinical organization now has to carry it to late 2027.

Blee

Legal teams used to inherit marketing's final draft. Now they inherit 50 variants, a video, 12 ads, and an AI agent capable of producing another batch before lunch. The old review queue did not acquire another day in the week.

That mismatch sits inside Blee's $20M Series A. Fin Capital and SMBC Fin Atlas Beyond Fund co-led, with Hannah Grey VC and National Bank of Canada participating. Y Combinator, Penny Jar Capital, Cardumen Capital, and Treasury continued their backing, bringing Blee's total disclosed funding to $27M.

Founder and CEO Guy Shahar built Blee around a problem that looks administrative until a claim reaches a customer. Marketing wants speed. Legal and compliance need context, precedent, evidence, and a record of who approved what.

Blee moves judgment earlier. Its AI pre-reads marketing, product, and sales materials against regulatory requirements, brand standards, internal policies, and a company's risk tolerance. Creators receive feedback before every issue lands on a lawyer's desk, while approvals, comments, versions, and decisions stay attached to an audit trail.

The platform keeps watching after the campaign goes live. Blee monitors websites, social channels, partner pages, and influencer content for changes, stale claims, brand drift, and material that skipped formal review. That matters because the enterprise remains accountable for every public claim after the content generator moves on.

Blee says customers cut average review time by up to 65%. The company reports more than 20M assets reviewed and revised, along with 10x ARR growth between June 2025 and June 2026. Those figures are company-reported, but the buying signal includes SoFi and work across financial services, travel, life sciences, and consumer brands.

The investor mix follows the product. Fin Capital brings a B2B fintech network, while SMBC's fund connects Blee to regulated financial institutions. The system still has to learn each customer's rules before it can earn trust.

The capital will deepen Blee's full-lifecycle capabilities, open more industries and geographies, and expand distribution across North America and Europe. That widens the obligation too: more rules, content formats, jurisdictions, and chances for context to disappear between creation and approval.

Blee is financing the system that keeps legal judgment attached to content while the volume moves. The next chapter will be written in all the claims its customers can govern before the queue owns the company.

Emberos

A brand can own the website, buy the media, brief the agency, approve every campaign, and still lose control of the answer that reaches the customer.

That answer may now arrive from ChatGPT, Gemini, Claude, Perplexity, or another AI system stitching together signals across the web. The logo belongs to the company. The assembled reputation does not.

Justin Inman built Emberos around that uncomfortable handoff. The Los Angeles company has raised an oversubscribed $5.5M Seed from a private consortium of strategic angel investors, bringing its reported total funding to $6.7M after a $1.2M pre-seed in January.

Most AI-visibility products begin with measurement: where a brand appears, how often it gets cited, and what sentiment surrounds the answer. Emberos is reaching for a larger operating loop. Its Brand Knowledge Graph connects entities, prompts, model responses, sources, recommendations, actions, and outcomes, then puts 5 agents on top of that record.

Scout watches how the brand is represented. Pilot forecasts where that representation may move. Flow coordinates interventions. Merchant carries the work into commerce. Echo audits what changed and whether the movement held.

That architecture matters because an AI answer is not a search ranking wearing a new jacket. A model can absorb the company site, press coverage, executive interviews, creator material, reviews, and old claims that never got the memo. Fixing one page may improve the source without repairing the story the system already assembled.

Emberos already has a serious proving ground. Stagwell says Search+ was built by Assembly in partnership with Emberos, whose agents provide the AI-search intelligence layer. Emberos also reports work with The Honest Company, RA Capital Management, Monigle, WHITESPACE, and BERO, though those relationships remain company-described here.

The new money is expected to accelerate CreatorGraph, an extension that maps creators, communities, and independent publishers alongside brands and products. The people influencing an AI recommendation may sit outside the channels a brand normally manages, while their work becomes part of the evidence a model trusts.

Emberos now has capital to prove that scattered representation can become an operating function. The hard work will live between diagnosis and correction, where a company must show what source moved, why the model changed, and whether the customer outcome followed.

That is where AI brand management starts earning a place in the enterprise stack.

Forus

Forus spent its first 2 years making no revenue, by design. Founder and CEO Sahir Jaggi kept the product free for doctors and patients while the team built distribution through the part of healthcare where a correct prescription can disappear into approvals, affordability programs, pharmacy routing, calls, and faxes.

That choice now has a $3B valuation attached to it.

Forus has raised a $150M Series C led by Bain Capital Ventures (BCV). Every existing institutional investor reinvested, including Thrive Capital, General Catalyst, Accel, Redpoint, BoxGroup, Pear VC, Avra, Human Capital, Neo, Vast Ventures, and SV Angel.

The capital accounting deserves a clean line. Forus said in May that it had raised more than $160M in total as the company emerged from its former Tandem name. This is new Series C capital, taking disclosed funding above $300M. The valuation is roughly 3 times the level reported 4 months ago; other terms remain undisclosed.

The product begins after the doctor makes the clinical decision. Forus assigns each prescription an AI agent that works through insurance coverage, appeals, financial assistance, pharmacy routing, and fulfillment. It must reason across incomplete information and act inside systems Forus does not control.

That last part changes the meaning of “AI agent.” A weak answer in a demo wastes a minute. A weak action inside medication access can delay treatment for someone whose physician already decided what should happen.

Forus says the network supports millions of people across all 50 states, reaches 85% of U.S. residential ZIP codes, and works with 9 of the top 15 global biopharma companies. Those company-reported figures explain the free product: every prescription adds another connection among providers, patients, payers, pharmacies, and drugmakers.

Biopharma is the commercial side. Forus can help drugmakers see where access breaks, prepare launches, and understand how treatment moves after approval. Pricing and contract economics remain private.

The Series C will support more specialties, more care settings, broader technical capability, and a larger team. Bain is leading after backing Forus from the early stage, while the full institutional group is increasing its exposure only months after the last disclosure.

Forus wants to reach every doctor's office in the country. The quality of that ambition will show up far from the funding announcement, in the days a patient spends waiting and whether the treatment a doctor chose actually begins.

Tanda

Four university housemates built Tanda for 14 years without an outside investor. The first yes was never going to be a casual financing decision.

Jake Phillpot, Alex Ghiculescu, Tasmin Trezise, and Josh Cameron started the company after managing people at a university venue. Rosters, hours, and pay rules kept finding ways to disagree.

That problem became a workforce platform built around a serious promise: the schedule, clock-in, award interpretation, wage calculation, payroll, onboarding, and HR record should tell the same story.

Thoma Bravo has now made a strategic growth investment in Tanda, the company's first outside capital since its 2012 founding. The amount, valuation, stake size, control terms, and board arrangements were not disclosed. Tanda's co-founders will remain significant shareholders and continue leading the business, with Jake Phillpot staying CEO.

Taking capital after 14 bootstrapped years means choosing who enters the decisions about product speed, new markets, and the parts of the business that refuse to get compromised.

Tanda says it serves approximately 8,000 businesses across hospitality, retail, quick-service restaurants, healthcare, and other frontline industries. In Australia, the product grew around Modern Award interpretation and Fair Work compliance. Internationally, the company operates as Workforce.com.

For a shift-based employer, workforce software is where operational math becomes a human outcome. Demand becomes a roster. The roster becomes clocked time. Local rules change the calculation. Payroll turns the chain into money someone expects to arrive correctly.

Thoma Bravo Partner Carl Press described the risk in employers' patchwork of legacy systems. Principal Adam Kinalski pointed to the product-first founders and their operating execution. This software sits beside a recurring obligation, not an optional workflow.

The capital will support product innovation, Tanda's AI roadmap, and new markets. Alex Ghiculescu remains co-founder and CTO, while the founding group stays economically and operationally attached.

AI can build a roster quickly, but the answer still has to respect availability, labor budgets, employment rules, time records, and payroll. Each new jurisdiction adds another way for those records to drift apart.

The next evidence will arrive quietly: schedules published, clock-ins recorded, rules interpreted, and people paid correctly in markets Tanda has not reached yet.