The Check Comes With a Person

Moneyball.ai

Private-capital deals rarely die in one cinematic moment. They fade through missed follow-ups, investor context buried in somebody's inbox, and the assumption that the banker who opened the conversation will remember every detail 3 weeks later.

Moneyball.ai is building inside that administrative fog. The San Juan company has joined the super{set} portfolio, with the venture studio contributing capital, operators, developer talent, go-to-market support, and hands-on execution. The September 9 announcement leaves the check size, round total, security, valuation, ownership, and close structure undisclosed.

Co-founder and CEO Julia Whippo calls the product Moe Money, an AI deal-team partner for boutique investment banks and SPV sponsors. It is designed to carry investor research, matching, outreach, scheduling, NDA workflows, CRM tracking, and follow-up across one system.

Private capital is a relationship business with an administrative tail long enough to consume the senior people responsible for those relationships. Every introduction creates context, every meeting creates another promise, and every delay changes timing. Moneyball.ai is trying to make that institutional memory executable without pretending software becomes the institution.

Its disclaimer draws a useful boundary: Moneyball.ai is not a broker-dealer or investment adviser. The product can organize work around a transaction while regulated securities activity remains with principals and registered firms. Buyers still need to know who owns the judgment, permissions, record, and client relationship when automation reaches the edge of its authority.

A November 2025 partnership with FINRA-member Regiment Securities put the product in a boutique-bank environment. Co-founder Michel Bayan is helping build it without losing the human context that makes a match useful.

super{set} fits this stage because its model is more workshop than cap table. General Partner Peter Day described bringing capital, talent, company-building experience, and operators alongside founders. Moneyball.ai must connect product, integrations, compliance design, customer implementation, and go-to-market execution at the same time.

The opportunity is to stop spending human judgment on calendar repair, CRM archaeology, and the seventh follow-up that should have remembered the first 6. Moe Money will earn its seat when context survives from investor discovery to the next call months later, and the relationship keeps moving without feeling processed.

Eubrics

A training dashboard can show 100% completion while Monday's sales call still falls apart. The boxes are checked, the manager has a report, and the customer still hears a rep searching for the right answer in real time.

That gap is where Eubrics AI is building. On September 9, 2026, super{set} added the company to its portfolio with undisclosed early-stage backing described in funding coverage as Pre-Seed. The official announcement does not reveal a check size, valuation, or ownership terms. It does reveal the more useful part: capital and hands-on developer, operator, talent, and go-to-market support.

Nikita Jain, Co-founder and CEO, spent more than a decade around HR consulting, capability development, and enterprise transformation. Founder and CTO Maxim Dsouza brings engineering leadership from Apple and earlier startups. Their product turns training from a scheduled content event into repeated practice.

Eubrics creates AI buyer personas and work scenarios for sales teams. Reps can rehearse outreach, discovery, objection handling, negotiation, renewals, and expansion, then receive immediate feedback against the competencies and methods their organization actually uses. Managers see where a team is ready, where judgment breaks, and where coaching should go next.

That distinction matters because enterprise learning has spent years measuring what it can count easily: attendance, completion, hours, modules. The buyer funding the program cares about readiness, compliant execution, conversion, and revenue. Moving between those two scoreboards is the commercial job Eubrics has chosen.

super{set} names Kohler, Accenture, Konica Minolta, Moog, and Clifford Chance among Eubrics users. The company also reports an anonymous insurance deployment cut ramp time from 90 days to 15 days, increased closures 19%, and raised revenue per lead 25%. Useful evidence, but company-reported rather than independent proof.

The next operating test is North America. June hiring materials show Eubrics building a financial-services motion across insurance, wealth, and banking. There, an impressive simulation must survive procurement, compliance, buying committees, and a proof-of-value pilot.

super{set} is helping Eubrics build the machinery that turns product evidence into repeatable enterprise contracts. The deeper bet is that sales practice can become a measurable operating system, with every difficult customer conversation feeding the next rep before that rep has to learn it live.

Harvey / Guardrails AI

A legal AI agent cannot call a bad answer a learning experience. It may already be sitting in a client memo, a diligence file, or a contract workflow that ran for 3 hours before anyone noticed the system had drifted.

Harvey has acquired Guardrails AI, bringing co-founders Shreya Rajpal and Zayd Simjee and their team into its product and engineering organization. They will begin work on Applied AI, putting Guardrails AI’s reliability expertise underneath high-consequence professional work.

The price, consideration, closing mechanics, and retention terms were not disclosed.

Guardrails AI spent 3 years building tools for the part of AI adoption that arrives after the polished prompt. Its open-source framework helps developers validate inputs and outputs and control risky behavior. Snowglobe creates synthetic users that push agents through edge cases before real users discover them. Harvey reports more than 250,000 monthly downloads.

That experience fits the problem Harvey now owns. Its agents work across matters, documents, firm knowledge, and workflows that can continue for hours or days. A system that takes 12 reasonable steps, loses the instruction on step 13, and delivers the mistake with complete confidence is a different operating problem.

Harvey’s Legal Agent Benchmark found that frontier systems completed fewer than 10% of held-out long-horizon legal tasks under an all-pass standard. Harvey also earned AIUC-1 certification after testing across 86 risk categories. The acquisition connects those ambitions with a team built around specifying, simulating, and controlling behavior.

This is Harvey’s fourth acquisition of 2026. Hexus added product-demo talent. Lume added integration depth. Benchmark extended Harvey into asset-management decisions. Guardrails AI moves closer to the layer that decides whether those agents can be trusted after the demo ends.

Guardrails AI raised a $7.5M Seed led by Zetta Venture Partners in 2024. Harvey also names Bloomberg Beta, Pear VC, Factory, and Microsoft as investors, while revenue, customers, headcount, and exact lifetime funding remain undisclosed.

Shreya Rajpal and Zayd Simjee are bringing a way to rehearse failure before a lawyer, client, or regulator has to explain it in production.

The next handoff belongs to Harvey’s product teams. They have to turn simulation and guardrails into evidence a firm can inspect while the agent moves across documents, permissions, and judgment that cannot be reset with a clean browser tab.

HelmGuard

A compliance report can be accurate on Friday and misleading by Monday.

The vendor shipped a feature. A model changed. An agent gained another tool. The approved annual questionnaire now describes a system that no longer exists.

HelmGuard AI is building for that gap. John Daley, co-founder and CEO, and Jack Miller, co-founder and CTO, have raised a $7.3M Seed round co-led by Infinity Ventures and Frontline, with FinTech Collective, Stage 2 Capital, and Entrepreneur First participating.

Most GRC software helps teams collect the paperwork around a decision. HelmGuard wants its agents closer to the decision itself. Its platform connects risk, security, compliance, and vendor information, then applies specialized agents to third-party risk, control gaps, customer assurance, and AI-agent oversight.

That is more useful than faster form filling, and more dangerous to get wrong. An automated risk conclusion has to carry evidence, expose its reasoning, show uncertainty, and leave a human able to defend the decision to an auditor, regulator, customer, or board.

HelmGuard says citations, reasoning traces, confidence mechanisms, and human review create that chain. One U.S. insurer reportedly assessed 1,250 counterparties in less than a week and migrated from a legacy system in under 10 days. Another customer cut assurance first-response time from days to minutes. The results are company reported, but the job is clear: less evidence chasing, more current judgment.

The investor mix fits the work ahead. Infinity Ventures understands regulated fintech infrastructure. Frontline helps European B2B companies cross the Atlantic. Stage 2 Capital brings go-to-market operators to a sale involving security, compliance, procurement, technology, and business owners.

The funding will support hiring, U.S. expansion in New York and San Francisco, a runtime assurance layer for AI agents, and HelmGuard's Verified Risk Network. The network is meant to exchange current, verified claims instead of sending the same static documents through another diligence cycle.

John Daley spent 8 years at Palantir around the consequences of fragmented enterprise risk information. Jack Miller came through AI research and secure-model work. HelmGuard now has to connect those disciplines where speed is welcome, but explainability is the admission ticket.

The harder HelmGuard succeeds at making agents useful, the more often those agents will change the risk they are supposed to measure. That loop is where this $7.3M gets put to work.

Nanoramic

Nanoramic has spent 17 years commercializing around an industrial fact battery chemistry likes to ignore: the factory already has a job.

A new material can deliver beautiful cell data and still lose when adoption asks a customer to replace coating equipment, rework drying lines, interrupt throughput, and defend the capex before the first production order arrives.

Founder and CEO John Cooley built Nanoramic around a different handoff, alongside CTO and Chief of Strategy Nicolò Brambilla, who joined in 2010. Neocarbonix Slurry Precursor, or NXSP, replaces conventional electrode binders and additives with a conductive carbon structure designed to run through existing roll-to-roll manufacturing. The pitch is better battery performance without a factory identity crisis.

Nanoramic has announced the first close of its Series 2 equity financing at a $250M pre-money valuation. Catalus Capital led, with GM Ventures and returning investors participating. The company did not disclose the amount of this close.

A July Form D reported $9.01M sold toward a $15.02M equity offering, but the September announcement does not say this close equals that filing. The clean accounting is a first close at a $250M pre-money valuation, not a $250M raise.

The timing carries more information than the headline. Nanoramic said it made its first commercial NXSP shipments from Asian mass-production facilities during Q1 2026 to Tier 1 battery makers, auto OEMs, electronics manufacturers, and power-tool companies through a supply chain coordinated with ITOCHU. It says shipments grew over the past 6 months and programs now include defense contractors.

Those milestones are company-reported. Customer names, shipment volumes, revenue, yield, and independently measured economics remain private. The operating standard is now qualification, delivery, and repeat orders across factories that punish surprises.

Catalus is returning after co-leading Nanoramic's $54M Series 1 with GM Ventures. One brings a long view of the company; the other knows how expensive manufacturing change becomes inside an OEM.

Battery buyers want more energy, more power, lower drying costs, fewer fluorinated materials, and stronger supply chains while protecting the infrastructure already on the floor. Nanoramic is trying to make those demands coexist inside one material decision.

The next evidence will arrive in qualification cycles, purchase orders, repeat shipments, and the factory math behind somebody else's product.

Atrium

A consulting firm that makes implementation faster is volunteering to invoice fewer hours.

That used to sound like a pricing problem. Atrium and Sverica Capital Management see a growth model.

Sverica has made a strategic growth investment in Atrium, the Bozeman company Chris Heineken, Colin Gelfer, and Geoff Birnes founded in 2018. The amount, valuation, ownership, and deal structure were not disclosed. Chris Heineken will keep leading the company, while Sverica Managing Partner Jordan Richards and Principal Matt Johnson join the board.

Atrium works across Salesforce, Snowflake, data engineering, analytics, implementations, and managed services. It also built Andi, an AI agent that can help design Salesforce solutions, configure systems, write code, and run QA.

The interesting part is not that a consultancy uses AI. Every services company has discovered that sentence. The interesting part is what Atrium is willing to let AI compress.

Traditional consulting can hide inefficiency inside the invoice. More meetings, handoffs, and development hours can still produce more revenue. Andi pushes in the opposite direction. Atrium is arguing that the valuable unit is the working outcome, not the time it took a team to manufacture it.

That changes where the firm must be excellent. Faster configuration does not clean the data, settle ownership, win employee trust, or decide which workflow should change. Atrium still has to carry customers through the parts of transformation that refuse to become a prompt.

Sverica knows this corner. The firm backed Salesforce consultancy Coastal Cloud, which it says more than quadrupled in size before Tata Consultancy Services acquired it for $700M. Sverica looks for technology and business-services companies where repeatable economics can turn expertise into a larger platform.

Atrium now has to show how that playbook changes when the delivery engine contains its own AI worker. The opportunity is a consultancy that learns faster, ships more consistently, and stays close after launch. The risk is treating speed like the finished product while adoption, governance, and data quality keep charging by the month.

Enterprise AI has no shortage of models. It has a shortage of clean handoffs between the platform, the data, the workflow, and the person whose name remains on the result.

Sverica is financing Atrium at that handoff. The next chapter gets measured in how much complexity disappears for the customer without the relationship disappearing with it.

Savvy Wealth

Ritik Malhotra sold his first software company to Box, went looking for financial advice, and spoke with more than 100 advisors. The profession made sense. The machinery around it did not.

That search became Savvy Wealth, built with co-founder and former CTO Muller Zhang around an industry choice: independent advisors could keep control of their books, or gain the institutional support needed to scale. Founder and CEO Ritik Malhotra has spent 5 years arguing that they should have both.

Savvy Wealth has now closed an oversubscribed $100M Series C led by Halo Fund, the growth investor co-founded by Ryan Smith and Ryan Sweeney. Returning investors include Thrive Capital, Industry Ventures from Goldman Sachs, Canvas Prime, Index Ventures, House Fund, Euclidean Capital, Alumni Ventures, and Vestigo Ventures.

InvestmentNews and WealthManagement.com report a $600M valuation. The round takes Savvy above $200M in total capital after a $26.5M Series A and a $72M Series B.

The numbers explain why the investor roster returned. Savvy says it supports more than 150 advisors overseeing $9B in client assets, recruited more than $4B in assets during 2026, and is on pace to exceed $100M in annual recurring revenue.

Those are company-reported metrics. The operating appetite is still visible. Every additional advisor brings onboarding, planning, compliance, reporting, tax coordination, investment work, marketing, and the expectation that none of it should feel like a software scavenger hunt.

Savvy Intelligence is the operating layer built for that load. It connects CRM, investments, tax, and planning data so AI agents can prepare reports, analyze financial plans, and surface work for advisor review. Current CTO Eric Hurkman leads the technology effort.

The regulatory boundary matters. Savvy Wealth is the technology parent; Savvy Advisors is the SEC-registered RIA. The AI does not advise retail clients, place trades, manage allocations, or make investment decisions. Advisors remain responsible for the advice.

The new capital will extend Savvy Intelligence across more of the advisor's day, broaden tax and estate services, and grow the technical team. Malhotra told InvestmentNews the company expects to move from more than 100 employees to hundreds.

That creates a demanding handoff. Savvy must scale regulated service, software, and human judgment at the same time while preserving the independence it sells. The advisor who keeps the book will notice quickly if the infrastructure starts owning the day.

Overroute

J.B. Hunt gave a startup the part of freight technology most pilots never survive: live loads, legacy systems, impatient customers, and operators with no time to applaud a clever demo.

Overroute spent roughly 1 year building inside that environment before announcing a public launch. Now it has raised $5.5M led by UP.Partners, with EIC Rose Rock, 1834 Ventures, Tulane Ventures, and other investors participating.

The order matters. Enterprise AI usually arrives with a model, a polished dashboard, and a request for the customer to reorganize reality around it. Overroute started with the work already stealing hours from freight teams: monitoring loads, handling exceptions, scheduling deliveries, communicating with drivers, updating customers, and closing loops across systems that do not cooperate politely.

Its agents connect to transportation management systems, ERP software, visibility platforms, and carrier data. Skilled operators keep the decisions where context and judgment matter. Software takes the repetitive chase around those decisions.

That design was shaped alongside J.B. Hunt operators across 5 business lines. The companies report that Overroute now runs across the carrier's business units and touches millions of loads. That is meaningful production exposure, though not an independent ROI study.

Founding CEO Alex Reed learned another industrial version of this lesson at Fluence Analytics, the process-monitoring company he co-founded and led. The clean plan meets live operations, then the exceptions start charging rent.

The $5.5M will accelerate product development and staffing as Overroute moves into more large carriers and private fleets. The roadmap extends from individual-load coordination into freight planning and asset optimization.

That creates the harder commercial assignment. The next carrier brings different data, systems, rules, customer commitments, and institutional scar tissue. Overroute has to carry the operating logic without pretending every fleet is J.B. Hunt wearing another logo.

UP.Partners is backing a company that crossed the pilot threshold once. The case now lives in repetition: how quickly Overroute earns trust inside the next operation, when its agents know a human should step in, and whether preserved handoffs create better decisions.

Freight technology spent years making loads visible. Overroute is financing the layer that helps somebody do the next thing with what the network can already see.

DermaSensor

A suspicious skin lesion has a geography problem.

It appears in primary care, where a clinician must decide whether the patient needs dermatology. The specialist is somewhere else in the referral chain, and delay can begin before the first appointment is booked.

DermaSensor built a handheld second read for that moment. It uses elastic scattering spectroscopy and AI to evaluate a lesion already considered suspicious, giving a non-dermatologist physician an objective risk result for the referral decision. The FDA cleared it in 2024 for adults 40 and older. It is an adjunct, not a screening tool or standalone diagnosis.

Cody Simmons and the team closed a $4M Series B extension from institutional and strategic investors. A separate Wefunder community round has raised more than $1M on the same terms, bringing company-reported total investment to $48M. The current investors and lead were not named. The $4M is the institutional extension; the $5M-plus headline combines 2 financing lanes.

A week earlier, DermaSensor announced 4 U.S. commercial relationships. Henry Schein opens a major medical distribution channel. MedTech MedCare adds contract sales coverage. Medical Place is the exclusive reseller into the U.S. Department of Veterans Affairs. An unnamed national healthcare institution also creates a deployment path across its divisions.

Those agreements create access without disclosing purchase volume, utilization, or revenue. DermaSensor says hundreds of provider organizations use the device and more than 45,000 suspicious lesions have been scanned. Those are company-reported adoption figures.

Medical-device scale carries another ledger. The FDA lists an open Class II recall covering 9 of 343 units after a component issue could produce incorrect results or delay referral. DermaSensor notified customers and arranged inspections and corrections. Clinician trust is shaped by the response after a problem appears.

DermaSensor spent years converting Boston University spectroscopy into a handheld product, completing the FDA process, and building clinical evidence. A Class IIb CE mark and NHS feasibility collaboration added an international lane in June. The capital now meets procurement cycles, sales coverage, training, quality systems, and evidence that must survive wider use.

That is the wager inside this extension. The algorithm has to earn a place in a referral decision while the company around it supports every device, clinician, correction, and patient handoff moving through the new channels.

TwoStep Therapeutics

The science behind TwoStep Therapeutics needed 3 Stanford labs before it needed a cap table.

Jennifer Cochran’s lab engineered PIP, a peptide built to recognize 5 tumor-associated integrins. Caitlyn Miller, PhD, connected that scaffold to therapeutic payloads, then led the spinout through Stanford’s Innovative Medicines Accelerator. Carolyn Bertozzi and Ronald Levy added chemical-biology and oncology depth.

That multidisciplinary origin matters because a solid tumor is not a neat row of identical targets. Expression changes across tumor types, patients, and cells inside the same tumor. A therapy aimed at one antigen can be precise and still leave enough target-low cells outside its reach.

TwoStep has closed an oversubscribed $62.5M Series A as the company reports FDA clearance of the IND for TS-104, its first peptide-drug conjugate. Insight Partners and Medical Excellence Capital led the round. M Ventures and Pfizer Ventures joined as co-lead investors, with NFX, 2048 Ventures, Stanford University, and other investors participating.

The company says the financing brings total capital raised to $71.2M. Its 2024 launch followed a $6.5M Seed led by NFX. The latest round is expected to carry TS-104 into a first-in-human Phase 1 trial, advance a radioligand program, and support the next generation of peptide conjugates.

TS-104 links PIP to MMAE, a cytotoxic payload already familiar to oncology drug developers. The unusual part is the address label. PIP is designed to bind 5 integrins through one small, rapidly internalizing peptide, giving TwoStep a way to test whether multi-target recognition can reach across more of a heterogeneous tumor.

The investor group now sits closer to the operating work. Matias Porras Paniagua, John Prufeta, Oliver Lyth, and Rana Al-Hallaq joined the board from the 4 lead firms. They join Caitlyn Miller and NFX’s Omri Amirav-Drory as the company crosses into clinical development.

IND clearance is permission to begin learning in people. It does not supply the answers. Phase 1 will put safety, dose, pharmacokinetics, tolerability, and early antitumor activity ahead of every elegant diagram built in the lab.

TwoStep has assembled capital, pharma-connected investors, clinical leadership, and manufacturing expertise around one demanding handoff. Caitlyn Miller’s team now has to show where TS-104 travels, what patients can tolerate, and whether 5-target recognition gives heterogeneous tumors enough common ground to hold.

Onix

The internet spent 30 years teaching experts a bad deal: publish the work, feed the platform, and hope attribution survives.

AI made that bargain faster. A model can absorb years of books, interviews, clinical frameworks, and public explanation, then answer in a voice that sounds familiar while the person who built the judgment becomes a footnote.

David S. Bennahum and Nicholas Nadeau built Onix around a different exchange. An expert licenses the work, helps shape a dedicated small language model, governs how it operates, and keeps 70% of subscription revenue, according to BetaKit. The user brings the personal context, but Onix says the conversation stays encrypted in a vault on the device instead of becoming a central data asset.

Onix has raised $5M in Pre-Seed funding led by Alpha Edison, with Garage Capital, Ride Home Fund, Jeremy Zimmer, and JS Cournoyer participating. BetaKit reports the round was structured as a convertible note. Onix calls it the company's first institutional capital.

That is a large opening check in a Canadian market where the average 2025 Pre-Seed deal was C$0.88M. The size matters because Onix is solving several problems. It must recruit experts, ingest and evaluate private knowledge, protect user context, make the product useful across weeks of real life, and keep wellness guidance from wandering into medical care.

The company is starting where the stakes expose weak architecture quickly: health and wellness. More than 20 founding experts are live across areas including longevity, nutrition, sleep, stress, and family health. The app is in iOS early access, with commercial availability expected later this year and Android planned to follow.

There is useful friction in the model. Experts set their subscription prices. Users choose whose judgment they want. Smaller models are trained on licensed, specific bodies of work. Onix says it cannot inspect the private conversation that gives the system context.

The funding will accelerate platform development, expert onboarding, research, evaluation, and public availability. BetaKit reports the 10-person team plans to add 4 roles in go-to-market and sales.

The next stretch belongs to the awkward work behind the promise: proving expert fidelity, enforcing the line between guidance and care, earning repeat use, and making consent feel like a better product rather than another settings page. David S. Bennahum, Nicholas Nadeau, and Alpha Edison have financed that work before the market has decided how much trust is worth.

System

A 2-minute medical intake can end with a package at the door. The operational obligation starts where the checkout flow ends.

System is financing that longer job. Axios reports the San Francisco peptide platform raised $12.6M, taking total capital to $20.3M. That is the cleaner accounting than the company release, which describes $20M as new funding. Courtside classifies its investment as Series A, while the release does not name the round.

Axios says Vine led the financing. Patron Fund, Will Ventures, Courtside, Daybreak, SV Angel, and RiverPark Ventures are also in the investor group.

Founder and CEO Adam Steinle is trying to own more of the chain behind the prescription. System pairs patients with U.S.-licensed clinicians, routes appropriate prescriptions to American compounding pharmacies, and ships treatments including NAD+, glutathione, MIC+B12, and sermorelin.

That promise becomes harder as access expands. System says it has a compounding pharmacy in its network, is considering pharmacy acquisitions, and has an active-pharmaceutical-ingredient manufacturer on its roadmap. A consumer brand can acquire demand quickly. Clinical continuity, sourcing, sterility, dosing, fulfillment, and follow-up still have to work on every order.

The early demand is company-reported but substantial. In June, Steinle said System had served roughly 10,000 patients in 4 months and was building a 14,000-square-foot U.S. facility. Revenue, retention, clinical outcomes, valuation, and the acquisition timetable remain undisclosed.

The regulatory context deserves the same precision as the cap table. System’s announcement points to a July vote involving 6 peptides. The FDA’s Pharmacy Compounding Advisory Committee recommended 6 of 7 reviewed substances for the Section 503A bulks list. That vote was advisory and nonbinding. It did not approve those drugs or change the list by itself.

FDA also states that compounded drugs are not FDA-approved and do not receive premarket review for safety, effectiveness, or quality. That does not erase the role compounding can play for an individual patient. It makes System’s infrastructure argument more demanding, because brand trust cannot stand in for clinical and pharmacy evidence.

System now has $20.3M behind the attempt to close that distance. Each new patient turns the brand promise into a clinical decision, a compounded product, a shipment, a follow-up question, and a record that must hold together long after the ad has disappeared.

fundcraft

Private markets can spend months engineering a fund structure, then ask the back office to keep the same truth aligned across spreadsheets, service providers, and regulators. That hidden operating job is what fundcraft is financing.

The Luxembourg-rooted company has secured €12M in strategic growth financing co-led by Riverside Acceleration Capital and CCAP Investments. Existing investors 3VC, MiddleGame Ventures, and Aperture Capital also participated. Fundcraft says the round brings total capital secured since its 2021 founding to €40M.

Co-Founders Julien De Mayer and Olga Porro are building around a deceptively expensive idea: fund administration should operate from one data layer. Onboarding, accounting, compliance, portfolio operations, reporting, banks, depositaries, and auditors all touch the same fund. Every handoff gives the record another place to drift.

Fundcraft reports nearly 300 supported funds, more than 20,000 LP subscriptions, and nearly €20B in commitments and subscription targets. It signed as many new clients in H1 2026 as in all of 2025. Nearly one-third of new mandates came from existing clients.

Those figures are company-reported, not audited service results. RAC already supported Fundcraft's expansion into France, CCAP backs European technology, and the returning investors watched the company complete an €11M Series A in 2024.

The new financing will support additional jurisdictions, institutional private equity buyout capabilities, and more automation and AI across investor, fund, and portfolio operations. Fundcraft says selected workflows already cover onboarding, transfer agency, reporting, waterfall calculations, and cross-border distribution.

This is where the word platform earns a harder definition. A dashboard can look modern while the client still reconciles the work behind it. Fundcraft has to make migrations, regulatory controls, data lineage, exceptions, and human responsibility move together as fund structures become more complex.

Its French entity received AMF authorization in June 2026. Fundcraft says it is already supporting several French launches representing more than €1B in combined commitments and subscription targets. The €12M gives Julien De Mayer, Olga Porro, current COO Christian Senitz, and the team room to carry that model across more jurisdictions without stitching together disconnected systems.

The next client will arrive with a structure, a history, and an exception Fundcraft has not seen before. The record still has to hold.

Cymphony

Sequoia wrote Cymphony's first check before the founders had settled on a product. Two years later, the second check has enterprise customers, revenue, and a 2026 security problem behind it.

Cymphony launched publicly with $30M in total funding. That includes a $25M Series A co-led by Sequoia Capital and SMBC Fin Atlas Beyond Fund, an SMBC-@Fin Capital vehicle, plus a $5M Seed led by Sequoia. TechCrunch places the post-money valuation above $100M.

Co-founders Shy Dekel, Idan Berkovits, and Edi Gotlieb came through Israel's Talpiot program, then worked across Unit 8200, the Office of the Prime Minister, Apple, and the Israeli Ministry of Defense. Shy Dekel is CEO, Idan Berkovits is CPO, and Edi Gotlieb is CTO.

Their product follows a problem enterprise AI inherited from the human workforce. Employees, contractors, service accounts, and applications accumulate permissions across years. An AI agent can borrow those paths, cross systems, search sensitive files, and act before security can explain what it was allowed to reach.

Cymphony is building a workforce graph across people, agents, identities, systems, data, permissions, and activity. It gives security teams one view of access, prioritizes exposures with business context, and can recommend or automate remediation instead of adding another alert to the queue.

The early evidence is company-reported, but it is more useful than a launch-day logo parade. Cymphony told TechCrunch it signed a double-digit number of enterprise customers and reached seven figures in ARR within its first year of sales. Syngenta, KKR, Cass Information Systems, and Athennian are named customers. Sequoia also uses the product.

At one unnamed U.S. public company, Cymphony says it found roughly 85,000 files exposed to AI tools and agents, closed the access gap, and verified that those systems had not reached the files. The customer is unnamed, so the result remains a case study, not an audit. It still shows why permission debt becomes more dangerous when software can search at machine speed.

The Series A will fund product development and expand engineering and go-to-market teams. The harder assignment is commercial: prove agent security deserves a durable budget before Microsoft, Okta, CyberArk, Wiz, Varonis, and the rest of the existing stack absorb it as a feature.

Cymphony now has capital behind the graph and enterprise names behind the pitch. Its next contracts sit where AI ambition meets the access history companies forgot they were carrying.

Lightfield

Keith Peiris and Henri Liriani had the startup version of a champagne problem: 25M users, about $80M raised, and a product they no longer believed could become the company they wanted.

Tome made AI presentations go viral. Forbes reported that revenue still plateaued near $3M as much of the audience stayed on free or low-priced plans. By March 2025, the founders shut it down, cut a roughly 70-person team to six, and started again.

Today, Lightfield announced a $47M Series A led by Andreessen Horowitz. Maverick Capital, Coatue, Audacious, Alumni Ventures, Greylock, and Lightspeed Venture Partners also participated, putting new capital behind a second act that is substantially harder than the first.

Most CRMs know whatever someone remembered to type after the meeting: a stage, a close date, a few notes. That becomes expensive when an AI agent is expected to prepare the next call, draft the follow-up, or explain why the opportunity moved.

Lightfield captures email, calendar, calls, Slack, and LinkedIn activity as it happens. It connects that history across people, accounts, and deals so humans and agents work from the same current picture. The company calls it a world model of the business, with an agent SDK, code sandbox, evals, API, MCP, and CLI.

That architecture is the product thesis. The commercial job is earning permission to become the system of record.

More than 5,000 companies have signed up since Lightfield launched in November 2025, according to the company. Forbes separately reported 1,000 paying customers in July. Those populations differ and are not audited, but the momentum explains why a16z sees an opening before new companies harden around an incumbent.

The capital accounting deserves the same discipline. Tome raised about $80M before the pivot and reportedly still had roughly half available. Lightfield did not disclose a reconciled lifetime total or valuation, so adding every historical round into one number would be cleaner than the real story.

The real story is better anyway. Keith Peiris and Henri Liriani did not pivot because nothing worked. They pivoted after usage worked, attention worked, and the business still pointed somewhere they did not want to spend the next decade.

Now Lightfield has to turn that judgment into infrastructure. Richer context and safer agent execution matter, but each customer migration asks for something code cannot award itself: trust with the memory behind the pipeline.

Sparkle Grooming

A franchise agreement can make a map look crowded long before a customer can book an appointment.

That distance sits inside Sparkle Grooming Co.'s $6M strategic growth financing. The Scottsdale company reports more than 600 franchise licenses awarded in 24 months, with 10 salons open. It expects at least 20 by year-end and plans more than 30 additional openings in 2027.

Selling the territory is the invitation. Opening the salon is where the promise starts collecting receipts.

Ben Crawford, co-founder and CEO, and Joe Aeppli, co-founder and COO, built Sparkle around a sharp observation: grooming is treated as an occasional transaction, but skin, coat, nails, teeth, and ears do not run on an occasional calendar.

Sparkle calls its answer Quick-Service Pet Care: recurring memberships, standardized systems, proprietary technology, and a hospitality-led salon experience. Pet parents get routine care with fewer surprises. Franchise partners get a business designed around repeat visits instead of repeatedly reacquiring the same customer.

That logic is attractive on a development spreadsheet. It becomes expensive in the field. Hundreds of licenses create future obligations: sites, training, recruiting, opening support, quality control, local marketing, technology, and an experience that cannot feel like headquarters sent a PDF and wished everybody luck.

Companion Fund led the financing. Digitalis Ventures manages it in partnership with Mars Petcare, giving the investor side a direct view into animal health, pet services, and the machinery surrounding both.

Digitalis partner Ben Jacobs pointed to a differentiated consumer proposition and a team that understands franchise scale. Sparkle's next phase is less about proving operators will sign agreements than turning those agreements into salons that earn recurring trust.

The capital is meant to strengthen franchise support, expansion, and system discipline. No numbered round, valuation, prior funding total, or participating-investor list was disclosed.

Sparkle is professionalizing a fragmented local service without sanding away the relationship that makes grooming work. Each opening adds another team that must deliver the care, explain the membership, and send the dog home looking like the routine was worth keeping.

Sparkle's next year will move through leases, hiring, training rooms, appointment books, and the first repeat visit in every market the franchise map has promised.

A-Rent

A crew can stand beside a finished electrical system with the schedule already burning and still be one calibrated instrument away from proving it works.

The system can be installed. The schedule still waits for somebody to prove it works safely, and that proof needs the right instrument, current calibration, technical support, and logistics that understand the word urgent.

That operating job has turned A-Rent Test Equipment into more than a rental counter. The company rents electrical test equipment, calibrates customer instruments, and manages fleets for utilities, contractors, manufacturers, data centers, and renewable-energy projects.

Now Bregal Sagemount has made an undisclosed majority growth investment in the business.

Paul Seppanen founded A-Rent in 2019 with backing from Sterling Partners. The company reports more than 1,000 active customers across four distribution locations, with a fifth scheduled for Hillsboro, Oregon, in late 2026.

The growth plan is physical: add test equipment, open distribution locations, expand the load-bank offering used in data-center commissioning, and build the commercial and technology infrastructure that tracks what is calibrated, moving, returning, and ready.

Every asset makes that system more useful and more demanding.

A-Rent says rentals are NIST-traceable, functionally tested, cleaned, and delivered with calibration certificates. Its asset-management service stores customer equipment, tracks maintenance, and prepares it for shipment.

The customer is not renting metal and electronics. The customer is renting fewer reasons for a crew, commissioning window, or energization plan to sit still.

That becomes a larger market as the power economy expands. Berkeley Lab estimates data centers could account for 11.8% of U.S. electricity use by 2030. More power demand creates more systems that must be tested before a facility carries real workloads.

Sagemount's Gene Yoon pointed to A-Rent's availability and service levels. Partner Will Breskman connected the investment to data centers, grid modernization, and electrification. The thesis is clear without a disclosed price: put more calibrated capacity closer to more projects without losing responsiveness.

Fleet expansion is the visible move. Keeping calibration, utilization, logistics, maintenance, and technical service synchronized across that footprint is where the investment earns its name.

Perry Weather

The hardest weather decision is not always when to stop. It is when to let everybody start again.

A coach has athletes waiting. A superintendent has crews burning daylight. A safety manager has a policy and a choice that gets expensive in either direction. Move late and people are exposed. Move early and a site loses time it may not have needed to surrender.

That decision used to live mostly inside somebody's judgment. Perry Weather has raised a $110M Series C growth investment to make more of it repeatable, visible, and documented.

Silversmith Capital Partners led the round, with Arthur Ventures returning after leading Perry Weather's $6.25M Series A and $15M Series B. The 3 announced rounds total at least $131.25M. Silversmith Managing Partner Sri Rao will join the board.

Colin Perry founded the Dallas company in 2013 around a useful distinction: a forecast can describe the weather, but an organization still has to turn its own safety policy into action at the exact place where people work, practice, or play.

Perry Weather puts on-site stations beside that decision. The hardware measures lightning, heat, wind, precipitation, and air quality. The software applies customer thresholds, alerts phones, sirens, PA systems, and strobes, distributes the all-clear, and retains a record.

The company says more than 3,000 organizations use the platform, including schools, cities, manufacturers, professional sports, and 24 of the top 25 contractors on ENR's Top 400 list. Perry Weather also reports that revenue has nearly doubled annually for 6 years. Those company-reported figures explain why growth investors see more than a weather dashboard.

The capital will fund AI capabilities, deepen the use of data from thousands of on-site stations, and support more than 50 Dallas hires through 2027. The AI work still has to prove how much earlier and more precisely customers can see disruption coming. The operating system around it is already clear.

Heat rules and athletic safety requirements are pushing organizations toward better measurement and records. OSHA's federal standard remains proposed, while state and athletic-association rules already create obligations. Compliance may start the conversation, but customers still have to protect people without treating every cloud as an invoice.

That is the commercial handoff Perry Weather is financing: from a policy somebody wrote to a decision an organization can execute, defend, and eventually reverse without losing another hour to uncertainty.

Carputty

Carputty spent its first chapter giving drivers leverage before they walked into a dealership. Its Series B is financing a harder second act: putting that leverage inside the systems dealers, lenders, and automotive platforms already use.

The undisclosed round included funding from an unnamed vehicle technology company. Returning investors TTV Capital, Fontinalis Partners, and Kickstart participated, bringing total capital invested in Carputty to more than $40M.

The missing number matters, but the use of the money says more about the company Patrick Bayliss and Joshua Tatum are building. Carputty will expand loan origination, servicing, and embedded-finance integrations across the automotive transaction.

That is a turn for a business that began with a consumer question: why should every vehicle purchase require a new financing scramble? Flexline provides reusable automotive credit for purchases, refinancing, and lease buyouts. Flexloan, created with Westlake Direct, serves a single vehicle.

The infrastructure has to work for everyone around the buyer. A dealer wants approval without losing the sale. A lender wants disciplined underwriting and servicing. A software provider wants financing inside the workflow before the customer starts wandering.

AutoManager CTO David Aeschbacher says Carputty integrates with several AutoManager products and creates dealer efficiencies. AutoManager supports approximately 5,000 independent dealerships, according to the announcement, giving the round distribution context.

The capital history deserves honest accounting. Carputty raised a $12.3M Series A in 2022, then added $5M of equity alongside a separate $75M warehouse facility from Silicon Valley Bank in 2024. That facility funds lending capacity; it is not venture equity wearing a bigger number for the photo.

The returning investors cover both sides of the market. TTV knows financial infrastructure, Fontinalis works where software meets mobility, and Kickstart has backed Carputty since the early rounds. The unnamed strategic investor adds a signal, but anonymity is not an invitation to guess.

Carputty now has to earn a place inside customer relationships owned by other companies. Every approval, valuation, payment, and service interaction must work well enough that the partner looks smarter for letting Carputty into the room.

The operating ambition becomes visible each time a dealer, lender, or automotive platform makes Carputty part of the transaction rather than another destination outside it.

Botsi

A paywall meets the customer before the company understands them.

That mismatch is expensive. High-intent, price-sensitive, and unsure users can see the same offer because a growth team needs one answer it can ship. The average wins the test. The individuals disappear inside it.

Botsi has raised a $1.5M pre-seed round from Telegraph Ventures, Plain Sight Capital, Atlanta Technology Angels, VentureSouth, and angel investors to build a different pricing system for subscription apps.

Jacob Rushfinn, Jason Schubert, and William Schubert founded Botsi. Jacob Rushfinn is CEO, and Joshua Jarvis is CTO and founding engineer. Their backgrounds span subscription growth, mobile monetization, operations, and production engineering.

Botsi trains a custom model for each app, reads behavioral and contextual signals, and chooses the price, paywall, or offer for an individual user in real time. The model keeps learning as traffic, products, and customer behavior change. That turns pricing from a monthly experiment into infrastructure running inside the product.

Raise one price for everybody and revenue per buyer may climb while price-sensitive customers leave. Lower it for everybody and willing buyers keep money the company might have earned. One winning paywall asks a single answer to carry too many people.

Botsi says customers including Fitness AI, Sleepiest, and Reading.com have recorded revenue and LTV improvements ranging from 15% to 75% since Q4 2025. Its Sleepiest case study reports a 20% subscriber ARPU lift and about $185K in incremental revenue against a live holdout. Those are company-reported results, but the design matters: a different offer must create incremental value, not claim a purchase already coming.

That is also where the harder company starts. Personalized pricing can respond to willingness to pay, geography, intent, and product behavior. It can also become a trust problem when customers cannot understand why prices differ or the model learns faster than the business can govern it.

The investor group is backing Botsi as the product graduates from a revenue experiment into a dependable commercial system. App publishers will expect uplift, clean controls, privacy discipline, and an answer when the model makes a choice the growth team did not anticipate.

Subscription businesses have spent years polishing the paywall. Botsi is working on the decision behind it, where each new customer becomes both a revenue opportunity and another test of whether the system deserves to keep learning.

Clay

Clay's market found the profession before the company finished naming it.

Revenue teams already had data vendors, CRM records, enrichment tools, spreadsheets, outbound systems, and people stitching the gaps together. What they did not have was an operating layer that could remember which decisions produced an outcome and improve the next move.

That is the larger wager inside Clay's $115M Series D.

Wellington Management led the round at a $7.1B valuation. Sequoia Capital, StepStone Group, Andreessen Horowitz's Perennial Venture Fund, Meritech Capital, DST Global, CapitalG, BoxGroup, Boldstart Ventures, Bloomberg Beta, and Evolution Equity Partners also participated.

Clay began as a flexible canvas for data enrichment and prospecting. Teams could connect providers, research accounts, score opportunities, personalize messages, and push the result into the systems where sellers already worked.

Now the company is trying to turn that canvas into a self-learning revenue engine.

The distinction matters. An automation can complete a task. A useful GTM agent has to preserve its sources, expose its reasoning, connect actions to replies and revenue, and carry that learning into the next workflow. The value is not just more activity. It is a revenue system that can explain what worked.

That product shift is arriving alongside a new profession. GTM engineers sit between revenue operations, sales development, data, and automation. They translate a market thesis into systems that can be tested, inspected, and improved.

Clay says its community now includes more than 50,000 people, and it is launching a $1M GTM Engineering Scholarship to help more operators enter the field.

The business has moved quickly enough to make the platform argument credible. Clay reported passing $100M in ARR in December 2025, growing more than 4x during 2025, serving 17,000+ customers, and reaching 80% of the Forbes AI 50.

Those are company-reported figures, but they explain why the round is being priced as infrastructure rather than another outbound tool.

Founders Kareem Amin and Nicolae Rusan started Clay in 2017. Varun Anand joined as a cofounder in 2021. Years of product exploration passed before the market, the profession, and the company converged.

The round sharpens the test: can Clay connect fragmented data, workflows, and outcomes well enough that each campaign teaches the next one something useful?

That is a much bigger job than automating prospecting. It is an attempt to make commercial judgment programmable without making it opaque.

Luminary

A family can spend decades building an estate, then hand its logic to the next adviser as a stack of PDFs.

The documents may be legally precise. The operating record is usually less elegant. One firm knows the investments, another wrote the trusts, another models the taxes, and somebody still has to explain how it all serves the people named inside it.

Luminary raised a $22M Series A to make that handoff less dependent on spreadsheets, document searches, and institutional memory. Ten Coves Capital led the round. BNY joined as a strategic investor, and 8VC returned after leading Luminary's $9.5M 2023 seed financing. Previous investors also participated. Luminary says it has now raised nearly $32M.

David Barnard spent more than 20 years in wealth management, including leading AllianceBernstein's private-wealth business and overseeing more than $100B in client assets. He started Luminary with Joe Lonsdale after watching sophisticated estate work create value while remaining painfully difficult to repeat, update, and show to the client.

Luminary turns estate and ownership documents into structured, source-linked data. The data feeds scenario modeling, beneficiary and tax analysis, exemption tracking, reporting, trust and entity workflows, and collaboration among wealth managers, attorneys, accountants, and trust companies.

The source-linking detail deserves more attention than the AI label. In estate planning, a clean summary that cannot show which provision produced the answer becomes a new risk wearing better typography.

Luminary says its platform supports more than $500B in client assets and names Caprock, IEQ Capital, and Wealth Enhancement Group among its customers. It also serves tax practices, AM Law 100 firms, and trust companies. Those are company-reported markers, but they show how many professional boundaries the product is expected to cross.

The new capital will deepen AI capabilities and integrations, move Luminary further into administration workflows, and expand the commercial team. A plan keeps changing after the client signs it. Families add assets, sell businesses, revise intentions, navigate tax changes, administer trusts, and eventually ask a different set of people to carry it forward.

The $22M buys Luminary room to connect planning with administration. The harder obligation sits inside every future handoff: keeping a family's instructions attached to their sources while advisers, attorneys, accountants, and fiduciaries change around them.