The Agreements We Never Meant to Make

Skan AI Raises $63M Series C for Enterprise AI Agent Context

Skan AI just raised $63M in Series C funding to attack a problem hiding in plain sight: enterprise AI cannot automate work it does not understand at scale.

Cathay Innovation and Dell Technologies Capital co-led the round, with Citi Ventures, Bloomberg Beta, State Farm Ventures, and Wipro Ventures participating. The financing brings Skan AI's total funding to approximately $120M, following a $14M Series A and $40M Series B.

That capital stack says plenty about the market. The backers are not betting on another assistant that writes polite emails and calls it transformation. They are betting on the messy sequence of people, systems, decisions, exceptions, and handoffs that makes a real business run.

Skan AI observes how work actually happens across enterprise applications, then turns those interactions into process intelligence and a Context Graph of Work. The point is not to document the official procedure. It is to capture the version employees execute after the policy deck meets reality.

Agents trained on static instructions inherit static blind spots. A process can look clean in a diagram and become unrecognizable once it crosses six systems, four teams, a compliance check, and the judgment call of someone who has done the job for 12 years. Skan AI is building the layer that gives automation access to that operational memory.

Its platform connects Skan AI Blueprint, Skan Intelligence, and Skan AI Agents. Blueprint maps AI opportunities, Intelligence exposes how work performs, and AI Agents uses observed context to execute with enterprise rules and guardrails.

Skan AI reports more than $500M in identified customer value and more than 300% year-over-year growth for two consecutive years. Those are company-reported figures, not audited public results, but the direction is hard to miss: buyers are moving from AI demonstrations toward systems that can survive production.

Avinash Misra, CEO, and Manish Garg, COO, have worked this problem since co-founding Skan AI in 2018 after building Endeavor Software, which Genpact acquired in 2015. Their thesis came from years inside transformation projects where documented processes and real work rarely match.

Skan AI, Avinash Misra, Manish Garg, and the team now have $63M more behind that thesis. The sharper takeaway is this: model intelligence gets attention, but operational context earns trust. For agents making decisions inside regulated workflows, understanding how work really moves is the foundation.

Blacksmith Raises $45M Series B to Scale AI Code Validation

Blacksmith just raised a $45M Series B led by Peak XV Partners, with Y Combinator and GV tripling down, at a $550M valuation. The round was raised in March and announced today, an infrastructure-company way to handle news: close the money, keep the machines running, tell everybody later.

AI made writing code dramatically cheaper, then handed engineering teams a fresh bill for proving that code works. Blacksmith says CI jobs on its platform have grown 5–10% week over week since the start of 2026. More than 6,000 companies now run CI through Blacksmith, including Supabase, Clerk, Ashby, and Mercury. The code arrives faster. Validation still has to clock in.

Blacksmith was founded in 2024 by Aditya Jayaprakash, Aditya Maru, and Aayush Shah, three University of Waterloo graduates with experience at Faire, Cockroach Labs, and Superblocks. Jayaprakash is CEO; Maru and Shah are co-CTOs. They did not start with an AI agent looking for a problem. They started with the unglamorous machinery every software company needs and few want to operate.

That machinery is a CI cloud built around purpose-built hardware, bare-metal compute, ephemeral VMs, caching, observability, and runners designed to make GitHub Actions faster and less expensive. Blacksmith reports more than 30,000 users and claims up to 2x faster runtimes than GitHub-hosted runners. Those are company-reported numbers, but the signal is clear: when software production accelerates, every slow test, flaky check, and overloaded queue gets expensive.

Now Blacksmith is moving up the stack. Its [code]smith agent can build features, fix bugs, diagnose CI failures, and automatically address failed checks. The company is also building [code]smith QA to test changes before they merge. The valuable layer may not be code generation alone. It may be the system that decides whether machine-made code survives contact with production.

Peak XV is betting that Blacksmith can own that validation layer. GV and Y Combinator returning after Blacksmith's $10M Series A and $3.5M seed round adds another signal: this is conviction compounding, not a new investor touring the factory floor.

The money is going primarily into compute. Blacksmith says it manages hundreds of thousands of cores and plans to expand that footprint tenfold while hiring across New York City and San Francisco. Aditya Jayaprakash, Aditya Maru, Aayush Shah, and the Blacksmith team forged a business around the part of software development nobody brags about, right before AI made it impossible to ignore.

Alta Smiles Closes Seed Extension for Hidden Orthodontics

Alta Smiles just closed a Series Seed extension, and the most interesting number is the one the company did not disclose.

The round amount is private. The strategy is not. Run Ventures led, Stokes Stevenson McMaster Health Ventures participated, and Dental Innovation Alliance returned after Alta Smiles' first institutional financing in 2024. The capital will support commercialization, market awareness, and platform expansion.

That platform starts with a simple truth: people want straighter teeth without visible hardware or a daily compliance assignment. C5 Hidden Orthodontics uses patented nickel-titanium wire bonded behind the teeth, without brackets. Unlike removable clear aligners, it does not depend on patients remembering to wear trays. The mouth is already complicated enough without giving it homework.

Alta Smiles is not building around one clever wire. The Fort Washington, Pennsylvania, company also offers Opti-AI for treatment planning and indirect bonding, plus V5x Dental Images for in-office photography and remote monitoring. Dental adoption is rarely won by novelty alone. Providers need a system they can integrate and explain without turning every case into an operational science project.

Founder and CEO Gary Giegerich now has more than capital around the table. Dan McMaster, a former 3M Health Care executive, and Gene Dorff, a former Dentsply Sirona executive and current CEO of Solmetex, joined the board. Those appointments show the phase: technology must translate into distribution, clinical support, and commercialization.

Run Ventures is underwriting that transition, while Stokes Stevenson McMaster Health Ventures and Dental Innovation Alliance add oral-care relationships and operating context. For an early dental technology company, that network can matter as much as the check. An elegant product can still disappear if providers are not trained and confident enough to offer it.

Alta Smiles is aiming at the gap between traditional braces and removable aligners: a concealed, provider-managed treatment designed around actual patient behavior. It is a narrow wedge, but good healthcare companies often start with a specific frustration and keep solving outward.

The next proof will come from broader provider adoption, repeatable clinical delivery, and whether C5 can turn early patient and clinician interest into a durable category. Gary Giegerich, the Alta Smiles team, Run Ventures, Stokes Stevenson McMaster Health Ventures, and Dental Innovation Alliance are backing technology patients are not supposed to see.

Silicon Data Raises $30.5M to Price AI Compute

Silicon Data just raised a $30.5M Series A to put a price tag on the part of AI everybody treats like weather: compute.

Valor Atreides AI Fund led the round, with ME Group, DRW, F-Prime, Samsung, VanEck, Further, Jump, Tectonic, Wintermute, Breed, Hack, Blank VC, Sancus Ventures, and SoGal Ventures participating. With its $4.7M seed, Silicon Data has disclosed $35.2M across two rounds.

Founder and CEO Carmen Li built Silicon Data around a blunt problem. GPU capacity can cost wildly different amounts across providers, regions, hardware generations, and contract terms, yet buyers have been expected to negotiate as if the market came with a clean sticker price. It did not. The bill arrived anyway.

Silicon Data is building the data layer underneath that market. Its platform spans daily GPU rental-price indices, SiliconMark benchmarking, SiliconPriceIQ predictive pricing, carbon analysis, token economics, a GPU Forward Curve, and SiteIQ for underwriting AI data-center capacity. The company says its indices cover 95% of neo-cloud GPU providers and all major hyperscalers.

AI infrastructure is moving from engineering expense to financial exposure. In May, CME Group and Silicon Data announced plans for compute futures based on Silicon Data's daily GPU benchmarks, pending regulatory review. If those contracts launch, cloud providers, AI builders, traders, and financial institutions gain something the market has been missing: a shared reference point for pricing and hedging compute risk.

The Series A syndicate reads like a clue. Trading firms, asset managers, strategic technology companies, and venture investors are backing the same thesis: compute is becoming a market that needs price discovery, not another invoice nobody can explain. DRW and Jump returned after co-leading the seed, while Wintermute, Sancus Ventures, and SoGal Ventures also came back for the next round.

Carmen Li reported in June that Silicon Data had grown revenue 5x since January 2026. That is a company-reported figure, but it fits the larger signal. Silicon Data is no longer only describing an opaque market. It is building the benchmarks, forward curves, and risk tools that could make it legible.

Carmen Li and the Silicon Data team are pursuing a useful business thesis: not every messy market needs more software, but the company defining the measurement standard can become infrastructure before the rest of the industry realizes it has been using guesswork as a pricing model.

Infinimmune Raises $75M for Human-First Antibody Trials

Infinimmune just closed a $75M Series A, and the interesting part is not the size of the check. It is what that check is being asked to prove.

Regeneron Ventures and Playground Global co-led the round. RA Capital Management, Goldcrest Capital, Forge Life Science Partners, Godfrey Capital, and Wild Tree Ventures joined returning investors including Merck Global Health Innovation Fund, Pear VC, Everbright Biofund, and Civilization Ventures. That is a serious syndicate around a company still facing the difficult transition from preclinical science to human data.

Infinimmune starts with antibodies found in the human immune system. Its Anthrobody platform screens millions of single memory B cells for natively paired, fully human antibodies, while its GLIMPSE language models help select and optimize candidates. The idea is simple to understand and hard to execute: begin with human biology, then engineer it into medicine without losing what made it useful.

The first clinical targets are in atopic dermatitis. IFX-101 targets IL-22, a pathway with no approved biologic, while IFX-201 targets IL-13. Both programs are preclinical, and Infinimmune expects both to enter first-in-human studies in 2027.

Atopic dermatitis is not waiting for somebody to discover inflammation. It has effective drugs, familiar targets, and room for better efficacy, convenience, and durability. Infinimmune is arguing that a different starting point can produce a better medicine.

In March, Infinimmune signed a multi-target antibody discovery collaboration with Merck carrying an undisclosed upfront payment and up to approximately $838M in potential milestones. That figure is contingent, not cash in the bank, but it shows pharmaceutical interest before the wholly owned pipeline reaches the clinic.

The leadership fits the assignment. Co-founders Wyatt McDonnell, Katie Pfeiffer, and Mike Gibbons serve as CEO, CTO, and CSO, respectively. Srikanth Pendyala leads clinical and regulatory strategy as CMO. Discovery built the company. Clinical execution now has to earn the next chapter.

The $75M will push IFX-101 and IFX-201 toward human studies, advance additional programs, and keep the Anthrobody engine moving. No valuation was disclosed. Infinimmune has disclosed at least $87M across its seed and Series A, but capital is only permission to run the experiment.

The Infinimmune team, its founders, and investors now have more runway behind a clear thesis. The round funds the transition. The clinic will decide whether human-first antibody discovery becomes a durable advantage.

Neros Raises $250M to Scale Autonomous Drones

Neros Technologies just raised a $250M Series C at a $2.5B post-money valuation, because modern warfare has turned the cheap drone from a tactical accessory into industrial policy with propellers.

Sequoia Capital and American Strategic Technology Fund co-led the round. Interlagos, Valor Equity Partners, Allen & Company, Thiel Capital, Spark Capital, and Dylan Field participated. Add the company's prior disclosed financing, and Neros has now raised more than $370M since 2023.

That capital is not going toward a shinier version of the same aircraft. Co-founders Soren Monroe-Anderson, CEO, and Olaf Hichwa, CTO, are expanding Neros from one high-volume FPV platform into a broader family of low-cost autonomous and interceptor systems.

Archer AI adds terminal guidance and GPS-denied position hold to Neros's FPV architecture. Bandit is being developed as a counter-UAS interceptor for Class 2 and 3 threats, including Shahed-style systems. Both are being designed with the hardware and compute needed for coordinated multi-asset control.

Neros says it plans to put Archer AI and Bandit into combat theaters by the end of 2026. That is a target, not a victory lap. The company still has to convert capital into reliable hardware, production volume, training, sustainment, and systems that work when the environment refuses to cooperate.

In June, the U.S. Army awarded Neros an IDIQ contract with a $500M ceiling for Archer systems and related support. A ceiling is not guaranteed revenue, but it creates a serious procurement lane for a company founded by two former competitive drone racers three years ago. It also gives the next product generation a manufacturing base instead of a PowerPoint promise.

The signal inside the round is larger. Western defense has spent decades buying a small number of exquisite systems at exquisite prices. The drone era keeps presenting a ruder equation: capability has to survive contact with volume, supply chains, electronic warfare, and replacement cost.

Neros is betting that vertical integration can make that equation work across several missions. Manual FPV strike, autonomy-enhanced strike, counter-drone interception, and multi-asset coordination can share components, manufacturing discipline, and lessons from the field.

Soren Monroe-Anderson, Olaf Hichwa, and the Neros team are building around a business lesson bigger than defense. When the market changes the unit economics of the problem, the winning company rarely sells the old answer faster. It rebuilds the factory, product family, and cost curve around the new reality.

CodeRabbit Raises $143M Series C for Agentic Change Management

CodeRabbit just raised $143M in a Series C at a $1.5B valuation, and the timing says more than the number. AI made code abundant. It did not make judgment abundant. Somebody still has to decide what should ship, what might break, and which pull request deserves a human brain before production.

Atomico and Smash Capital co-led the round. BMW i Ventures, Datadog, Hirtle Callaghan, SineWave Ventures, Scenic Management, CRV, Scale Venture Partners, Flex Capital, Pelion Venture Partners, Harmony Partners, and Engineering Capital also participated. With $88M raised through its Series B, CodeRabbit has now pulled in roughly $231M.

That capital is not chasing a prettier review bot. CodeRabbit is moving into Agentic Change Management: a control layer that validates changes, prioritizes attention, explains the intent and risk inside massive pull requests, and monitors code after it ships.

Coding agents can generate thousands of lines before a team decides whether the work is valuable, ready, or connected to the priority. The backlog is moving from tickets to code. Implementation got cheap. Informed attention did not.

Harjot Gill and Gur Singh saw that pressure early. CodeRabbit says it handles more than 2M code reviews each week, serves over 17,000 customers, and supports 150,000 open-source projects. More generated code creates more review volume, security exposure, and demand for a neutral quality gate.

The platform organizes that approach into three parts. Triage scores changes by value, urgency, risk, dependencies, readiness, and reviewer fit. Change Stack turns sprawling diffs into an explanation of architecture and impact. CodeRabbit Security carries the same context beyond merge to find reachable security and business-logic risks.

The distinction is independence. The tools writing code increasingly want to review it too. CodeRabbit is betting teams will want a separate system that questions outputs, preserves evidence, applies organizational standards, and keeps humans responsible for intent, architecture, product judgment, and acceptable risk.

The Series C will fund international expansion and deeper investment in research, infrastructure, and product development. CodeRabbit also plans to invest more than $10M over the next 12 months to keep AI code review and agent capabilities free for open-source maintainers.

Harjot Gill, Gur Singh, and the CodeRabbit team now have $143M more behind the thesis, with Atomico and Smash Capital backing judgment at scale. If agents can create change faster than organizations can understand it, the company that makes change legible may own a much larger piece of the development stack.

Bridge to Life Raises $110M Series C for VitaSmart Expansion

Bridge to Life just raised $110M in Series C equity and debt financing, and the timing says more than the headline. This is commercialization capital for a product that crossed the hardest line in medtech: moving from clinical promise to an FDA-authorized system transplant teams can adopt.

Soleus Capital led the equity financing. Lauxera Capital Partners joined alongside Bridge to Life directors, officers, and employees. Soleus Capital Credit Opportunities Fund provided the debt. The split and valuation were not disclosed.

Bridge to Life will use the capital to expand the field teams behind VitaSmart, fund U.S. transplant-center deployment, advance an investigational organ-viability tool, broaden its multi-organ pipeline, and scale globally. Part of the financing refinanced the Perceptive credit facility, reducing leverage and its interest rate.

Growth rounds are usually presented like the entire check is headed for a product roadmap and a victory lap. Here, management is funding expansion while cleaning up the capital structure. Less cinematic, more disciplined.

VitaSmart received FDA De Novo classification in January 2026 for hypothermic oxygenated perfusion of specified donor livers after static cold storage and before transplantation. Bridge to Life says the early U.S. rollout has produced master service agreements and repeat orders. Those commercial figures are preliminary and unaudited, but the direction is clear: the next phase is adoption, not invention alone.

More than 100,000 people are waiting for organ transplants in the United States. Every preservation decision carries clinical and operational weight. Every viable organ lost is a problem nobody can explain away with a prettier pitch deck.

President and CEO Don Webber is betting Bridge to Life can turn an established preservation franchise into a broader perfusion platform. Founder and Chairman Steve Schweighardt built the company from preservation products into machine perfusion. CFO Brian Lenz now has a financing structure designed to support that expansion.

Soleus Capital partner Ben Lund framed the thesis around an established franchise, FDA-classified perfusion, field adoption, and a credible pipeline. Lauxera co-founding partner Samuel Levy brings organ-transplantation experience. These investors know the clinical terrain.

Bridge to Life, Don Webber, Steve Schweighardt, Brian Lenz, Soleus Capital, and Lauxera Capital Partners are now entering the harder phase: making better organ preservation ordinary enough that the technology disappears into standard practice while more patients make it home.

Gravity Raises $30.5M Series A for AI Advertising

Gravity just raised a $30.5M Series A to build the advertising rails for a world where the buyer might be a person, a chatbot, or a software agent with permission to spend. Lightspeed Venture Partners and Committed Capital co-led the round, bringing the San Francisco startup's total funding to $38.5M.

Co-founders Zach Oldham, CEO, and Leo Martinez, CTO, are building the full stack: tools for advertisers to buy placement, tools for AI platforms to monetize conversations, and the exchange connecting both sides. Gravity also handles contextual matching, auctions, creative, measurement, conversion tracking, and publisher payouts.

The execution is less simple. An ad inside an AI conversation cannot arrive two seconds late, miss the user's intent, or feel like a pop-up wearing a lab coat. Gravity's system reads recent context, retrieves campaigns, runs an auction, and generates creative for that moment.

Gravity says it serves millions of ads per day, with placements across AI products including ChatGPT, Codebuff, EcoGPT, Daimon, Magneta, Runable, and CTO.new, and advertisers including Best Buy, Target, Vercel, and MongoDB. Those names show access, not proof of scale. Gravity has not disclosed revenue or campaign economics.

The stranger, more important move is agent-to-agent advertising. Gravity is testing a model where advertisers feed product catalogs, features, and current promotions to software agents making recommendations. The next planned step is payment integration, allowing an agent to complete a purchase after the user approves it.

That puts discovery, ranking, attribution, and checkout into one commercial loop. It also raises the question every serious operator should ask: when an agent recommends a product, where does helpful context end and paid influence begin?

Google, OpenAI, and established adtech companies did not forget advertising exists. Gravity is betting that an independent network connecting fragmented AI inventory with advertiser demand can learn faster across both sides than any single application can alone.

Prior seed backers Caffeinated Capital and Basis Set Ventures saw the opening early. Lightspeed Venture Partners and Committed Capital are now backing Gravity's attempt to turn that opening into durable market structure.

Zach Oldham, Leo Martinez, and the Gravity team are building around a business lesson bigger than the funding: when a new interface changes how decisions get made, the valuable company may be the one building the market underneath it.

Chatham Financial Announces Orogen, Atairos and GIC Deal

Chatham Financial just put a fresh piece of evidence on the table: independence and outside capital do not have to be sworn enemies.

The Kennett Square capital-markets advisor announced a definitive strategic investment agreement with Vikram S. Pandit and The Orogen Group, Atairos, and GIC. The amount and valuation were not disclosed, and the transaction has not closed. It is expected to close in late Q4 2026 or early Q1 2027, subject to customary conditions and regulatory approvals.

This is not a conventional venture round dressed up in expensive vocabulary. It is a long-term partnership with a 35-year-old, employee-owned firm that says it serves more than 4,500 companies across 12 offices and handles $2T in annual transaction volume.

Chatham Financial will remain majority-owned by its employees. Matt Henry continues as CEO. Amol Dhargalkar continues as Chairman. Vikram S. Pandit, Chairman and CEO of The Orogen Group, is set to join the board. The structure says plenty: bring in serious institutional partners, keep control close to the people doing the work, and add board-level experience from somebody who has lived through the sharp end of global finance.

The investor lineup is unusually coherent. The Orogen Group focuses on long-term strategic investments in financial services. Atairos describes its approach as patient partnership. GIC manages Singapore's foreign reserves. Different institutions, same preference: durable businesses with room to compound.

Chatham has spent decades working in the plumbing that makes capital markets function: assets, debt, derivatives, hedging, valuation, reporting, and advisory. That work becomes painfully important when rates move, currencies swing, debt stacks get complicated, or somebody discovers that risk management was not supposed to be a decorative slide.

The timing also carries weight. Chatham recently completed its acquisition of Hodes Weill & Associates and launched Chatham Onyx, its next-generation technology platform. The new capital arrives as the firm expands its reach across the capital stack while trying to preserve the independence that made the franchise valuable in the first place.

Mike Bontrager, Matt Henry, Amol Dhargalkar, and the Chatham Financial team have built a company capable of attracting The Orogen Group, Atairos, and GIC without selling away its center of gravity. The real test comes after closing: whether long-term capital can accelerate Chatham's next chapter without sanding down the employee-owned culture that got it here.

Mindgard Raises $30M in Series A Funding

Mindgard just raised a $30M Series A, and the money is landing where enterprise AI has been weakest: the gap between moving fast and knowing what models, agents, and applications can do when somebody attacks them.

Album VC led the round, with Karma Ventures, @.406 Ventures, Atlantic Bridge, IQ Capital, and Lakestar joining. It follows Mindgard’s $8M round disclosed in January 2025, when .406 Ventures backed its U.S. push.

AI security is graduating from a research problem to an operational budget line. Companies are moving AI systems into production while the attack surface changes beneath them. Prompt injection, agentic manipulation, model theft, unsafe tool access, and hidden integrations do not wait for the next quarterly review.

Mindgard attacks the problem from the offensive side. Its platform maps AI exposure, automates red teaming, assesses exploitable vulnerabilities, and supports runtime protection across models, agents, and applications. The point is not another dashboard telling a CISO to worry harder. It is evidence of how an AI system can fail before an attacker turns the lesson into an incident.

CEO James Brear is leading the company’s scale-up from Boston and London. Co-founder and Chief Science Officer Dr. Peter Garraghan brings more than a decade of AI security research from Lancaster University, work that became Mindgard’s attack library and shaped its attacker-aligned approach.

Mindgard says its platform is used by Fortune 500 security teams, and a March 2026 product announcement reported more than 80 publicly disclosed vulnerabilities identified across leading AI technologies in 90 days. Those are company-reported metrics, but they show the pressure enterprise buyers are managing.

The investor mix offers another useful signal. Album VC and Karma Ventures are joining investors who already knew the company, including .406 Ventures, Atlantic Bridge, IQ Capital, and Lakestar. New conviction is useful. Returning conviction after product and leadership changes is harder to fake.

AI adoption creates value and exposure in the same motion. Companies that secure models only after deployment are treating production like a laboratory with customers inside it. Mindgard is betting the market will pay to move adversarial testing earlier, automate it, and keep it running as systems change.

James Brear, Dr. Peter Garraghan, the Mindgard team, and its investors now have $30M behind the next phase. The funding is the headline. The deeper signal is that securing AI is becoming part of building AI, not an apology written after the breach.