Venture Capital & Startup Funding Roundup, June 23, 2026

It’s Tuesday, June 23, 2026, and the last 12 hours of startup financing did not look like a broad market rebound. They looked like a market making a narrower, harder-edged bet. Capital went to software that sits inside real operating systems, to infrastructure that decides who gets served, monitored, dispatched, or protected, and to a handful of physical-world technologies tied to defense, energy, and medical supply chains. The clearest signals came from Peregrine’s $250 million Series D for government and enterprise operations software, Cadence’s $100 million Series C for chronic-care automation, Stark’s €500 million defense-tech financing, and a cluster of smaller but strategic rounds in fusion, semiconductors, and building-efficiency execution.

That matters because it marks a continued shift away from generic “AI for X” funding and toward systems with tighter claims on budgets, workflows, and outcomes. Investors are showing far more appetite for products that can live near a customer’s cost center or mission center than for software that simply adds another model layer at the edge. Today’s winners were not selling abstraction for its own sake. They were selling dispatch, chronic disease management, secure decision support, drone production, isotope supply, insulation deployment, and power-protection hardware.

The investor mix also stood out. Spark Capital, Andreessen Horowitz, Sequoia, Amplify Partners, Founders Fund, and the NATO Innovation Fund all appeared in rounds whose core argument was not about experimentation but about system importance. That is the through line in today’s market: venture money is still plentiful for companies that can plausibly become infrastructure, while everything else is pushed into a much tougher proof-of-value cycle.

The Macro Environment: Capital Is Chasing Systems, Not Features

The broader backdrop helps explain why today’s deal mix looks so concentrated. Stanford’s 2026 AI Index said global private AI investment reached $344.7 billion in 2025, up 127.5% year over year, while U.S. private AI investment hit $285.9 billion. In other words, the absolute amount of money available to AI startups remains massive, but investors are no longer distributing it evenly. They are funneling it toward businesses that can own a workflow, a regulated process, or a technical choke point.

PitchBook-NVCA’s Q1 2026 Venture Monitor makes the concentration story even sharper. It found that the top five venture deals in Q1 captured nearly three-quarters of total venture investment, that 90.9% of VC dollars went to hub markets, and that 88.8% of Q1 deal value went to AI-related companies. It also noted that public software multiples had fallen to 10-year lows as agentic AI raised real questions about software durability. That is the pressure shaping private rounds now: if software is easy to copy, investors want a reason it will still matter when the model layer gets cheaper.

That is why today’s rounds cluster around three defensibility models. The first is embedded workflow ownership, as seen in Cadence, Probook, Attention, and Peregrine. The second is physical or sovereign capacity, as seen in Stark, Astral Systems, VARM, and AlpSemi. The third is trusted data infrastructure, which is the heart of Allium’s pitch. These are not identical businesses, but they all argue for durable budget authority rather than incidental usage.

Investor behavior is also getting more top-heavy. PitchBook-NVCA said six managers accounted for 76.2% of fund capital raised in Q1 2026. That concentration matters because the firms writing the biggest checks now have more power to define what “venture-scale” looks like. Today’s lineup reflects that power: firms are rewarding companies that can plausibly become category control points, even in markets that used to look too operational, too regulated, or too industrial for classic software investing.

The Funding Rounds

Stark raises €500 million in funding to scale sovereign European defense manufacturing

Berlin-based drone startup Stark Defence has raised €500 million in a funding round led by Sequoia and Founders Fund, lifting its valuation above €3.5 billion just two years after launch. Stark was the largest strategic round in the group, not only because of its size but also because of what it says about the direction of European venture capital. The Berlin defense-tech startup makes weaponized drones and plans to use the new financing to build electronic-warfare research facilities, scale production, and push deeper into what it describes as sovereign defense capability. More than 80% of the capital is earmarked for R&D and manufacturing, which is exactly the opposite of the asset-light software playbook that dominated the last cycle.

Investors care because Stark sits at the intersection of three forces: Europe’s defense rearmament, venture capital’s renewed comfort with dual-use technology, and the realization that “sovereign tech” now includes factories as much as code. The company’s cap table underscores that point. Sequoia and Founders Fund are involved, but so are the NATO Innovation Fund, Project A, and other backers willing to underwrite manufacturing scale as part of a geopolitical thesis. The valuation is high enough to show genuine scarcity value, though reported figures differ: Tech.eu said €3.5 billion, while the Financial Times reported €3.2 billion. That discrepancy does not change the signal. A Europe-based drone maker is now being financed like a strategic platform company.

Funding Details
Startup: Stark
Investors: Sequoia, Founders Fund, NATO Innovation Fund, Project A, Air Street Capital, 201 Ventures, Advent, Döpfner Capital
Amount Raised: €500 million
Total Raised: Not disclosed in today’s announcement
Funding Stage: Undisclosed venture round
Funding Date: June 23, 2026
Headquarters: Berlin, Germany
Sector: Defense tech and autonomous systems

Peregrine Technologies raises $250 million to push AI deeper into government and high-stakes operations

Peregrine’s Series D is one of the day’s cleanest examples of where AI funding still feels rational at scale. The company helps governments and other complex organizations unify siloed data inside a permission-aware operational system, then use that context to support decisions in public safety, emergency response, and related use cases. This round values the company at $6.8 billion and expands a business that has moved from state and local agencies into federal, enterprise, and international markets.

The investor case is not hard to see. Unlike many AI startups, Peregrine does not need to invent demand. Governments already run fragmented data systems, and they already pay heavily when those systems fail to talk to each other. The product is sticky because it sits inside environments where governance, auditability, and access rules are not nice-to-haves. They are the product. The round was led by existing investors, including Fifth Down Capital, Sequoia Capital, OG Venture Partners, Goldcrest Capital, XYZ Ventures, and Godfrey Capital, and it follows Peregrine’s $190 million Series C at a $2.5 billion valuation in March 2025. That valuation jump shows how aggressively investors are repricing AI companies that have moved from pilot mode to mission-critical software.

Funding Details
Startup: Peregrine Technologies
Investors: Fifth Down Capital, Sequoia Capital, OG Venture Partners, Goldcrest Capital, XYZ Ventures, Godfrey Capital
Amount Raised: $250 million
Total Raised: Not disclosed in today’s announcement
Funding Stage: Series D
Funding Date: June 23, 2026
Headquarters: San Francisco, United States
Sector: Government software, data infrastructure, and enterprise AI

Cadence raises $100 million in funding to automate chronic care with regulated AI

Cadence’s $100 million Series C is one of the strongest examples of healthcare AI moving from pilot theater into scaled operating infrastructure. Cadence manages chronic disease care for older adults through a model that blends supervised AI agents, remote monitoring, EMR integration, and direct health-system partnerships. The company says it now works with more than 20 health systems, treats more than 100,000 active patients, tripled annual recurring revenue in 2025, and saves Medicare roughly $2.7 million every week. Spark Capital led the round, with participation from Thrive Capital, General Catalyst, Coatue, B Capital, Corewell Health Ventures, Memorial Hermann, and Duke Health.

Why investors care is straightforward: Cadence is not selling AI as a productivity overlay for clinicians. It is trying to become part of the care-delivery stack itself. That gives it a better shot at durable reimbursement and deeper integration, but it also subjects the company to more scrutiny. Stat reported that the round values Cadence at $1.23 billion, and that tension is important. In healthcare, companies that promise labor leverage and cost reduction can command premium valuations, but only if they survive policy, billing, and quality tests. Cadence’s financing says investors believe the company has crossed enough of those thresholds to justify more aggressive expansion.

Cadence also embodies a bigger venture pattern. The company launched in 2021 with $41 million, raised a $100 million Series B later that year, and has now added another $100 million, implying roughly $241 million in total disclosed funding. That is the kind of financing progression investors reward when a startup can tie AI directly to lower cost of care, better outcomes, and a large payer-backed market.

Funding Details
Startup: Cadence
Investors: Spark Capital, Thrive Capital, General Catalyst, Coatue, B Capital, Corewell Health Ventures, Memorial Hermann, Duke Health
Amount Raised: $100 million
Total Raised: Approximately $241 million disclosed
Funding Stage: Series C
Funding Date: June 23, 2026
Headquarters: New York, United States
Sector: Healthtech and clinical AI

Allium raises $40 million in funding to become the data layer for onchain finance

Allium’s $40 million Series B says something important about where blockchain funding has settled. This is not a retail-crypto momentum bet. It is an infrastructure bet on the idea that financial institutions will keep moving payments, issuance, and analytics onto blockchain rails, and that someone has to clean, standardize, and serve the underlying data. Amplify Partners led the round, with Kleiner Perkins, Theory Ventures, and Pruven Capital participating.

The company’s pitch has become more persuasive as blockchain has moved from speculation toward narrower institutional uses. Allium says it supports data across more than 150 chains, and it points to use by organizations including Visa and Boston Consulting Group; The Block reported that clients include Visa, Stripe, Coinbase, and a16z crypto, and that the Federal Reserve has cited its data. That kind of customer and citation base matters because it suggests Allium is not fighting for crypto-native mindshare alone. It is trying to become middle-office infrastructure for institutions that need trusted data before they can deploy capital or build products.

There is also a practical venture lesson here. Allium raised a $16.5 million Series A in July 2024, which brought disclosed funding to $21.5 million at the time. Today’s Series B lifts total disclosed funding to roughly $61.5 million. In a sector that has seen painful valuation resets and consolidation, that is a sign that investors still have an appetite for picks-and-shovels businesses tied to institutional adoption rather than token volatility.

Funding Details
Startup: Allium
Investors: Amplify Partners, Kleiner Perkins, Theory Ventures, Pruven Capital
Amount Raised: $40 million
Total Raised: Approximately $61.5 million disclosed
Funding Stage: Series B
Funding Date: June 23, 2026
Headquarters: New York, United States
Sector: Blockchain data infrastructure and financial data plumbing

Probook raises $40 million in funding to own dispatch in home services

Probook’s announcement is smaller in absolute dollars than the day’s megadeals, but it may be one of the sharpest reads on where vertical AI investing has gone. The company announced $40 million in funding, comprising a $34 million Series A led by Andreessen Horowitz and a previously completed $6 million seed round led by Sequoia Capital. Sequoia also joined the Series A. Rather than chasing the crowded market for voice agents and top-of-funnel automation, Probook built around dispatch, the decision engine that determines who goes where, when, and with what economic result.

That matters because dispatch is not a feature. It is a control point. If a startup can own the scheduling and matching layer inside HVAC, plumbing, and electrical businesses, it can expand outward into intake, customer messaging, and workflow automation with far better data and much stronger retention. a16z’s own investment memo makes that case explicitly, and the company says it already runs across hundreds of locations nationwide. This is classic vertical software logic updated for the agent era: don’t bolt AI onto the edge of a workflow, own the part the rest of the workflow bends around.

Funding Details
Startup: Probook
Investors: Andreessen Horowitz, Sequoia Capital
Amount Raised: $40 million announced, including a $34 million Series A and a previously completed $6 million seed
Total Raised: $40 million disclosed
Funding Stage: Series A
Funding Date: June 23, 2026
Headquarters: New York, United States
Sector: Vertical AI and home-services software

Attention raises $30 million in funding to move sales AI from observation to execution

Attention’s $30 million Series B fits the same pattern as Probook, but in enterprise go-to-market software. The company says most AI sales tools still watch calls, summarize notes, and update dashboards. Attention wants to act inside the workflow instead: drafting and sending follow-ups, updating the CRM, and running the next play. RTP Global led the round, with Aglaé Ventures, Eniac, Alven, Linea Ventures, and customer angels participating.

Investors are responding to a specific wedge in the market. Observation-only tooling is getting cheaper and more interchangeable as large models improve. Execution layers, by contrast, can become more valuable because they hold the workflow, the business logic, and the closed-loop data on which actions actually drive revenue. Attention says it is now running more than 20 million agent actions per month, serving more than 500 customers, and has grown ARR 4x year over year. Whether those numbers endure will depend on enterprise retention and outcome quality, but the financing makes the current investor thesis easy to read: the next category leaders in sales AI will be the companies that do the work, not the ones that merely describe it.

Attention had previously raised a $14 million Series A in 2024, after a $3.1 million seed round reported by Business Insider, implying about $47.1 million in total disclosed funding. That is enough capital to push further upmarket, where CRM quality, forecasting accuracy, and workflow compliance matter more than demo appeal.

Funding Details
Startup: Attention
Investors: RTP Global, Aglaé Ventures, Eniac, Alven, Linea Ventures, customer angels
Amount Raised: $30 million
Total Raised: Approximately $47.1 million disclosed
Funding Stage: Series B
Funding Date: June 23, 2026
Headquarters: New York, United States
Sector: Enterprise AI and revenue operations

Astral Systems raises £23 million in funding to bring fusion economics into the medical supply chain

Astral Systems is the kind of deeptech round that tells you investors still want exposure to hard science, but only when the commercialization path is clear. The Bristol company raised a £23 million first close led by Mercia Ventures, with Tees River, Daphni, Blast Club, Speedinvest, and Playfair involved. Rather than positioning fusion strictly as a long-horizon power story, Astral is using its multi-state fusion technology to target medical radioisotopes, where shortages already affect diagnostics and treatment.

That near-term wedge is why the company matters. Fusion startups often have no credible bridge between laboratory promise and present-day revenue. Astral is trying to cross that gap by addressing a medical and industrial supply chain problem that customers already understand. Tech.eu reported that the company has operational reactors, plans to bring critical isotopes to market by early 2027, and has generated more than £3 million in revenue through research contracts. This is the sort of hardtech profile venture firms increasingly prefer: technically ambitious, yes, but anchored in a nearer revenue model than “grid-scale energy, eventually.”

Funding Details
Startup: Astral Systems
Investors: Mercia Ventures, Tees River, Daphni, Blast Club, Speedinvest, Playfair
Amount Raised: £23 million
Total Raised: More than £28 million disclosed
Funding Stage: Series A first close
Funding Date: June 23, 2026
Headquarters: Bristol, United Kingdom
Sector: Fusion, biotech infrastructure, and industrial deeptech

Rapalogix Health raises $20 million in funding to turn longevity science into a dermatology business

Rapalogix Health’s $20 million Series A sits at the junction of biotech, longevity, and consumer-adjacent clinical care. The California company is building around skin health, but the underlying claim is broader: that aging biology can be translated into prescription therapeutics and science-led professional skincare products. The round was led by Deerfield Management, with participation from WovenEarth Ventures, leads from Apple Tree Partners and Vasama Capital, and a group of healthcare executives and physicians.

Why investors care is less about skincare per se than about the commercial strategy. Longevity science has often struggled to find products that are scientifically legible, commercially sellable, and not trapped in the supplements market. Rapalogix is trying to use dermatology as a wedge where customers already pay, physicians already exist, and clinical claims can potentially be tied to visible outcomes. That is still a risky route, but it is a more practical one than building a company around generalized anti-aging rhetoric.

Funding Details
Startup: Rapalogix Health
Investors: Deerfield Management, WovenEarth Ventures, representatives from Apple Tree Partners and Vasama Capital, healthcare executives, and physicians
Amount Raised: $20 million
Total Raised: Not disclosed in today’s announcement
Funding Stage: Series A
Funding Date: June 23, 2026
Headquarters: Carlsbad, United States
Sector: Biotech and longevity-focused dermatology

VARM raises €17.5 million to industrialize home insulation, not just talk about it

VARM’s Series A is one of the day’s most founder-useful rounds because it shows what climate-tech investors are rewarding right now. The Berlin company did not raise for a new material or a novel heat-tech stack. It rose to execute insulation at scale. ABN AMRO Sustainable Impact Fund led, GET Fund co-led, and Aurum Impact joined existing backers Emerge, Partners, and Pale Blue Dot.

The thesis is blunt: Europe already knows it needs more efficient buildings, but the bottleneck is workforce and execution, not awareness. VARM’s answer is a workforce model that trains career changers into insulation installers, runs quality control in-house, and offers fixed-price installation. That may sound operationally heavy for a venture. It is. But it also sits closer to measurable savings and policy targets than many climate software pitches. In a market that increasingly wants proof that climate companies can turn adoption into throughput, VARM is offering a model built around installation capacity rather than consumer persuasion.

Funding Details
Startup: VARM
Investors: ABN AMRO Sustainable Impact Fund, GET Fund, Aurum Impact, Emerge Partners, Pale Blue Dot
Amount Raised: €17.5 million
Total Raised: Not disclosed in today’s announcement
Funding Stage: Series A
Funding Date: June 23, 2026
Headquarters: Berlin, Germany
Sector: Climate tech and building efficiency

AlpSemi raises €17 million in funding to build power protection for AI-era electrical systems

AlpSemi may be the most quietly strategic round of the day. The French semiconductor company raised €17 million to industrialize power switches for solid-state circuit breakers, with Yotta Capital leading and SE Ventures, Navitas Semiconductor, and Cycle Group participating. The company is targeting buildings, industrial systems, and 800V direct-current AI data centers.

This round matters because it sits under a problem most AI app investors prefer not to think about: electricity constraints. As AI workloads continue to drive data-center demand, the winning infrastructure companies will not be just model hosts and chip designers. They will also include firms making the electrical systems more controllable, efficient, and fault-tolerant. AlpSemi’s bet is that power protection is moving from electromechanical hardware toward semiconductor-based control. If that thesis holds, the company is not just a component supplier. It is a beneficiary of AI’s knock-on demand for smarter power architecture.

Funding Details
Startup: AlpSemi
Investors: Yotta Capital, SE Ventures, Navitas Semiconductor, Cycle Group
Amount Raised: €17 million
Total Raised: Not disclosed in today’s announcement
Funding Stage: Undisclosed
Funding Date: June 23, 2026
Headquarters: Grenoble, France
Sector: Semiconductors and power infrastructure

What Today’s Funding Activity Reveals

The biggest pattern is that investors are paying for ownership of decision loops. Peregrine owns operational context. Cadence owns patient monitoring and intervention logic between visits. Probook owns dispatch. Attention owns next-action execution inside revenue teams. Allium owns standardized onchain data for institutional workflows. In each case, the startup is closest to the moment where software changes an outcome, not just a dashboard.

The second pattern is that physical systems are back in force, but only where there is a sharp revenue or policy wedge. Stark is selling into Europe’s defense rearmament. Astral is attacking isotope shortages through fusion. VARM is solving the install-capacity problem behind home efficiency. AlpSemi is building the electrical protection layer that AI-era infrastructure will need. Investors are still selective about hardware and deep tech, but they are willing to back it when the problem is large, current, and tied to real bottlenecks.

The third pattern is that “AI” by itself no longer explains why a company gets funded. Virtually every software company in this roundup uses AI somewhere in the pitch, but the financing logic is now more specific: embedded workflow, proprietary data position, regulated deployment, or infrastructure leverage. That shift is consistent with the wider market, where PitchBook-NVCA reported extraordinary concentration in AI financing and declining public software multiples simultaneously. Venture is still paying up for AI, but mostly when AI sits inside something far harder to replace.

Comparative Funding Table

Strategic Takeaways for Founders and Investors

Founders should take one lesson from this batch of financings: the easiest AI story to finance is no longer “we use models.” It is “we own the workflow where money, risk, or service quality actually moves.” Startups that can point to dispatch, billing, compliance, uptime, patient outcomes, electrical efficiency, or defense readiness have a better chance of escaping feature-level competition and procurement fatigue. That is true whether the product is software, hardware, or a hybrid execution model like VARM’s.

Investors, meanwhile, are signaling that defensibility now comes from one of three places: system-of-record adjacency, physical capacity, or trusted data. The rounds for Peregrine, Cadence, Probook, and Attention reward software that lives deep inside a customer operation. The rounds for Stark, Astral, VARM, and AlpSemi reward physical execution in markets where capacity itself has become strategic. Allium shows that data normalization can still command capital when the buyer is institutional and the use case is operational.

There is also a pricing signal here. Capital remains abundant for companies that can argue they are core infrastructure, but is far less forgiving for businesses whose AI layer is easy to reproduce. PitchBook-NVCA’s data on concentration, faster AI deal cycles, and pressure on public software multiples suggest the market will continue to reward category-control stories while punishing undifferentiated AI wrappers. For founders, that means distribution, compliance, deployment complexity, and data rights matter as much as model quality. For investors, it means the next premium companies may look more operational and less glamorous than the last cycle’s poster children.

Conclusion

Today’s funding activity points to a venture market that is still willing to move fast and pay up, but only when the startup sits close to a real bottleneck. In software, that means embedded systems that can own the next action. In hard tech, it means capacity that governments, hospitals, utilities, and industrial operators cannot easily source elsewhere. That is why a defense drone maker, a chronic-care automation company, a government decision platform, and an insulation operator can all belong in the same roundup without feeling mismatched.

If there is a single market insight to carry forward from this set of deals, it is that venture capital is becoming more literal about infrastructure. The winners are increasingly the companies that can make a workflow run, a supply chain move, a building consume less power, a hospital manage more patients, or a government act faster with better information. Money is still flowing to AI. It is just flowing to AI that can prove it belongs somewhere, and it is hard to remove.

Open Questions and Limitations

A few data points remain incomplete. Several companies did not disclose total funding or current valuation, so those fields are marked as not disclosed rather than estimated. Stark’s valuation was reported differently by Tech.eu and the Financial Times, so both figures are noted above rather than harmonized. And one round, AlpSemi’s, sits right on the 12-hour boundary based on the source’s “12 hours ago” timestamp rather than an explicit clock time.

Large financings that were still in the news cycle today but were announced earlier were excluded. The clearest example is Baseten’s $1.5 billion Series F, which was still being covered on June 23 but whose official announcement was published on June 22, outside this roundup’s 12-hour window.