The most valuable machine at Intelligence Ventures is not an algorithm. It is an inbox. Every week, another 20 to 30 founders send the Lakewood Ranch, Florida firm their plans for applying artificial intelligence to healthcare. Some want to read an electrocardiogram. Others want to watch a patient deteriorate without a cuff or cable, guide a physician inside an MRI scanner, or catch liver disease from a blood test. The pitches arrive with familiar promises and very different odds. Founder Doug Nissinoff and his small team read them anyway.
By July 2026, Intelligence Ventures said it had logged 828 applications and read 850 decks. For a fund writing initial checks of $25,000 to $50,000, that is a disproportionate amount of homework. It is also the strategy. Each deck is one more coordinate in a private map of early-stage AI healthcare: where founders are clustering, what valuations they expect, which claims repeat and which problems remain curiously untouched.
The firm calls the public version of that map the State of AI Healthcare. The annual report has become part research paper, part calling card and part sourcing funnel. Founders download it, recognize their neighborhood on the map and apply. More applications produce more data. The data sharpens the report. Then the fund invests behind what it thinks it has learned.
The $50,000 research department
Intelligence Ventures sits in a peculiar corner of venture capital. It is too small to win by outbidding large funds and too specialized to spray checks across whatever category is fashionable. Its answer is to narrow the frame. The company invests only at pre-seed and seed, only in the United States and Canada, and only across four verticals: biotech and pharma tools, diagnostics, medical devices and healthcare software. It does not invest directly in therapeutics.
This is not a product catalog dressed up as a thesis. The four categories come with opinions. Biotech tools offer deep moats but long timelines. Medical devices reward founders willing to endure hardware and regulatory complexity. Diagnostics is described as underbuilt relative to its market. Healthcare SaaS is the largest portion of the pipeline and the most crowded. A general workflow assistant has to do more than save clicks to earn attention.
“We didn't build a fund. We built a map. The fund is what we do with it.”Intelligence Ventures' investor thesis
That map is the real competitive claim. Any investor can say it knows healthcare. A large generalist can hire a specialist or subscribe to the same industry databases. Intelligence Ventures is betting that raw, very early deal flow reveals something those databases miss. A pre-seed pitch can show a market forming before revenue, regulatory filings or a priced round create a visible trail. Read enough pitches and repetition becomes information.
Own the data, or own a wrapper
The firm's decisive question is simple: who owns the data? In consumer software, a better interface or lower price can produce an opening. In healthcare AI, interfaces are easy to mimic and foundation models are widely available. Durable value, Intelligence Ventures argues, comes from proprietary clinical, genomic, imaging or operational data that continually improves the product. Remove the AI and the product should stop working. Remove access to the company's unique data and a competitor should not be able to rebuild it overnight.
Percentages come from the firm's February 2026 newsletter and are rounded. The imbalance explains its public fascination with diagnostics: the smallest slice may contain the least competed-for problems.
The filter is useful even when it rejects a company. A clinician using a generic model to summarize notes may have a helpful feature, but not an investment-grade moat. A diagnostic system that generates a growing archive of labeled signals while it works may be harder to copy. The distinction also forces an uncomfortable founder conversation: is AI the mechanism of the product, or the newest sentence in the pitch?
Five bets, five data loops
The publicly listed portfolio is compact enough to see the thesis at work. AGED Diagnostics is developing AI-powered blood tests for liver disease. HEARTio analyzes electrocardiograms for cardiac conditions. Atapir uses computer vision to monitor vital signs without contact. Directed Systems models physiology to anticipate instability in surgery and critical care. Medical Devices Corner is building a teleoperated robotic system for precise work inside an MRI scanner.
Blood-based liver-disease detection and monitoring.
AI interpretation of ECG signals at the point of care.
Contactless computer-vision monitoring for patient deterioration.
Physiological modeling for surgical and critical-care decisions.
Teleoperated robotics for MRI-guided cancer procedures.
These businesses sell to clinicians, hospitals, laboratories and life-sciences teams rather than to the fund itself, but they share a commercial problem: healthcare adoption is slow, evidence-heavy and regulated. A clever model is not enough. Buyers need clinical validity, workflow fit, reimbursement logic, security and a reason to change behavior. Intelligence Ventures' team combines biotechnology, pharmacy, device, regulatory and operating backgrounds, then extends that bench through its limited partners.
Limited partners with homework
The Executive LP Committee is the firm's most revealing cultural artifact. Once a month, participating limited partners receive decks and investment memos, then spend two hours grading four startups on a structured rubric. The group includes physicians, scientists, healthcare executives and operators. Participation is optional, but the design turns some investors into an auxiliary diligence and advisory team.
It also changes what the fund is selling. Accredited investors can put in a stated minimum of $25,000 and remain hands-off, or take a closer look at the machinery. Founders, meanwhile, gain access to people who may understand a clinical buyer, a regulatory pathway or a technical hire. The business model remains conventional venture capital - management of LP commitments and potential returns from portfolio equity - but the network is arranged to do practical work.
That same instinct now appears in the IV Accelerator. The 10-week program accepts ten pre-seed or seed companies, writes a $25,000 check on day one, opens with three days in Tampa and ends with a demo day in New York. Weeks two through nine are virtual. Rather than march every founder through the same startup syllabus, the team studies each company's previous fund application and diligence feedback, then builds sessions around the gaps. Regulatory strategy, fundraising, marketing and go-to-market support are pulled in as needed.
“Scientists first. Investors second.”The accelerator's compact promise
The accelerator is also an admission that promising companies do not always arrive ready for a check. Intelligence Ventures says more than 700 startups had applied by the time the program was announced, while fewer than one percent had been funded. An accelerator creates a place for the near misses. It gives the fund ten weeks of observation, gives founders a small amount of capital and gives the LP network something concrete to coach.
Where the map can fail
A pile of inbound decks is not the market itself. It is a sample shaped by the firm's audience, geography and application form. The loudest categories may simply be the easiest to pitch. The quietest may be empty because they are difficult, capital-intensive or commercially unattractive. Proprietary data can create a moat, but it can also create consent, privacy and interoperability problems. In medicine, rare data is not automatically good data.
Small checks bring another constraint. A $25,000 investment cannot finance a clinical trial or manufacture a medical device. Intelligence Ventures has to earn influence with speed, judgment and connections, then rely on other investors for the larger rounds. Its more than 350 co-investor relationships are therefore not a decorative statistic. They are part of the delivery system.
The firm is young. Nissinoff began shaping it in February 2023, Florida records show the LLC filed that May, and Fund I formally launched in May 2025. The team lists five portfolio companies, no exits and no public performance record. What it has built so far is an interesting information business wrapped around a micro-fund: a report that attracts attention, an application funnel that creates data, a committee that lends expertise and an accelerator that develops the companies just outside the portfolio.
A small fund's useful lesson
For founders, the practical value is clarity. A company can know before applying whether it fits: AI must be central, the business must operate in the United States or Canada, the round must be pre-seed or seed, the product must sit inside one of four verticals and therapeutics are out. That sharp boundary saves everyone time. For accredited investors, the offer is focused exposure plus the option to participate in diligence. For anyone studying AI healthcare, the public report is a free view into a market normally hidden behind private pitch decks.
The broader lesson is not to imitate the branding or publish another generic trend report. It is to turn the routine exhaust of a business into a useful asset. Intelligence Ventures had to read applications to invest. By structuring what it saw, comparing it over time and publishing selected findings, the firm made the reading itself valuable. The map may or may not produce exceptional returns; that takes years to know. Today, it already gives a tiny Florida fund a reason for founders and investors to keep sending coordinates.