Health Insurance Is Insurtech’s Biggest Boogeyman

Vivek Krishnamurthy is a documentary obsessive — for over a decade he kept a spreadsheet tracking the 400-plus documentaries he’d watched, calling it his “lazy man’s way of podcasting to himself” before documentaries became mainstream pop culture. That same instinct to actually dig into the mechanics of something before forming an opinion shows up directly in how he invests: as a Principal at Commerce Ventures, a nine-year-old, roughly $260 million fund across four vehicles investing at the intersection of retail, financial services, and insurance.

Before Commerce Ventures, Vivek spent several years in investment banking at FT Partners, a fintech- and insurtech-focused shop, where his first-ever deal was Servio (acquired by Solera, a claims fraud detection data business) and where he later worked on SquareTrade’s sale to Allstate. Five years into his time at Commerce Ventures, roughly 65% of the fund’s roughly 100 portfolio companies skew toward technology and infrastructure, with about 35% in full-stack financial services and insurance plays.

In Episode 61 of InsurTechTalk, Vivek and I covered why Commerce Ventures deliberately avoids pure distribution plays, why health insurance is the category he thinks insurtech investors most consistently ignore, and what a close read of Lemonade’s own S-1 metrics actually reveals.

About Vivek Krishnamurthy

Vivek Krishnamurthy is a Principal at Commerce Ventures, a venture capital fund investing across retail, financial services, and insurance. Before joining Commerce Ventures (now in his fifth year there), he worked in investment banking at FT Partners, focused on fintech and insurtech transactions. Commerce Ventures has invested in roughly 100 companies over nine years, including Mulberry, Kin Insurance, Amino, and Precision Gx.

Distribution vs. Product: Where Commerce Ventures Actually Plays

Vivek’s foundational investment framework treats financial services disruption as splitting into two distinct problems: the distribution of a financial product, and the product itself. Pure distribution plays, in his experience, tend to be capital-intensive and brand-driven rather than genuinely dependent on technical or actuarial expertise — real businesses, sometimes very successful ones, but not where Commerce Ventures’ own “insurance nerd” orientation adds distinctive value. The fund instead looks specifically for opportunities where technology and data are the actual reason a company wins — either by identifying and reaching customers others can’t, or by underwriting and servicing them meaningfully better.

That thesis is why the first wave of direct-to-consumer insurtech carriers generally wasn’t a fit for Commerce Ventures. Instead, the fund focused early on genuine infrastructure and SaaS selling into carriers — a notoriously difficult category to build in, given how entrenched policy administration and claims management incumbents are. More recently, as alternative data sources and the ability to underwrite against them have matured, Commerce Ventures has made direct carrier-adjacent investments (MGAs and MGUs) where that same tech-and-data edge applies directly: Mulberry, which sells warranty insurance at the point of sale, and Kin Insurance, which underwrites hard-to-place customers using satellite data instead of relying on zip-code-level proxies.

Health Insurance: The Category Insurtech VCs Keep Skipping

This was the sharpest thesis in the conversation. Vivek argued health insurance is something of a “boogeyman” among insurtech investors — under-discussed relative to P&C, partly because large health carriers function as genuine monopolies in many markets. But Commerce Ventures broke health insurance down by lifecycle the same way it approaches P&C, and found real, technology-driven opportunity specifically because of how much usable data exists in the category. Two portfolio examples: Precision Gx, which uses a proprietary correlation model to flag medical bills submitted to health insurers that were likely billed incorrectly; and Amino, which helps self-administered health plans inside large employers let employees discover, schedule, book, and pay for care through one unified experience.

His explanation for why health is structurally harder than P&C or life: those categories typically involve one dominant relationship (policyholder and insurer, often with a government or lender mandate simplifying the picture), while health insurance involves the employer, the carrier, the provider, and often a separate payment intermediary all at once — a genuinely convoluted, multi-party structure. He offered a memorable framing for the underlying dysfunction: healthcare is a retail experience that tries desperately not to behave like one, even though patients are functionally paying customers of their providers, just usually involuntary ones routed through several intermediary layers.

He also drew a clear distinction between two different responses to that complexity. The first wave of health insurtechs (his examples: Oscar, Bright Health) largely tried to vertically integrate everything themselves — buying clinics, building proprietary EMR systems — essentially building a parallel universe alongside the broken one. Commerce Ventures is more drawn to the alternative: fixing the existing, broken infrastructure directly rather than replacing the entire stack.

Neglected Life Insurance Categories

I raised a related point directly: within life insurance specifically, nearly all digital-first attention has gone to term life, while whole life, universal life, long-term care, and disability remain largely ignored by comparison — not because they’re genuinely harder products to build, in my view, but because the distribution and messaging problem (effectively having to ask “are you afraid of dying, and do you care about your family?”) is harder to solve than the underlying insurance mechanics. Vivek didn’t dispute this, tying it back to the same broader theme: categories get ignored less because of actual technical difficulty and more because of unsolved distribution and communication problems.

Crypto Insurance: A Rare Case of Unforced Demand

Vivek drew a sharp, necessary distinction between two entirely different things people lump together as “crypto insurance”: using blockchain as infrastructure to make the insurance industry itself more efficient (his example: an immutable record of title and coverage history), versus actually insuring crypto assets, which he characterized as fundamentally a form of cyber insurance — since the primary loss driver for crypto holders is theft, not some crypto-specific risk category. He described the underwriting approach he’s seen from existing players in that space as unusually hands-on: carriers hiring cybersecurity specialists to directly assess an insured’s actual operational security posture (device handling practices, risky network usage, and similar), rather than relying on traditional actuarial tables that simply don’t exist yet for this category.

That absence of actuarial history connects to a genuinely difficult structural problem he referenced from something he’d read: in a category with no actuarial tables, responsibly writing new coverage effectively requires being willing to absorb losses long enough to build real loss history — a tough proposition for an early-stage venture-backed company operating on typical seed-stage capital constraints.

What makes crypto genuinely unusual, in Vivek’s framing, is that it’s one of the only insurance categories where real demand already exists without any convincing required — institutional crypto holders are actively asking who they can offload this risk to, a stark contrast to how hard it typically is to sell insurance at all (his blunt framing: even SMB owners who intellectually know cyber insurance matters rarely buy it proactively). The open question isn’t demand — it’s whether the industry can respons­ibly underwrite and deliver the product yet.

His rough market-segmentation prediction: large institutional custodians (his example, Fidelity) will need dedicated large-scale crypto custody coverage relatively soon; individual retail crypto holders mostly won’t need dedicated coverage in the near term, aside from a small segment getting custom policies from carriers like Chubb; and a middle segment — small businesses with blended personal and business crypto exposure — is much further out, realistically only becoming relevant once crypto functions as everyday transactional currency rather than primarily an investment asset, which he estimated could be twenty-plus years away. He also expects a data and exposure-analysis layer to emerge in between (drawing a parallel to Cyence, the cyber risk modeling company later acquired by Guidewire), helping institutions understand crypto exposure they may not even realize they’re carrying, ahead of any insurance product actually being placed.

Where Decentralized Data Actually Matters

Responding to a friend’s skepticism that “anything blockchain can do, a good database can do,” Vivek reframed the real value as decentralized control of data, not the underlying technical mechanism — and pointed to two concrete examples where centralized data control actively harms people today.

The first, which he called merely annoying: title records. Closing on a home requires paying a title company, often around $3,000 and roughly ten added days, essentially just to confirm the seller is the legitimate prior owner — a service that’s expensive less because the verification itself is technically hard, and more because title companies control and monetize access to that data. A transparent, shared property-ownership ledger over time could remove much of that friction and cost.

The second, which he called genuinely harmful: health records. Large EMR vendors (he named Epic specifically) hold patient data that federal law already requires them to make available, but Vivek argued the practical experience of accessing it is often deliberately difficult — his example, receiving hundreds of pages of oddly formatted or even sideways-scanned PDFs rather than clean structured data, creating weeks of delay when a patient moves states or when providers on different EMR systems (his example, a physician on Cerner needing records from a patient on Epic) need to coordinate care.

On solving that specific problem, he laid out two paths: a full “blow it up” decentralized approach, requiring every hospital and stakeholder to buy into a shared, patient-controlled ledger (which he considers a beautiful but extremely difficult idea to actually execute given the sheer number of stakeholders required), versus companies working within the existing rules to build data-normalization layers on top of today’s messy EMR exports. He specifically praised Health Gorilla in that second category, along with Particle Health and Pluto Health, describing a personal example of getting COVID test results delivered near-instantly through Health Gorilla’s pipeline — faster than the testing clinic’s own official channel delivered them.

Reading the Insurtech IPOs: Hippo, Lemonade, and the LTV Question

Asked how the wave of insurtech carrier IPOs and their subsequent stock performance has shaped how Commerce Ventures thinks about the space, Vivek drew a direct parallel to the first generation of fintech lending IPOs (On Deck, Lending Club) — distribution-oriented businesses that didn’t sufficiently focus on unit economics or default rates, and were eventually punished for it. He argued the same dynamic is playing out with insurtech carriers: public markets will tolerate weak loss ratios as long as growth stays extraordinarily high, but give roughly two to three quarters of grace once growth decelerates, before comparing loss ratios unfavorably against traditional carriers and repricing the stock hard. He cited Hippo trading at roughly 10% of its IPO reference price at the time of recording as a direct example of that repricing.

He singled out Lemonade as having navigated this better than peers like Root, attributing it partly to smart acquisitions made using Lemonade’s elevated public multiple to buy assets at lower multiples, and partly to starting with a structurally simpler product (renters insurance) than competitors. But he offered a genuinely skeptical read of one of Lemonade’s own disclosed metrics: the company highlights that 10% of its homeowners policyholder base originated as renters graduating into a homeowners policy — a number that sounds impressive, but which Vivek argued is high mainly because Lemonade’s homeowners book is still quite small overall, meaning the actual renters-to-homeowners conversion rate is closer to roughly 1%. His broader point: as these “graduation model” insurtechs mature, they’ll increasingly face a genuine LTV question, needing to either meaningfully raise customer lifetime value (through acquisitions, as with Lemonade’s Metromile deal) or meaningfully improve their organic cross-sell/graduation rate, or risk simply trading like an ordinary mid-sized carrier over time.

His summary of what earns a real premium from public markets, even relative to a strong incumbent like Progressive: sustained growth, sustainable margins, a demonstrably lower-than-average loss ratio, and stable reinsurance relationships that aren’t at risk of pulling back — all four together, not any single one in isolation.

The VC Funding Paradox: Why the Headline Numbers Mislead

I raised the apparent contradiction directly: public insurtech performance has been rough, yet 2021 saw roughly $14 billion invested into the space, nearly double the prior year, much of it concentrated in large later-stage rounds. Vivek’s response was that the headline figure hides two important adjustments. First, roughly one in every three dollars invested in the space comes from corporate venture arms, whose incentives are strategic and balance-sheet-risk-oriented rather than purely return-driven — his example, a reinsurer like Munich Re being more concerned about whether a portfolio company turned out to be a sound reinsurance risk than about how its IPO ultimately performed financially. Second, of the remaining roughly two-thirds, he estimated perhaps half is concentrated in a small handful — maybe 10% — of capital-hungry, largely first-generation carriers (his example, Oscar’s roughly $1 billion raised), rather than spread broadly across the ecosystem.

His conclusion: the real amount of capital funding genuinely new, early-stage company formation (pre-seed through Series B) is meaningfully smaller than the headline number suggests — which he framed optimistically, as evidence there’s real room and appetite remaining to back a more disciplined “second generation” of insurtech companies that hasn’t fully emerged yet.

Who Should Reach Out

Asked who he wants to hear from, Vivek’s answer was broad but specific in spirit: founders, operators, and other investors who genuinely love insurance — particularly people who are optimistic about the role insurance plays in people’s lives but healthily pessimistic about how hard the path to improving it actually is, a mindset he associates with anyone who’s spent real time working in the category. That extends to operators at carriers, reinsurers, and brokers who enjoy wrestling with genuinely difficult ideas, and to founders considering pivoting into insurance from an entirely different background.

Advice: 14 Peaks

Asked for a closing recommendation, true to form, Vivek pointed to a documentary: 14 Peaks on Netflix, following a Nepalese mountaineer who climbed all fourteen of the world’s peaks above 8,000 meters in seven months, shattering a prior record that had stood at seven years. His takeaway extends well beyond climbing: when you anchor your own ambition to the goals of whoever came before you, you tend to achieve only incremental improvement over them — but a genuine first-principles rethink of what’s actually possible can produce results that dramatically exceed even your own original goal, a lesson he applies directly to how he thinks about venture investing and life more broadly.

Key Takeaways

• Commerce Ventures’ core investment filter distinguishes distribution-driven financial services plays (capital-intensive, brand-oriented, less differentiated by technology) from product-driven ones where technology and data genuinely determine who wins — the fund concentrates on the latter
• Health insurance is framed as insurtech’s most under-discussed major category, with real technology-driven opportunity once broken down by lifecycle the same way P&C is — despite being structurally harder due to its multi-party structure (employer, carrier, provider, payment intermediary) compared to P&C’s simpler single-relationship model
• Crypto insurance is genuinely two different things — blockchain-as-industry-infrastructure versus insuring crypto assets (functionally a form of cyber insurance) — and represents a rare case where real customer demand already exists without needing to be manufactured, unlike most insurance categories
• Decentralized data control (not blockchain technology for its own sake) has real, concrete value in categories like title insurance and health records, where centralized gatekeepers today profit from friction and complexity rather than from genuinely hard underlying technical problems
• A close read of Lemonade’s own disclosed metrics suggests its renters-to-homeowners “graduation” conversion rate is much lower than its headline 10%-of-homeowners-base figure implies, previewing an LTV challenge facing graduation-model insurtechs broadly
• Public markets reward insurtech carriers on four combined traits — sustained growth, sustainable margins, a genuinely lower-than-average loss ratio, and stable reinsurance capacity — rather than any single metric in isolation
• The headline “record” VC funding figures in insurtech overstate genuine early-stage innovation funding once corporate VC’s strategic (non-return-driven) capital and concentration in a small number of capital-hungry later-stage companies are accounted for