Insurtech AI startups captured 95.2% of Q1 2026's $1.63 billion in venture funding, the highest concentration Gallagher Re has recorded, and the money kept landing in the same three functions through the first half: submission intake, underwriting triage, and claims. Sixfold, Shepherd, and Outmarket each closed rounds there. None went to a standalone pricing or reserving engine.

That pattern is not an accident of timing. It is a statement about where venture investors believe defensible margin sits in the insurance stack, and it is a different statement than "AI is eating insurtech." Of the 68 AI-labeled deals Gallagher Re counted in Q1 2026, the average round size hit $25.79 million (Gallagher Re, May 2026), while CB Insights separately put total Q1 deal count at 81 (CB Insights, Q1 2026), which leaves roughly 13 non-AI rounds splitting the remaining $80 million, or about $6.2 million apiece. That is a rough four-to-one size gap between an AI insurtech raise and everything else, computed from two independently sourced totals rather than asserted outright. The gap holds a signal actuarial and pricing leaders should not skip past: investors are paying a premium for workflow automation around the underwriting decision, not for the decision engine itself.

Where the Money Actually Landed

Line up the named Q1 and Q2 2026 raises and a function map appears. Sixfold closed a $30 million Series B on January 29, 2026, led by Brewer Lane Capital with strategic participation from Guidewire, taking total funding past $50 million (Fintech Global, January 2026). The capital funded its AI Underwriter, an agentic system that automates end-to-end underwriting decisioning; it launched in production on June 15, 2026 across six carriers representing $270 billion in gross written premium (The Insurer, June 2026). Shepherd followed with a $42 million Series B on March 24, 2026, led by Intact Private Capital, bringing its total to $67 million (PR Newswire, March 2026). Shepherd insures physical infrastructure for AI and data-center builders, and the round funds autonomous underwriting that cuts quote turnaround from weeks to hours using live construction data feeds; the company has grown revenue more than sevenfold in 24 months across $400 billion of insured project value and 1,500-plus policies (Carrier Management, March 2026). Outmarket AI closed a $17 million Series A on May 13, 2026, led by Permanent Capital Ventures, lifting total funding to $21.7 million (PR Newswire, May 2026); it automates broker-side submission intake, policy comparison, and gap detection, and reports annual recurring revenue up fivefold year over year across more than 250 brokerages (The Insurer, May 2026).

Three raises, three distinct points in the workflow: Sixfold sits at underwriting decisioning, Shepherd at underwriting plus claims-adjacent construction risk monitoring, Outmarket at broker-side intake. Widen the lens to the top of the Q1 2026 league table and the pattern repeats. Corgi's $160 million Series B funded an AI-native carrier writing liability and cyber paper, not a rating engine (Gallagher Re, May 2026). Reserv's $125 million Series C funded claims triage and adjudication at $100 million in annual recurring revenue (Gallagher Re, May 2026). Counterpart's $50 million Series C funded SME cyber underwriting workflow (Gallagher Re, May 2026). The InsurTech.ME weekly tracker's July 5-11, 2026 report shows the concentration extending past H1: Norm AI, an AI governance platform with two insurance carriers already inside its cap table, closed a $120 million Series C at a $1.2 billion valuation (InsurTech.ME, July 2026), crossing the unicorn threshold in under three years.

The Function-by-Function Split

Function Representative H1 2026 raise Amount Actuarial core function present?
Submission intake / broker workflow Outmarket AI (Series A, May 2026) $17M No
Underwriting decisioning Sixfold (Series B, January 2026) $30M Adjacent (uses carrier rating output, does not set it)
Underwriting + risk monitoring Shepherd (Series B, March 2026) $42M No
Claims triage / adjudication Reserv (Series C, Q1 2026) $125M No
AI-native carrier paper Corgi (Series B, Q1 2026) $160M No (writes on its own paper; pricing is internal, not vendor-sold)
Model governance Norm AI (Series C, July 2026) $120M No
Rating / pricing engine None disclosed at comparable scale — N/A
Reserving engine None disclosed at comparable scale — N/A

The absence in the bottom two rows is the actual story, and it is worth being precise about what it does and does not show. It does not mean no capital has ever gone to actuarial pricing software; Akur8 has raised and continues to sell rate-filing intelligence tooling, and EXL has built a substantial AI patent position around insurance infrastructure. What the H1 2026 data shows is that neither appears among the quarter's largest, fastest-growing raises, and no comparable new entrant challenged them. The capital flowing into insurtech in 2026 is chasing surface-level workflow automation, not the rating and reserving math underneath it.

Why Investors Are Avoiding the Pricing Layer

The economics explain the avoidance. Intake, underwriting triage, and claims automation solve throughput problems: a carrier processing 370,000 submissions a year cares intensely about compressing a two-to-four-week underwriting cycle into minutes, because that compression captures business that otherwise lapses to a faster competitor. That is a volume and speed story with an obvious, demonstrable ROI metric investors can underwrite, and it is why Sixfold could point to $270 billion of in-production GWP within months of launch.

Pricing and reserving software sells a different, harder story. A rating engine's value is measured in loss ratio points and reserve adequacy, both of which take years of earned exposure to validate, not quarters. A vendor pitching a reserving model has to convince a chief actuary to trust its output against ASOP-governed sign-off requirements that no outside software vendor can absorb liability for; the actuary of record remains personally accountable regardless of which tool produced the indication. That accountability structure caps how much of the pricing and reserving workflow can plausibly move to a third-party black box, and it correspondingly caps how much venture capital will chase the category. Workflow automation vendors sell speed to a buyer who can act on the metric immediately. Pricing and reserving vendors sell to a buyer who cannot outsource the judgment the software claims to replace.

That divide is exactly why Guidewire's PricingCenter still gets built on top of, rather than replaced by, third-party rating AI, and why the platform incumbents keep absorbing point solutions through acquisition, as Duck Creek did with SEND for agentic underwriting, rather than ceding the pricing core. Sixfold's strategic investor is instructive here: Guidewire took a position in the underwriting-decisioning layer that sits adjacent to, not inside, its own PricingCenter rating engine. That is a platform incumbent hedging the workflow layer while keeping the actuarial core in-house.

The Build Question for Actuarial Core Tooling

For a chief actuary weighing build versus buy on pricing and reserving infrastructure in the second half of 2026, the funding data argues for a split decision rather than a uniform one. Buy the workflow layer: intake, triage, and claims automation are now backed by companies with $50 million-plus in total funding, six-figure GWP deployments, and multi-year runway, the kind of capitalization that made carrier procurement teams nervous about vendor survival in 2022 through 2024. Sixfold, Shepherd, Reserv, and Corgi are no longer fragile enough to disqualify on vendor-risk grounds alone.

Build, or at minimum tightly govern, the pricing and reserving core. No venture-scale competitor emerged in H1 2026 to challenge the position that the data moat around a carrier's own loss experience is what actually prices risk correctly, and the capital markets are implicitly agreeing: nobody is funding a challenger to that moat at Series B scale. MGA submission-intake AI already tests exposure data lineage rather than pricing judgment, which is the workflow-versus-core distinction playing out in a single product category. The rating and reserving engine remains the place where a carrier's proprietary loss data creates a defensible edge no outside vendor can replicate, which is precisely why no outside vendor is trying to raise money to replicate it.

Vendor Concentration Risk Nobody Is Pricing Yet

The flip side of "buy the workflow layer" is that a carrier's underwriting throughput now depends on a small set of venture-funded companies whose survival is not guaranteed past the next funding cycle. Gallagher Re's Q1 2026 data shows 82% of insurtechs that raised in the quarter will not be ready to raise again for at least six months, 17 percentage points longer than the broader venture market (Gallagher Re, May 2026). If Sixfold's AI Underwriter is embedded in a carrier's bind decision for $270 billion of premium and Sixfold hits a funding gap, that is an operational continuity exposure with no obvious internal fallback, because the whole point of buying the workflow layer was to avoid building the fallback.

That risk compounds with model concentration. Foundation model labs are increasingly selling directly to carriers alongside these insurtech vendors, which means a single upstream model provider's outage, price change, or model deprecation can propagate through multiple nominally independent underwriting and claims vendors simultaneously. Vendor-patented AI governance controls and platforms like Norm AI exist precisely because carriers and regulators recognize this concentration risk is accumulating faster than governance frameworks are maturing to address it. An enterprise risk actuary evaluating vendor dependency in 2026 is not asking whether any single insurtech vendor will fail; the 82% re-raise delay rate makes some vendor failure close to a base-rate expectation. The actuarial question is which workflow functions can absorb a vendor's exit without disrupting bind, quote, or claims-payment continuity, and which cannot.

Second-Order Read: Software Over Balance-Sheet Risk

Step back from the function-by-function map and a broader capital-allocation pattern emerges. Every one of the H1 2026 raises examined here, Sixfold, Shepherd, Outmarket, Norm AI, funds software that sits alongside an insurance balance sheet rather than software that takes on underwriting risk itself. Corgi is the partial exception, holding its own carrier license, but even Corgi's headline value proposition is speed-to-quote infrastructure rather than a proprietary pricing algorithm competing on loss-ratio performance. Venture capital is underwriting the thesis that insurance technology margin comes from selling software to risk-bearers, not from bearing risk directly, which is the same logic that has kept most insurtech unicorns of the 2015 to 2021 cohort as MGAs and distribution platforms rather than full-stack risk carriers.

That has a durability implication worth naming plainly. A software vendor's revenue scales with the number of carriers it sells to and is largely insulated from underwriting-cycle losses; a risk-bearing insurtech's revenue is exposed to exactly the loss volatility that sank several 2021-vintage full-stack insurtechs. The 2026 capital concentration into software-layer AI, rather than balance-sheet AI, suggests investors have learned that lesson and are deliberately routing around it. For actuaries, that means the vendors most likely to still exist in three years are the ones selling picks and shovels to the underwriting process, not the ones betting their own capital on loss ratios, and procurement decisions should weight vendor durability accordingly.

Further Reading

Sources