← back to ranking

ACGL · Arch Capital Group Ltd.

Insurance - Diversified · mkt cap $30.6B · calls: Q1 FY2026 vs Q4 FY2025
43.0 conviction · conf-adj 43

conf –

enthusiasm:18.0 · trend:8 · quantifies:0 · impact:0 · under_radar:14 · credibility:0 · business_impact:4 · disruption:0 · commitment:-4 · confirmation:3

Enthusiasm latest 6 / prev 4 (rising)

Arch frames AI as internal ops/underwriting enablement (10+ years of ML, recent MCE migration code/testing acceleration) and as a cyber systemic-risk amplifier, not a growth or margin lever they size. Enthusiasm rose call-over-call with more concrete examples, but management still says results are hard to prove, AI is early innings, and they cite no P&L-linked AI metrics—credible on risk and integration work, thin on measurable business impact.

PAST (realized)
CURRENT (now)
FORWARD (guidance)
TRACK RECORD — PROMISE vs DELIVERY

—/100 (no quantified promises)   no-quantified-promises  6 calls reviewed

Across six calls (Q4 FY2024–Q1 FY2026) Arch cited data/analytics and risk-based pricing only qualitatively; no quantified, time-bound AI/ML/automation target was set. The only explicit AI mention (Q1 FY2026) was retrospective credit for finishing a mid-market systems migration in 18 months—not a prior measurable promise to track.

PRICED-IN (REFINED)
LOW (room left)

Est. revisions flat  ·  Fwd P/E 6.0  ·  EV/Sales 1.6x

AI claim maps to Insurance Segment, Reinsurance Segment, Mortgage Segment

Analyst rating counts are unchanged month-over-month (3 strong buy, 7 buy, 9 hold) with no upward migration, and forward EPS steps down from ~14.6 to ~8.8 then only modestly recovers to ~9.4 while revenue grows mid-single to low-double digits—revision momentum is flat, not rising. Valuation is not stretched: ~6x forward P/E (on the provided next-FY EPS), ~6.5x TTM P/E, ~1.6x EV/Sales, and price ~87.6 sits below ~103 average targets, so the market is not paying a rich multiple for AI-driven upside. AI underwriting, pricing, and claims efficiency would most plausibly flow through Insurance and Reinsurance (and mortgage analytics in Mortgage), but flat estimates on a cheap multiple imply that thesis is not yet baked in.
COVERAGE — ENTHUSIASM TRAJECTORY + CATALYSTS
3Q4 FY20243Q1 FY20253Q2 FY20253Q3 FY20254Q4 FY20255Q1 FY2026

AI enthusiasm across 6 calls — trend ↗ rising

Data-and-analytics underwriting talk dominated; Q1 FY2026 added explicit AI for MidCorp systems migration only.

RECENT AI CATALYSTS & NEWS
BUSINESS IMPACT - QUALITATIVE MATERIALITY

5/10 qualitative impact   moderate  medium-term · mixed evidence

Where AI matters: underwriting productivity & systems integration

Arch has decade-long ML in mortgage/P&C plus credible ops wins (MCE migration testing/code acceleration, underwriter triage tools) but management calls AI early innings, cannot tie results to P&L, and cites no quantified revenue or margin uplift.

Caveats: No disclosed AI ROI or expense-ratio targets; Cyber AI may raise frequency/scale of systemic losses; Efficiency gains may compress margins via competitive pricing; Scaled AI still gated by data and systems integration

AI DISRUPTION / CANNIBALIZATION RISK  tailwind · 3/10

Specialty underwriting judgment and risk selection remain hard to replicate quickly; AI mainly pressures cyber tail/systemic loss scenarios and industry-wide pricing pass-through of efficiency, not automating away Arch's core insurance/reinsurance product.

OPTIONS / MARKET STRUCTURE

option liquidity: fair

ATM IV
—
TYPICAL BID-ASK
—
OPEN INTEREST
—

proxy inputs — dollar-ADV $189M · beta 0.33 · px $87.62

source: proxy (no options chain on FMP)
FMP /stable/ exposes no options-chain endpoint on this key, so ATM IV, bid-ask spread and open interest are unavailable. Liquidity below is a PROXY from dollar-ADV, beta and price level (a stand-in for option depth), not measured option-market data.

CONFIRMATION — INSIDERS · 13F · LANGUAGE
Mixed — insiders selling, institutions adding, management language 3/10 hedged.
INSIDERS selling 8 open-market sell(s) vs 1 buy(s) — net distribution
INSTITUTIONS (13F) adding as of 2026-03-31: 115 new / 122 closed positions; 436 increased / 351 reduced; institutional ownership -1.95pp; -9 net 13F holders
MGMT LANGUAGE 3/10 hedged AI mentioned once, retrospectively on migration; no metrics, rollout, or forward AI plan.
commit “also represents a strong use case for artificial intelligence in accelerating systems and platform transformation.”
VERBATIM AI QUOTES
“Earlier this month, our team successfully completed the data and system migration of the acquired businesses from Allianz to Arch on systems. The ability to complete this effort in just 18 months speaks not only to the dedication of our teams but also represents a strong use case for artificial intelligence in accelerating systems and platform transformation.”
— Nicolas Alain Papadopoulo, Q1 FY2026
“In terms of the recent AI, Anthropic Mythos, we see it as a real current threat. But we don't really see it's changing the cyber product. I think we see the cyber product as more of a the cyber market as more of an arm race between attacker and defender. And certainly, Mythos is accelerating that trend. But Mythos can help the attacker, but the defender can also reinforce in deference using the same model. So we think it's really an acceleration of the speed at which maybe cyber attacks can be conducted. And it's -- and to your point, it's also an acceleration of the scale. So I would think because the scale would be larger I would think we see it more as an increased systemic risk. So we are taking a very careful approach to that in our RDS scenarios, so.”
— Nicolas Alain Papadopoulo, Q1 FY2026
“we are focusing on building new tools to really help our underwriters with battery selection, triage and so on that will make them more productive.”
— Nicolas Alain Papadopoulo, Q1 FY2026
“In terms of AI, I mean, I don't -- I mean, it's certainly something that is coming at us really quickly, really fast. We're trying to think of ways where we can kind of, again, automate things, and we're doing some of that, but I think there's -- it's still very early innings, very early days of that. So I think we'll -- that will evolve, and we'll see where it goes.”
— François Morin, Q1 FY2026
“we've been investing in AI for the last 10 years, both in mortgage and P&C. So we've deployed a bunch of AI and machine learning models and -- but it's changing really fast. I think the industry in our struggle is really to really show results while at the same time, working on our data strategy and our integration of our system to really support AI at scale. And third, really figure out what AI would look like 3 years from now because it's changing so quickly. If you look at the Entropic model, they open huge opportunities to do certain things, but what's next. So I think you really have to take -- and it's a lot of investment. At the same time, you're trying to create productivity and the insight for your underwriters to be able to compete.”
— Nicolas Alain Papadopoulo, Q1 FY2026
“the way it really help us and speed up the process is to write some of the codes. I think we old didn't do it out there. But when we did, it was really helpful. And the big help was on the testing. A lot of the testing was done by AI, and that really accelerated the time to market.”
— Nicolas Alain Papadopoulo, Q1 FY2026
“that's where I think to Nicolas' point, the AI kind of capabilities really came through and helped speed up the process.”
— François Morin, Q1 FY2026
“Running scenario to make -- is running scenarios to make sure that every time you create -- we created a new platform to a good point, Francois for context. And so every time you create a new software you have a lot of testing that to make sure that the software is doing what it's supposed to do. And a lot of it today can be done through AI as opposed to individuals going in and asking the underwriter to test, the guys that collect the cash to test that -- what they answers get to the right places and so on.”
— Nicolas Alain Papadopoulo, Q1 FY2026
“we think of AI as more of an opportunity for efficiency rather than a threat. But ultimately, the beneficiary of AI will be the consumers as most of the savings and efficiency will be passed on to the insurer. So -- but yes, I think the advantage of being in the specialty market is it's complex. I think it will -- I'm not saying it's impossible, but it will take time for models to learn, to replicate the behavior of the underwriters. So I think what we're seeing is personal lines or SME may be happening there faster than in the space that we are playing.”
— Nicolas Alain Papadopoulo, Q4 FY2025
ANALYST QUESTIONS ON AI
Q (Q4 FY2025, Cave Montazeri (Deutsche Bank)): Given yesterday's move in the market, I was going to ask you about the risk of disruption to your business model from AI and whether you're more likely to be a net beneficiary from AI via improved efficiencies and smaller risk selection rather than at risk of disruption, which I suspect is probably more limited to some distribution platforms or maybe the carriers lines are more commoditized.
A: Nicolas Papadopoulo agreed with the premise: AI is more an efficiency opportunity than a threat; savings likely pass to consumers/insurers; specialty complexity means models will take time to replicate underwriter behavior; personal lines/SME may adopt faster than Arch's markets.
Q (Q1 FY2026, Cave Montazeri (Deutsche Bank)): Structurally, like given the recent developments in AI and the potential for cyber attacks to become more frequent and more destructive, does that change your view of tail risk, aggregation risk or even the long-term insurability of the product?
A: Nicolas Papadopoulo: cyber ~3:00 p.m. on the underwriting clock; recent AI (referenced as Anthropic Mythos) is a real current threat but not changing the cyber product; attacker–defender arms race accelerated; AI increases attack speed and scale → higher systemic risk; careful approach in RDS scenarios.
Q (Q1 FY2026, Taylor Scott (Barclays)): Is lagging into an artificial intelligence investment and doing it that way to try to achieve growth something that you think is achievable?
A: François Morin: AI is moving very fast; automating some work but still very early innings. Nicolas Papadopoulo: investing in AI/ML for ~10 years in mortgage and P&C; deployed models; industry struggles to show results while building data strategy and systems integration for AI at scale; hard to forecast AI in 3 years; heavy investment aimed at underwriter productivity and competitive insight.
Q (Q1 FY2026, Matthew Heimermann (Citi)): Follow up on your call related to using AI in the technology rollover of mid-corp. And just curious how that experience has been different than past.
A: Nicolas Papadopoulo: AI helped write code and especially accelerated testing/time to market. François Morin: greenfield platform build after MCE acquisition (no inherited systems) is where AI capabilities sped the process. Nicolas: AI runs test scenarios vs manual underwriter/ops testing.