New Ask Claude or ChatGPT about any stock, using Bullpen. Bullpen in Claude or ChatGPT. 7 days free

Friday, November 13, 2026

Back to the front page

The Spy Who Shagged The Market

Nvidia’s leadership scandal puts its AI-market role under scrutiny, with comparisons to AIG’s pre-GFC liquidity influence.

From X

Fajasy on X

Root's $ROOT annualized premium run rate (premiums in force) fell 5.4% y/y in Q3, nearly double Q2's 2.8% drop, as larger carriers cut prices and Root gave up direct-to-consumer growth to protect returns. Policy growth offset the lower premium per policy in Q1, but policies in force were flat from June at 484k (+3.8% y/y) while premium per policy fell 8.9% y/y, to $1,441. $ROOT's gross accident period loss ratio (losses from crashes in the quarter over premium earned) rose 4.6 points y/y, to 61.1%, but only 1.6 points from Q3 2025's first estimate of 59.5%, which $ROOT has since revised down to 56.5%. The 10-Q blames (1) higher repair and medical costs per claim, (2) more frequent claims, and (3) a shift in channel mix, so claims costs are rising while premium per policy falls, though 61.1% is still inside $ROOT's 60-65% target range. The net combined ratio (losses plus expenses over premium; under 100% is an underwriting profit) improved 4.9 points y/y, to 97.2%, but ~4.2 points of that is Carvana warrant expense falling $15.3M after a one-time catch-up in Q3 2025, so the underlying gain was under a point. $ROOT's Q3 letter to shareholders says it repurchased $36.2M of stock "during the third quarter," but the 10-Q shows $15.5M in Q3, down from over $20M in Q2; the $36.2M is the total since the May authorization, with $38.8M left as of September 30. $ROOT also began deploying its new pricing model in Q4, which it expects to be its largest improvement in predictive power to date, and which it says lets it cut rates for better drivers while keeping its target returns. At ~$56/share, $ROOT trades at ~2.6x its $327.0M book value as of September 30, the low end of the $55-80 base case in my August 10 update, on a Q3 annualized ROE of ~18%, the low end of that case's 18-22%. So I read Q3 as $ROOT playing the soft market the way I expected, holding its margins and shrinking rather than buying growth, which keeps the stock priced as if it never grows again. What I'd want to see from here: 1) Policies: Whether policies in force grow from the flat 484k once the new pricing model reaches more of the book. 2) Partnerships: Whether new business through partners and independent agents, up 19% y/y in Q3 and now 52% of the total, keeps growing. 3) Loss ratio: Whether 2026's first estimates come down a year later, like Q1-Q3 2025 did.

Nvidia’s Future | My Question Marks

Nvidia’s future is framed around alleged China GPU diversion, tighter export controls and supply-chain risk after Jensen Huang’s claimed…

Qualcomm Custom ASICs?

Qualcomm’s Alphawave SerDes, CPU cores and HBC could support custom AI ASICs as hyperscalers build heterogeneous fleets beyond GPUs.

QXO Inc Earnings Update: Q3 2026

QXO's Q3 sales rose 79% to $4.9 billion after TopBuild and Kodiak acquisitions, while acquisition debt lifted interest expense to $163…

Cerrado Gold: Research Note

Cerrado Gold will acquire Argentina’s Manantial Espejo mine from Pan American Silver for $20 million and targets a 2027 restart funded by…

Publications we follow

Every publication on this site, hand-picked.

130 publications

Newsletters 110

X accounts we follow

Every account on this site, hand-picked.

91 accounts

Investor letters we follow

376 firms

Sectors

Ordered by how often our editor reaches for them.