Every Chip, Every Ride

Here's the description with numbers converted:

Missiles in the Persian Gulf. Oil at a hundred dollars a barrel. Consumer sentiment near its prior trough. And somehow, the Nasdaq closed the week at an all-time record — carried almost entirely by AI earnings. Two companies reported on the same day. One posted the best quarter in its 35-year history and got sold. The other missed its headline revenue number and rose sharply. Both are asking the same question in the same week: what happens when a toll collector starts driving on its own road?

In this episode:

• Why Arm's record quarter — $1.49 billion in revenue, 49% operating margins, data center royalties doubling for the 4th consecutive quarter — still produced a reversal after hours, and what the RPO miss actually means for a licensing business whose fastest-growing revenue line doesn't show up in backlog at all
• What AGI actually stands for in "Arm AGI CPU" — it is not what you think — and why the chip that generated more than $2 billion in customer demand 6 weeks after announcement may be the product of a $6.5 billion acquisition made 4 months before anyone announced it
• The revenue capture math that makes the silicon move significant: Arm's traditional licensing generates roughly $0.10 to $2.00 per chip; a complete data center CPU sells for thousands
• Why Arm lost round one of its lawsuit against Qualcomm — and why Qualcomm's separate countersuit, which directly targets Arm's intent to compete in silicon, is the legal risk that actually matters for the long-term thesis
• The optionality tension most analysts are not naming clearly: operationally, the AGI CPU is optionality on top of a royalty engine that already works — but at roughly 100x forward earnings, the valuation is not optional at all
• Why Uber's 14.4% revenue growth was 9 percentage points lower than it would have been under prior accounting treatment, and why the $2.3 billion in free cash flow in a single quarter is the only number that actually matters
• The AV data from the markets where autonomous vehicle competition is most advanced — what it says, why Deutsche Bank and MoffettNathanson read it in opposite directions, and what that disagreement tells you about where the thesis actually stands
• The Morgan Stanley 2032 model: the most specific and uncomfortable version of the Uber bear case, stated fairly and answered directly

Read the written version — with card breakdowns, segment data, and the sourcing that doesn't translate to audio — at quietvelocity1.substack.com, the companion Substack to Conviction Bet.

New episodes weekly. Subscribe on Apple Podcasts, Spotify, YouTube, or Amazon Music.

Conviction Bet is independent investment commentary. Nothing in this episode is investment advice.