MBI Deep Dives

MBI Deep Dives

Meta's Heterogeneous Fleet

Last month, I mentioned the following on Meta’s Chip Resilience: “…in an earlier blog post last year, Meta’s own engineers mentioned dealing with 5–6 hardware SKUs a year makes it harder to move workloads around and can create underutilization and software friction. From Meta’s own blog

Token Cost Conundrums

A couple of weeks ago, I highlighted a couple of pieces by Anjali Shrivastava who made the point that a token is not a fixed unit of cost. Today’s piece is a follow-up driving the point home that even though every model charges its API customers per token, tokens

The Counterintuitive Truths

In my investing career, there are perhaps couple of truths that felt counterintuitive to the extent that my first instinct was to fight it. Not only didn’t it feel right, what complicated it further is accepting these truths felt deeply uncomfortable. The first was the paradox of skill. As

Huang Isn't Planning to Wake Up a Loser

You may have already listened to Jensen Huang on the recent episode of Dwarkesh podcast. If you haven’t, I highly recommend you listen to the entire episode and not just the bits and pieces clips. This wasn’t the kind of episode that you want an AI to summarize

TSMC 1Q'26 Update

Yesterday’s TSMC earnings call was yet another confirmation that the AI momentum right now is one-way street: up and to the right! TSMC’s CEO C.C. Wei in his prepared remarks encapsulated the sentiment around AI (emphasis mine): “AI-related demand continues to be extremely robust. The shift from

ASML 1Q'26 Update

A new earnings season has arrived. If ASML’s earnings is any guide, be prepared to hear another quarter of “demand is outstripping supply” from all the players in the semi value chain. ASML’s CEO Christophe Fouquet started his prepared remarks with the following which set the tone: “…for

AI Economics in the East: Part 2

Yesterday, Flo Crivello tweeted something that caught my attention: We've tested new OSS models the moment they're released for a while at Lindy. Inference is our #1 cost by a lot (more than payroll) — cutting it by 2-5x would be transformative. Last year, OSS models were

Frontier AI's Economic Engine

It should not surprise anyone that coding has found the most product market fit in AI, but it is notable that how much it surpasses everything else in enterprise segment so far which explains so much why Anthropic has been on a tear. a16z recently highlighted AI’s adoption and

AI Economics in the East

In the most recent All-in episode, Brad Gerstner from Altimeter (which owns stakes in both OpenAI and Anthropic) pushed back at the gross margin concerns for AI labs. From Brad Gerstner (emphasis mine): “Their gross margins are exploding higher. Like the fastest increase in gross margins I've probably

Too Cheap to Meter, Too High to Measure

Programming Note: As a reminder, every Sunday, I write pieces that are predominantly based on personal experiences which may or may not be loosely connected with investing. If you are reading MBI Deep Dives everyday, I think it would be rather useful for my readers to understand my personal lens