Chon Tang

This account is verified. Learn more

@chontang

Berkeley/MIT. Former quant, always entrepreneur. VC launching startups at UC Berkeley @skydeck_cal, Silicon Valley’s global accelerator. Opinions mine!

Berkeley, CA SkyDeck.berkeley.edu Joined June 2009

762 Following 4K Followers 5.5K Posts

Latest posts 5

Chon Tang @chontang 5h

Unfortunate outcome. The raise wasn't done correctly and the ask probably wasn't right - but no earnest founder should be bullied out of the arena. We go through this 40 times a year with founders that look like him (deeply technical, first time founder, no SV cred)... So tough.

Taelin Verified account @VictorTaelin 10h

guys I get it, my deck is bad, I'm retarded and a terrible person

I don't have anything to say in my defence

I'll take a break from social media, think about all of this and be back only if I have something worthy of your time

sorry :(
View quoted post on X

Chon Tang @chontang 3d

Uh, 7% MoM growth is not gonna cut it... especially with a full month of Opus 5.5.

Note Openrouter token traffic was up roughly 20% MoM.

TickerTrends 🔬 Verified account @tickerplus 3d

Anthropic’s October ARR topped $80B, per WSJ, consistent with TickerTrends tracking

We tracked $82.33B as of Oct. 30, up 7.2% from our Sept. 24 weekly release

View quoted post on X

Chon Tang @chontang Oct 28

If this is true, then everyone focusing on inference demand leading to mass scaling of infra is completely off base...

I have been telling people demand for tokens would easily 100x over the next 10 years, but if thats only 20k racks.. Thats only 2.7 GW.

Wow.

James Rothmann Verified account @JamesRothmann Oct 28

Though Experiment:

If all inference tokens in the world today ran on GB300 NVL72 Racks and DeepSeek v4.1 Flash, we would only need 200 Racks.

Let's say, Opus 5.5 needs 15 times this, then we need 3000 racks.

Nvidia is cranking out like 60,000 (equivalent) racks this year.

View quoted post on X

Chon Tang @chontang Oct 27

This is really interesting.

All of OpenRouter's Deepseek demand (roughly 30% of total OR token traffic) could be satisfied by 5-10 GB300 NVL72 racks. Even if that 10x over the next year, it's 50-100 racks.

NVIDIA is shipping 90k racks next year? What am I missing?

dax Verified account @thdxr Oct 27

ok to put what i'm saying about demand in context

openrouter is the most successful model router, you can almost serve 100% of their deepseek traffic yesterday on a single GB300 rack

this is worth 6-7M...even if there's 10x this demand secretly that's still < 100M of compute
View quoted post on X

Chon Tang @chontang Oct 8

I only want my math ethically sourced.

Ellie Sleightholm Verified account @elsleightholm Oct 7

Important statement from the Association for Human Mathematics

ahmath.org/statements

View quoted post on X