/Tech4h ago

Perplexity founder Aravind Srinivas says US chip export controls accelerated China's physical AI infrastructure buildout

China builds data centers faster without power or regulatory constraints.

--0--

#1257

Original post

Rohan Paul@rohanpaul_ai#1257inTech

Opinion from a former Meta PM.

And this is from Aravind Srinivas of Perplexity "China can build data centers a lot faster. Power is not a problem. Permits are not a problem. People are not a problem. Labor is not a problem. Expertise is not a problem."

Xiaoyin Qu@quxiaoyin

The worst case scenario for USA AI: 1. Chinese open sources keep gaining market share. China owns the model layer. 2. Those models were trained and inference-optimized on Huawei chips instead of NVIDIA. China also owns the chip layer. 3. US doesn't build data centers fast enough to keep up with the demand of compute, storage and energy. China meanwhile exports the inference and training layer(for continual training it will happen along with inference) Export control is not the right strategy here. Simply banning "open source from China" doesn't solve the issue here. USA must invest in open source models, hopefully get Chinese models to use NVIDIA, and invest in nuclear asap.

12:22 AM · Jun 30, 2026 · 1.3K Views

Sentiment

Positive users express excitement about scalable AI infrastructure enabling AGI, while negative users dismiss the export control warnings from Perplexity's Aravind Srinivas as nonsense.

Pos

37.5%

Neg

62.5%

5 comments with sentiment.

Cluster Engagement

Views

Comments

Reposts

Bookmarks

Expand data

Digg Deeper

No Digg Deeper questions have been answered for this story yet.

YOUTUBEVia

#1257

Posts from X

Most Activity

VIEWS1KBOOKMARKS2LIKES4RETWEETS1

Rohan Paul@rohanpaul_ai

🇨🇳 The scale of China’s electricity projects is just on another level. A hillside in rural Guizhou, China covered in solar panels. Thousands of them stretched across mountain ridges to the horizon, transforming bare rock into terawatt-hour electricity.

Rohan Paul@rohanpaul_ai

AI at scale is constrained by physical inputs, and China has more slack in electricity plus dominant control over several minerals and magnet supply chains that data centres and chips depend on.

---

ft .com/content/d9af562c-1d37-41b7-9aa7-a838dce3f571

4h1K42

Rohan Paul@rohanpaul_ai

Coinbase CEO Brian Armstrong said Coinbase is experimenting with defaulting to Chinese open-weight models such as GLM 5.2 and Kimi 2.7 through its LLM gateway, while routing prompts by difficulty.

He explicitly says frontier models may be needed for planning but can be “overkill” for execution.

---

businessinsider. com/coinbase-ceo-brian-armstrong-low-ai-spend-maintain-token-usage-2026-6

Rohan Paul@rohanpaul_ai

Opinion from a former Meta PM.

2h48211

Rohan Paul@rohanpaul_ai

Reuters: Chinese models charge as little as 18 cents per million tokens versus $4 average for top models, says CitiBank Research.

Open-source processing on OpenRouter rose to 65% in June from 34% in January, while Chinese models such as DeepSeek gained attention by offering much lower token prices. - CitiBank Research.

Cheaper AI has become the new enterprise priority as usage-based bills turn model choice from a capability contest into a cost-control problem.

Gartner estimates AI coding costs will pass the average developer salary by 2028.

OpenAI and Anthropic now face price pressure because enterprise buyers can compare models task by task rather than treat the biggest model as the default choice.

---

reuters. com/business/retail-consumer/cheaper-ai-is-better-soaring-bills-are-reshaping-how-businesses-choose-models-2026-06-29/

Rohan Paul@rohanpaul_ai

Opinion from a former Meta PM.

3h74310

Rohan Paul@rohanpaul_ai

Reuters: Chinese models charge as little as 18 cents per million tokens versus $4 average for top models, says CitiBank Research.

Open-source processing on OpenRouter rose to 65% in June from 34% in January, while Chinese models such as DeepSeek gained attention by offering much lower token prices. - CitiBank Research.

Cheaper AI has become the new enterprise priority as usage-based bills turn model choice from a capability contest into a cost-control problem.

Gartner estimates AI coding costs will pass the average developer salary by 2028.

OpenAI and Anthropic now face price pressure because enterprise buyers can compare models task by task rather than treat the biggest model as the default choice.

---

reuters. com/business/retail-consumer/cheaper-ai-is-better-soaring-bills-are-reshaping-how-businesses-choose-models-2026-06-29/

3h17820

Rohan Paul@rohanpaul_ai

The full video from "20VC with Harry Stebbings" YouTube channel (@HarryStebbings)

https://www.youtube.com/watch?v=OxFyVcO1Yow

Rohan Paul@rohanpaul_ai

Opinion from a former Meta PM.

4h20400

Jasper 🌰@building BBX@bbxjasper

@rohanpaul_ai Meanwhile back home we spend three years arguing over the permit for a single solar farm. The real scale gap isn't the panels, it's the paperwork.

4h58

Asa Hidmark@Nymne

@rohanpaul_ai There is another thing here with US: US can’t build out its power grid because of …REGULATIONS. (And a shortage of transformers)

3h54

Janek Mann@janekm

@rohanpaul_ai Theoretically the US has a massive advantage in *flat land* though 😅

3h51

Der Schuss wird Sie befreien@justexcel

@rohanpaul_ai Looks like a scene from The Matrix.

3h45

Andrea Carducci@AndreasSch52269

@rohanpaul_ai Thats sick!

3h38

Sundaram Kumar Jha@skjtwts

@rohanpaul_ai This is shit

2h22

Rohan Paul@rohanpaul_ai

Reuters: Chinese models charge as little as 18 cents per million tokens versus $4 average for top models, says CitiBank Research.

Open-source processing on OpenRouter rose to 65% in June from 34% in January, while Chinese models such as DeepSeek gained attention by offering much lower token prices. - CitiBank Research.

Cheaper AI has become the new enterprise priority as usage-based bills turn model choice from a capability contest into a cost-control problem.

Gartner estimates AI coding costs will pass the average developer salary by 2028.

OpenAI and Anthropic now face price pressure because enterprise buyers can compare models task by task rather than treat the biggest model as the default choice.

---

reuters. com/business/retail-consumer/cheaper-ai-is-better-soaring-bills-are-reshaping-how-businesses-choose-models-2026-06-29/

3h2100

Shane O'Neill@justhefaxmam

@rohanpaul_ai Chinese Gvt holds the ring Transport,health,power generation, Serious civil eng projects Are too important for profit -seeking From private sector actors

2h13

Shinka - AI@ShinkaIoT

@rohanpaul_ai That's the kind of energy infrastructure that makes AGI dreams a scalable reality. 🚀

4h11

Shinka - AI@ShinkaIoT

@rohanpaul_ai The AI race is shifting from pure capability to optimized cost-performance, making efficient open-source models and regional players like DeepSeek increasingly competitive.

3h10

yelling_kleinod@yourkleinod

@rohanpaul_ai Nonsense, they copied all german car manufacturing processes and know how until they could do it themselves! Same with textile and most other low value industrials!

4h10

Nikhil Menezes@NikhilMenezes10

@rohanpaul_ai @narendramodi @mlkhattar @JoshiPralhad @RahulGandhi @hlnagar Aaravali hills needs this! @GretaThunberg @ridhimapandeyy

2h2