Exponential View 🔮 @ExponentialView

Unlocking the near future • AI • hardware • energy • economy • geopolitics • climate change • business → newsletter 📧 podcast🎙️ community👐 exponentialview.co Join 100,000+ (it's free) → Joined May 2016

Tweets

5K
Followers

15K
Following

2K
Likes

4K

Raghav Agarwal @0xRaghav

6 days ago

Once you truly understand what part of the process is increasing token consumption, you will be forced to use AI in a completely different fashion. According to the analysis presented by @ExponentialView "A coding agent that operates over 10 turns might need to re-read its full context every turn. That repetitive reading of context could use as many as 55x more tokens than a single-turn query for the same task. Actual active inference is probably only 15-20% of the total token consumption. The rest is invisible work that you, as a user, and possibly the company paying for it all, haven’t modelled." Proper documentation standard and knowing how to provide context effectively helps you reduce your token consumption. Think of token usage like fuel for your car. Don't waster it regardless of what model you are using.

Raghav Agarwal @0xRaghav

3 weeks ago

x.com/i/article/2062…

1 3 18 2K 6

0 2 1 601 2

View Details

Foundation Capital @FoundationCap

6 days ago

We’re excited to share that AI in the Real World with Foundation Capital is now live on all major podcast platforms! Tune into our latest episode with @azeem of @exponentialview, where he and @joannezchen discuss the ways in which the great reorg is only just getting started, now on Spotify, Apple Podcasts, YouTube, and more. youtu.be/gYGyfHRwGTU

1 2 3 932 0

View Details

Navin Kabra @NGKabra

3 weeks ago

@aparanjape Want to maximize your use of AI but still give your brain enough exercise? Maybe buy a good quality notebook and fountain pen, as @ExponentialView suggests, and spend a few hours every week thinking with paper and pen. cc: @docbhooshan

0 2 7 1K 1

View Details

GigaTrion @GigaTrion

4 weeks ago

Azeem and his prop research is so good. Been subscribed to substack for almost a year. Truly novel and informative

Azeem Azhar @azeem

4 weeks ago

A general-purpose technology can sit inside firms for ages before we see the results. It happened with electricity, it’s happening with AI.

17 51 331 160K 209

0 1 5 675 0

View Details

Dan Shipper 📧 @danshipper

4 weeks ago

Extremely smart take on the tokenmaxxing panic and why it won’t last:

Azeem Azhar @azeem

4 weeks ago

A general-purpose technology can sit inside firms for ages before we see the results. It happened with electricity, it’s happening with AI.

17 51 331 160K 209

11 32 269 54K 224

View Details

Epoch AI @EpochAIResearch

4 weeks ago

Are we nearing a compute crunch? In our latest Gradient Update, @luke__emberson and @Jsevillamol estimate how many tokens all the Blackwell chips on Earth could serve, and compare this to total token demand. Direct comparisons are difficult, but it appears demand is growing much faster than supply.

7 34 270 29K 137

View Details

Exponential View 🔮 @ExponentialView

4 weeks ago

📈 Why AI bills rise as costs fall exponentialview.co/p/data-to-star…

1 0 0 274 0

View Details

Shanu Mathew @ShanuMathew93

a month ago

.@ExponentialView estimates that the number of tokens processed per quarter has grown by around 17,000x over four years.

2 10 35 6K 31

View Details

Azeem Azhar @azeem

a month ago

Tokens are exploding. Enterprise AI bills are getting bigger. And both trends are accelerating at the same time. 🧵

2 1 2 1K 2

View Details

Daily Chartbook @dailychartbook

a month ago

"In the US, the average monthly spend on AI by large enterprises grew 36% to $85,000 between 2024 and 2025" @azeem @ExponentialView

0 6 24 5K 6

View Details

Ibrahim Ajami @IbrahimAjami

a month ago

The cost killers are coming Via @azeem

3 14 129 21K 58

View Details

Azeem Azhar @azeem

a month ago

Uber & ServiceNow blew through their 2026 token budgets in four months. AI is diffusing fast and finance is scrambling: exponentialview.co/p/monday-data-…

3 1 9 1K 5

View Details

Matt Stoller @matthewstoller

a month ago

The U.S. is losing the AI race, because China has competition whereas we have monopolies. exponentialview.co/p/inside-chine…

7 10 64 5K 10

View Details

Vincent Colbert @vcolbert

a month ago

It’s a fascinating period when you take a step back… a very good article on the Chinese approach. #AI #China @hannahjpetrovic @ExponentialView exponentialview.co/p/inside-chine…

0 1 3 257 0

View Details

Azeem Azhar @azeem

a month ago

A couple of weeks ago, @Bloomberg reported that TSMC won't buy ASML's newest chipmaking machine until at least 2029. ASML's High-NA EUV is the most advanced lithography tool ever built. When the world's #1 foundry passes, that's a strong signal. Today's @exponentialview analysis w @hannahjpetrovic exponentialview.co/p/the-broken-b…

0 7 11 2K 8

View Details

Azeem Azhar @azeem

2 months ago

Over the past week, I have been in China meeting AI and robotics teams, including Zhipu and MiniMax (the two publicly listed foundation model companies), as well as Kimi, Alibaba, Xiaomi, Bytedance, and others. Demand is booming. Zhipu, for example, is serving 5.5 trillion tokens per day. Developers are rushing to the platform, joining at about ten a minute. Keeping up with inference loads is a struggle in China, as it is in the US. Teams universally acknowledged compute constraints, particularly the shortage of Nvidia chips. But that isn’t stopping innovation under those constraints. Researchers and developers are free to use whichever model they want. Anthropic’s Claude was the preferred choice for technical teams, but it was clear that “dog-fooding” is commonplace. I'll write something in more depth in the next couple of weeks. For now, here’s a short video from one of the many demos we saw at Unitree:

8 22 172 27K 79

View Details

Azeem Azhar @azeem

2 months ago

Here's @ExponentialView analysis of the solar supercycle exponentialview.co/p/solar-superc…

0 1 0 768 1

View Details

Daily Chartbook @dailychartbook

3 months ago

"Golf courses in the US consume more than 30x as much water as the entire country’s data center industry uses for cooling" @azeem @ExponentialView

17 55 250 20K 76

View Details

Navin Kabra @NGKabra

3 months ago

Pakistan's people installed solar panels in record numbers because their government couldn't be trusted to keep the electricity on... As a result, Pakistan hasn't been as badly hit by the Iran war as other similar countries, argues @ExponentialView

1 3 14 1K 4

View Details

Rui Ma @ruima

3 months ago

Chinese renewable energy is far more ambitious than just EVs or grid electrification. This interactive chart from @ExponentialView by @azeem is a great way to see it: as solar gets cheap enough, it does not just take share in power generation. It starts to make entire incumbent systems vulnerable. Readers can play with the assumptions on the original site and see how quickly the timing changes. A few examples of what ultra-cheap solar can start to unlock: Grid power is the most obvious one. Global electricity generation reached about 30,900 TWh in 2024, and solar still provided only 6.9% of it, which shows how early we still are. Aviation fuel is another. Global airlines’ fuel spend was projected at about $291 billion in 2024 and $248 billion in 2025, so this is already a roughly $250 billion to $300 billion annual market at current fuel prices. Industrial heat may be even more important. Industry accounts for nearly 40% of global final energy demand, and much of that energy is used in heat-intensive processes. That implies an enormous incumbent fossil energy market, plausibly well above $1 trillion a year on a rough back-of-the-envelope basis. Green hydrogen is especially interesting because it first displaces existing fossil hydrogen demand in refining and ammonia, and then could move into much larger industrial systems like steel. It is already being piloted at scale, and there are strong reasons to think it will ramp seriously over the next decade. And then there are areas like desalination and carbon capture, where the value is harder to summarize as a single market number but could be enormous in social and economic terms. Link to the chart in the comments. It is worth playing with the assumptions yourself to see when different cost thresholds start to make these incumbent systems vulnerable.