From Limited to Limitless
Issue 01: The binary flip / $17 billion for everyone / Apple's sad reality
Hello Limitless nation,
Welcome to the Limitless newsletter. We’re happy you’re here.
We live in unprecedented times, and we’re all insanely lucky to get to witness them.
Limitless is here to bring you dispatches from the frontier as humanity pushes forward into its next era of progress. We’re going limitless, baby.
Every week, you’ll get a rundown on the week’s buzziest frontier tech headlines with some of our candid takes and an essay charting a topic we’re feeling passionate about. We’ll also make sure you see our latest podcast episodes.
Our goal is to serve you insight you can’t find anywhere else.
We’re excited to start this journey, and we’re thrilled you’re here with us from the beginning.
Sponsor: KGeN — KGeN is building the Verified Distribution Protocol for the next Web3 wave.
ESSAY: The Binary Flip
Josh on what Limitless is all about.
So, for all of human history, everything we built was limited by two constraints: how much energy we could capture, and how much intelligence we could apply.
With the advent of AI, we’ve invented a way to directly convert energy into intelligence.
It's a monumental shift.
Over the next few decades, this converting mechanism is going to radically redefine everything around us that we call life.
Limitless exists to explore this phenomenon up close – to understand what happens to reality as the cost of energy races to zero, while the amount of intelligence ascends parabolically to levels incomprehensible to our simple wetware brains.
The binary flip from limited, to limitless.
For billions of years a rock was just a rock until one day someone with enough knowledge understood that one rock was different from the others. This rock was iron ore from a hematite rock. Smelt it, and you get iron which leads to cities built of steel. Even the average backyard stone holds, in theory, more energy than coal because of trace uranium and thorium.
The bottleneck isn’t the rock, it’s our know-how to unlock its potential energy.
The pattern repeats: sand becomes silicon becomes a microchip, nitrogen from the air becomes fertilizer that feeds billions.
Knowledge turns things into resources.
What we call scarcity is often just a gap in understanding.
The solution lies behind a beautiful flywheel: more energy yields more intelligence yields asking better questions yields solving harder problems.
AI is the ultimate enabler for this progress.
With the advent of the transformer we've created a way to inject energy into GPUs on one side, and get intelligence out of the other enabling the abundance flywheel. It's likely one of the last great breakthroughs we'll make as humans and the one that ultimately serves as a bootloader for this new intelligence we refer to as artificial.
A remarkable future lies on the other side of AI generating new science and physics and here we sit today right on the cusp of it.
There’s a quote by Steve Jobs I've always loved that says, “Everything around you that you call life was made up by people that were no smarter than you.”
For the first time we get to ask the question, what happens if it was?
Thank you for joining us on this journey. Welcome to Limitless.
RUNDOWN: Everyone Gets $17 Billion?
Ejaaz’s candid takes on this week's huge developments across AI and frontier tech.
Ejaaz here. Welcome back to the Limitless High Five, let's get straight into this week's wildness. Here are the week's five headlines that I can't stop thinking about 👇
5️⃣ Elon is spending $17B to shoot microwaves at you
Yes, you read that right and no, I’m not being dramatic… okay, well maybe a little.
This week Elon announced a $17 billion purchase agreement with a company called EchoStar who will provide 50MHz of S-band spectrum to Elon’s satellite company Starlink. This spectrum access will underpin a brand new product called Direct-To-Cell that will beam down high-speed internet access straight to your mobile phone!
Does this sound familiar? Are you getting PTSD flashbacks to being on hold for two hours with Verizon? Well, fear no more – this new service is dead set on killing traditional cellular providers and bringing the 5G mobile network industry into the 23rd century :)
My take:
If Elon’s behind it, I’m excited. I am TIRED (tired) of shitty airplane wifi, losing 5G access in the middle of a random city and my FaceTimes looking like a grainy black and white movie from the Stone Age. If there's one thing Elon has proven he’s good at – it's building and scaling infrastructure. Look at what he’s done with accessing space, electric vehicles and underground transport. I may not WANT to livestream my life 24/7 anywhere in the world in full HD, BUT I want the OPTION to.
Also, he’s totally going to launch a mobile phone (or some sort of next-gen consumer device). He’s already built the chip facility, building the operating system and he owns the number one distribution platform via X. It’s a no-brainer to own the consumer hardware layer in my opinion. Now whether he chooses to compete at the smartphone level or an AI-device level remains to be seen. Either way you can count me in. (btw my co-host Josh completely disagrees with me that he’ll build a mobile phone, we get into a heated debate on this week’s episode, check it out!)
4️⃣ MSFT just pumped this company’s stock 50% in one day
So the trend this week appears to be $17 billion dollar investments.
Yet ANOTHER company (Microsoft) committed $17B to buy the compute services of a company called Nebius who, at the time of announcement, was valued at $15B (yep they're paying more in services than the market cap of the entire company).
Nebius is a European-based datacenter group specializing in bare-metal GPU clusters used to train AI models and service inference loads (aka all your ChatGPT prompts!). The deal spans over 7 years with Microsoft paying a gargantuan $2.4B PER YEAR which is 3X more than what Nebius currently makes in an entire year.
Talk about hitting the jackpot.
My take:
If anything proves we’re NOT in an AI Capex bubble, it’s this. For months now people have been speculating we’re in a massive spending bubble in AI where the amount invested in training models hasn’t produced nearly enough of the goods they promised. But Microsoft splashing this sort of cash on a random no-name company (sorry Nebius) signals there's simply not enough compute to meet current and expected demand.
Remember Microsoft owns a HUGE stake in OpenAI and is their most high-profile partner. ChatGPT is plugged into every single one of Microsoft's core products & cloud services, and it’s their weapon when it comes to expanding enterprise sales and customer bases. If Satya is spending this kind of money, he knows something. This also makes me bullish to see what kind of applications OpenAI is going to launch soon – I have a feeling it’ll require a lot of inference compute.
Actually, speaking of OpenAI applications…
FRIEND & SPONSOR: KGeN
KGeN is building the world’s largest Verified Distribution Protocol, a.k.a VeriFi — delivering AI & Human Data-driven intelligence through KAI. With over 200+ enterprise partners, and global human-verified data at scale, KGeN has achieved a massive $33.6M+ ARR as of July 2025.
3️⃣ OpenAI’s inaugural AI application.. is LinkedIn??
Yeah I cringed as well.
Fidji Simo, OpenAI’s CEO (of applications) announced this week they’re launching a LinkedIn competitor called the OpenAI Jobs Platform “to connect AI-ready workers with companies who need AI skills, and OpenAI-Certified for workers to learn and demonstrate their AI skills.”
Here’s their pitch: everyone in the world is going to use AI in some shape or form over the next decade and businesses need to figure out how best to integrate it into their products, services and workforce before AI does it for them (and bankrupts them in the process).
Their solution? Connect businesses with AI professionals that are “OpenAI certified” via their academy training system. The platform will also use AI to teach, train and certify professionals who are looking for work.
I can only pray that they don’t emulate the cringe feel-good posts that frequent the LinkedIn timeline.
My take:
I’m pretty unimpressed. I mean, I wasn’t expecting a Steve Jobs iPhone moment but they could’ve at least tried. After a lackluster GPT5 launch, OpenAI needed to convince existing (and future) users that they’re the real deal but all this has done is sow more doubt. I’ve spoken to a number of people who’ve noticed the slower model improvement plus lack of app innovation as suggesting that Sam is getting desperate.
Still, I think it’s too early to call and I’m optimistic about what Simo and Vijay (CTO) could start to cook up.
2️⃣ This robot company is IPO-ing at $7B
High-performance robot civilian manufacturer, Unitree, has announced plans to IPO on the Chinese stock exchange at $7 billion making it one of the largest Asian IPOs in a while.
Their annual revenue currently sits at $140M with 65% of that coming from a robot dog (a market which they own 70% of btw…), 30% from humanoid robots and 5% from sales of sensors, actuators and controllers.
My take:
Look I’ll admit – I didn’t expect robots to progress this rapidly. I kinda thought this market was all some sort of fantastical gimmick for Star Trek nerds that got bullied in high school (and maybe college), but I’ve started to eat my words. These robots are more than just performative, they’re functional – just search “Unitree robot folds laundry” or “Unitree robot helps unload and load a truck,” and you’ll start to quickly realize these things are closer to replacing humans (and maybe even our dogs?) at certain manual jobs sooner than we think.
It also does not make light of the fact that China has become a really convincing opponent to the U.S. empire. I shared on an episode of Limitless a while back that over 50% of the world's top AI researchers reside in China, of which most are paid handsomely by their government to build on-shore. This is what has led to massive leaps in innovation such as DeepSeek’s breakthrough in reinforcement learning and Kimi K2’s insane AI model. Oh, and did I mention they dominate in hardware manufacturing and energy production (cheaper, plentiful) now? They’re a bigger threat to the future of the western economy than you think – it's no longer a meme.
1️⃣ You can now live multiple lives through ChatGPT
If you logged on to ChatGPT this week you might’ve noticed a new feature: Fork your conversation.
This neat new button lets you duplicate any conversation with ChatGPT and run a parallel conversation with it. So instead of having an hour-long conversation and realizing you should have framed the discussion in a different way or added new information earlier - you can just branch the conversation out at any point in the chat and have a separate thread!
My take:
This is really cool because you can experiment with different scenarios and conversation types that you previously weren't able to do. For instance when I’ve used ChatGPT to help me figure out an investment thesis, it often frustrates me that I can’t create a duplicate chat and tweak a few details to get it to be more performant. The new customizability is so freeing and will lead to a more representative and personalized experience for the user – not to mention it will probably 10X OpenAI’s valuable data moat per person as they start to gather even more niche details about their user’s habits and needs.
Thanks for joining us for our first issue. Now, go listen to our podcast :)
