Genspark's Jing Kun: What Does Unlimited Token Actually Mean?
Once again, we find ourselves at a crossroads in history.
Last week, we launched the ZhenFund Token Grant, topping up tokens for more AI founders who have ideas and have already started building, helping them ship their first product faster.
Recently, Genspark founder Eric Jing also shared in a talk: a 50-person company where every employee has unlimited access to AI can already outperform a 500-person company without that capability.
Not 10% faster. 10x faster.
The efficiency gap between companies is no longer measured in percentages. In the AI era, companies that give every individual full token access will operate at 10x, 20x, even 100x the speed.
In his first six essays, Jing wrote about the arrival of AGI, how to build AI-native teams, the changes brought by vibe working, why multimodality is the future, and how we should prepare for the next generation. These were all about how humans evolve alongside AI.
But this one is different. It's no longer just about people; it's also about companies. He's become increasingly convinced of one thing: over the next five years, what truly separates the winners from the losers is whether every individual has unlimited access to AI resources.
Below is the original sharing, compiled by ZhenFund:
Every week, CEOs, founders, or executives pull me aside in different settings. At conferences, over dinner, on phone calls — they all ask me one question: "How exactly should we use AI inside our company?"
I understand why they ask. It sounds like the right question, strategic and responsible, showing they're taking AI seriously.
I've always valued the sincerity behind the question, but every time I hear it, I feel a slight unease. Because the question itself reveals a fundamental misconception: it treats AI as a project to be "implemented," with a start date, scope, rollout plan, completion time, and assumes there exists a linear path from today to becoming an AI company.
Our job is simply to find that path and execute.
But the real question is much simpler: "Have we provided unlimited AI resources to everyone in the company to think, create, and build?"
If the answer is no, everything else is noise.
Whether it's AI strategy, task forces, pilot projects, or governance frameworks — none of it matters until this question is resolved. Most companies I encounter still answer "no." Without this, you're not deploying AI; you're performing AI usage. There's an essential difference between the two.
What "Unlimited Tokens" Actually Means
When people hear "unlimited AI access," what comes to mind is often a vague concept, like a more open culture toward AI. But it's not. It's something very concrete and measurable.
Tokens are the basic unit of AI work. Every request you send, every document you have AI analyze, every piece of code it generates, every agent you launch — all of it fundamentally consumes tokens.
Tokens are the raw material of AI productivity, just as electricity was in the industrial age and bandwidth was in the internet era.
When a company sets monthly token caps, requires IT approval for advanced models, blocks certain AI tools on the corporate network, or has 20 people share one account, it's fundamentally doing one thing: limiting employees' ability to use AI, placing a valve on their cognitive output, directly controlling how much intelligence they can summon to complete their work.
"We are limiting how much intelligence our employees can use" — saying this out loud sounds absurd.
Yet this is exactly what most companies are doing today, not out of malice but out of inertia. People are accustomed to treating new technology as a cost to control, rather than a capability to unleash.
I myself went through a similar moment in the early 2000s. Some companies gave every employee full internet access, saying you could use it to do your job better; others chose to block websites, monitor usage, and institute strict corporate policies about what was and wasn't allowed.
Ten years later, the first group had largely become industry leaders; the second group had failed. They failed not because of their internet strategy per se, but because of their attitude toward new technology.
At a moment when they should have been expanding capability ceilings, they were over-optimizing for control.
Going further back, in the early 20th century, when electric motors entered factories, most factory owners did what seemed natural: they replaced the central steam engine with a motor, but kept the existing belts, drive shafts, and factory layout, simply swapping the power source. Costs did come down somewhat, but the real transformation never happened.
The factories that truly transformed did something else: they tore out the entire transmission system, installed motors directly at each workstation, and rebuilt the whole production system from scratch around this new architecture.
The result wasn't a 10-20% improvement; it was 3-5x. Processes previously constrained by physical transmission distances could now be reorganized around logic and efficiency. Production methods that were simply impossible under the old structure began to emerge.
Economist Paul David called this phenomenon the "dynamo paradox" in a 1990 paper. From the first commercial power station to the real productivity explosion, nearly half a century passed.
Because to use this technology well, factories had to rebuild their entire systems. Many factories thought they had completed their technology upgrade, when in fact they had merely layered new capabilities onto old structures, puzzled why their inputs and outputs didn't match. They were using new technology with old thinking.
The situation today is identical. Most companies' "AI deployment" is like replacing the steam engine with an electric motor — a shared subscription, a few approved use cases, a quarterly AI review meeting. The "belts and drive shafts" of the old organizational structure remain firmly in place.
"Unlimited tokens" represents another choice.
It's equivalent to installing an independent motor at every workstation. This is a structural decision, signaling to the entire organization that you are rebuilding this factory, not merely swapping out the power source. Like those factories back then, companies that make this choice won't just improve by 10%; they will enter a completely different productivity tier.
We stand at the same crossroads right now. Only this time, the investment is greater, and the distance opened up will be even wider.
The Era of AI Employees Has Arrived
In my earlier article "AI Will Make the Future a 3-Day Work Week," I mentioned "vibe working" — the psychological and work-style shifts that occur when people stop treating AI as a tool and start treating it as a true collaborative partner.
At the time, this was still in transition. But today, it is complete.
The era of AI employees has arrived.
AI is no longer a tool that helps you write emails faster, nor a search engine you open when you're stuck. It's a colleague, a partner, an expert team on call at any moment — with engineers, researchers, analysts, strategists, designers, writers inside.
AI is online 24/7, no vacations, no emotions, no organizational friction. It doesn't clock out at six, doesn't lose motivation, doesn't need onboarding, and doesn't need an annual salary negotiation.
But there's one prerequisite: this team only appears when you open the door.
Since launching in April 2025, Genspark went from zero to $200 million ARR in 11 months — a pace with virtually no precedent in enterprise AI. Our product hit $10 million ARR in just 9 days, faster than ChatGPT, Claude, and every AI product in history.
All of this was achieved by a team that, by traditional standards, is shockingly small.
100% of our code is AI-generated.
One engineer built an AI browser in three months. One PM delivered AI Slides in two weeks. One designer who had never written code built a browser download site from scratch in three days.
In these 11 months, we shipped AI Workspace 3.0, the first fully autonomous AI employee Genspark Claw, plus Workflows, Teams, Meeting Bots, Realtime Voice, and more.
This wasn't accomplished by a group of geniuses. It was accomplished by a group of ordinary people, using an extraordinarily capable AI team.
A company with 50 employees but unlimited AI capability won't operate like a 50-person company. It will operate like a 500-person, even 5,000-person company. This multiplier effect is real. We experience it every day.
Consider the reverse: a 500-person company with strictly limited AI usage, with token budgets, IT approvals, quarterly evaluations, and carefully designed rollout plans. It's still just a 500-person company, nothing more.
But that 50-person company will ship products at 10x the speed, iterate at 10x the frequency, learn at 10x the pace, and experience 10x more failures.
Every week that passes, this gap widens further.
This Time, the Gap Is Light-Years
From PCs to the internet, to mobile internet, to cloud computing, in every technology wave there have been gaps between early movers and laggards. But these gaps had boundaries — maybe 1.5x, 2x, in extreme cases 3x — and could be closed. A company that was two years behind in 2012 still had a chance to catch up by 2015, though at significant cost, but at least it was possible.
This time is completely different. The gap isn't linear; it's exponential. And I'm not sure it can be closed at all.
Imagine two ships leaving port on the same day: one nuclear-powered, the other rowed by oars. On day one, the gap isn't large. After a week, it's growing distant. After a month, the rowed ship can no longer see the other. After a year, the distance becomes incomprehensible, no longer measured in miles but entering an entirely different dimension.
This is exactly the gap that tokens create.
On one side: a company where employees have unlimited access to the strongest frontier models, where engineers engage in real-time multi-turn dialogue with AI to design entire systems, where product managers iterate research reports in minutes, where executives have already used AI to simulate multiple competitive scenarios before a single PowerPoint slide is written — where every person is compounding cognition.
On the other side: a company where using advanced models requires filing an IT ticket, where AI tool budgets are discussed quarterly, where employees start using their own credit cards to bypass corporate restrictions, where management is still debating whether to expand permissions from engineering to marketing.
The gap between these two companies isn't 10%, isn't 50%. It's the gap between "one is running" and "one is standing still." Every day, the former exponentially pulls away. Faster iteration brings better products and revenue, which in turn accelerates iteration further.
This isn't competitive advantage. This is elimination.
What "Truly Embracing AI" Actually Looks Like
"Fully embracing AI" is easy to say, but also easy to turn into empty words, so let me be more specific.
First, eliminate all token caps and AI usage budget restrictions immediately. Not next quarter, not after the security assessment is complete — now. There will be costs, but they are trivial compared to the productivity gains, and far smaller than the cost of being left behind by the era.
Second, stop treating AI as an IT cost. AI should be in the headcount budget. When you move this expense from IT to headcount, everything changes. It means AI is no longer a software tool to be controlled; it's part of the team.
Every agent you deploy should have its own position, its own workspace, its own reporting line, its own responsibilities, its own output. When AI appears in the org chart, not just in vendor contracts, your team will take it seriously.
No CFO looks at the payroll and says "how do we cut this cost." Salary is the price of human capability; AI access cost is the price of AI capability. In a world where AI already handles 80% of the work, this investment deserves the same level of importance.
Third, build a culture where "using AI for everything" is the default, not the exception. At Genspark, we don't ask when to use AI. We ask the inverse: "Why didn't you use AI for this?"
This inversion matters. It represents organizational-level seriousness, and it drives group-level learning acceleration.
If your company is still evaluating AI or using it only in limited scope, this isn't caution — it's slowness. What truly leading companies are doing now is full deployment, rapid iteration, continuous compounding.
Every extra month you spend evaluating is another month your competitor spends executing.
The New Corporate Wealth Divide
It's easy to feel urgency; it's hard to truly understand the mechanism behind it.
The token gap isn't just a gap in current output. It's fundamentally a gap in learning speed. This is what makes it dangerous.
A company that has provided unlimited AI access to every employee for the past two years hasn't just done twice as much. It has accumulated two years of organizational learning — in ways of working, intuitive judgment, muscle memory, and internal culture.
These things cannot be made up by throwing money at them. You can't acquire your way to being AI-native, and you can't hire your way there in six months. Organizational capability compounds in a way that is almost invisible.
Until one day, the gap is no longer about performance but about entirely different tiers of capability.
The earliest AI adopters have entered a flywheel that is almost impossible to stop. Their products are better, attracting more users; users bring more data and feedback, so products improve faster. Faster iteration brings faster learning; learning in turn supports more AI investment, further accelerating iteration.
Meanwhile, the best talent who can maximize their potential in AI-native environments will naturally flow to these companies. No ambitious engineer or designer wants to waste their career waiting for IT approvals.
Companies that fall behind face a compounding "liability." They aren't just behind in output; they're behind in intuition, culture, talent density. One day, this gap will cross a critical threshold. At that point, the question is no longer whether you can catch up, but whether you're still in the game.
You can't re-enter a race where your opponent uses nuclear power by rowing harder. Falling three years behind on tokens is likely a permanent gap.
This isn't a metaphor. This is reality.
The Final Choice Is Yours
Over the past few months, I've been observing two types of companies.
One type is riding the wave. Not necessarily perfect, but moving forward. They decide quickly, accept uncertainty, embrace the chaos AI brings, and accumulate experience every week. The other type is still standing on shore, watching the wave approach bit by bit, holding meetings to discuss whether to get in the water.
When I wrote my first article "Seeing AGI Arrive," I felt like a father worried for his 12-year-old child. Now I have the same emotion, but directed at the founders reading this article. Because I've seen what happens next, and I truly don't want anyone to be swept away by this wave.
When a tsunami comes, it doesn't wait for you to finish your board meeting, and it doesn't pause while you conduct your review. It simply arrives.
Organizations already in the water, moving with the force, will survive. Those still on shore discussing will be directly submerged.
The window is still open, but it's closing.
Tonight, every founder, every CEO, every operator needs to answer one question: Have you provided unlimited AI resources to everyone in your company to think, create, and build?
If not, then ask yourself one more thing: What are you waiting for?
I've been in tech for nearly 20 years. I've seen market cycles, company rises and falls, paradigms flip overnight. But I've never seen change this fast, with impact this deep.
What truly keeps me up at night isn't the technology itself. It's an image: smart, hardworking founders who have poured years of effort into their companies, waking up one day to find the gap with their competitors is already unbridgeable.
Not because they weren't smart enough, not because they didn't work hard enough, but because at one critical moment, they hesitated, waited for one more data point, held one more meeting, asked for one more quarter to evaluate.
I'm not writing this to create anxiety. I truly believe most people haven't yet felt the weight of this. When they truly realize it, there may no longer be time to act.
So I want to leave you with the single most important point of this article.
The efficiency gap between companies is no longer primarily determined by talent, strategy, or capital, but increasingly by whether you have opened up "unlimited tokens."
Companies that answer "yes," even if imperfect, even if chaotic in process, are compounding their advantage every day. Companies still discussing aren't standing still — they're being pulled away at a speed never before seen in history.
This gap used to be measured in percentages. Now it's measured in multiples. Soon, in some industries, it won't even be measurable, because one side is no longer in the game.
I hope you stand on the side of the future.
If you've read this far, felt a moment of resonance, heard a quiet voice saying "this might be us" — then don't wait until the next board meeting to confirm the answer.
The wave is already hitting you. The real question is only one: Are you in the water, or on the shore?
There is still time, but not much.