The $1.3 Million Token Burn: What Peter Steinberger’s AI Spending Spree Means for B2B SaaS and GTM Strategies
In the fast-moving world of AI and B2B SaaS, one number is turning heads and sparking debates: $1.3 million. That’s the monthly token bill Peter Steinberger, creator of the viral AI agent OpenClaw, posted to X on Friday. The screenshot, showing nearly $20,000 in daily token spending on OpenAI’s API, has ignited a firestorm of commentary about the cost of AI development, the value of free compute, and the emerging culture of “tokenmaxxing” in Silicon Valley.
But behind the jaw-dropping price tag lies a playbook for revenue teams, product leaders, and GTM strategists. Steinberger’s case isn’t just about token waste—it’s about how access to unlimited compute is reshaping the talent wars, the economics of AI, and what it means to build software in the future. As chief editor of B2B Pulse, I’m breaking down what this means for your bottom line, your team, and your growth strategy.
The Shocking Numbers: How One Engineer Spent More Than a Small Startup
Let’s start with the facts. On Friday, Steinberger posted a screenshot from CodexBar, a tool that visualizes token spending across AI coding tools. The image revealed:
- Daily token spend on OpenAI’s API: Just under $20,000
- 30-day total: $1.3 million
- Primary usage: Development of OpenClaw, the viral AI agent that sparked a Mac Mini buying craze and has since become the fastest-growing open-source product of all time
Steinberger, now an employee at OpenAI, doesn’t pay a dime out of pocket for these tokens. As he put it, the funds are “perks of OpenAI supporting OpenClaw.” When asked if OpenAI charged him, he responded simply: “ofc not.”
Still, the figure has left many stunned. One X user commented, “Someone’s burning through enough tokens to bankroll a small startup.” Others questioned whether that budget could be better allocated to hiring new engineers.
But here’s the reality: Steinberger’s spending is a direct result of a deliberate experiment. He wrote that he was testing the question: “How would we build software in the future if tokens don’t matter?” That’s not just a philosophical exercise—it’s a strategic pivot that B2B companies need to pay attention to.
Tokenmaxxing: The New Competitive Edge in AI Talent Wars
What Steinberger’s token bill reveals is a broader shift in the AI talent landscape. Tokenmaxxing—the practice of maximizing token consumption to accelerate development—is sweeping Silicon Valley. OpenAI is reportedly one of several companies with competitive token leaderboards, where engineers compete to use (and often waste) tokens as a badge of status.
For B2B SaaS companies, this isn’t just a tech curiosity. It’s a signal. Access to free compute has quickly become a selling point in the AI talent wars. If you’re trying to attract top AI engineers, offering unlimited tokens is as compelling as a six-figure salary—sometimes more so. Engineers want to build without constraints. The companies that provide that freedom win the best talent.
This is a lesson for your GTM strategy: the value you offer your engineering team directly impacts your ability to ship innovative products. If your competitors are giving their engineers $1.3 million in token credits per month, and you’re rationing tokens like a limited resource, you’re already behind.
The Real Cost of AI: What $1.3 Million Buys You (and What It Doesn’t)
Steinberger’s spending is concentrated on OpenClaw, a project that has already proven its viral potential. The AI agent sparked a Mac Mini buying craze and became the fastest-growing open-source product ever. That’s not just a vanity metric—it’s a signal of product-market fit.
But here’s the sobering truth: intensive use of AI is expensive, and the costs are not slowing down. At $1.3 million per month, Steinberger’s token burn is equivalent to the full-year salary of about 10 senior engineers (assuming $150k each). Many commenters asked whether that budget would be better spent on humans.
Yet the counterargument is equally compelling: without the token spend, OpenClaw wouldn’t exist. The product is building a new category of AI agents that automate complex workflows, make phone calls, and integrate deeply into existing systems. That innovation requires heavy token usage—and the payoff could be massive.
For B2B revenue teams, the takeaway is clear: AI costs are an investment, not an expense. If you’re building AI-powered features for your SaaS product, don’t nickel-and-dime your token spend. Instead, align your budget with the potential ROI of the features you’re building. A $1.3 million token burn that leads to a product that generates $10 million in new ARR? That’s a good trade.
How B2B SaaS Companies Can Compete Without OpenAI’s Deep Pockets
You might be thinking: “My company doesn’t have OpenAI’s budget. How can I compete?” That’s a fair question. Steinberger’s situation is unique—he works at the company that provides the compute. Most B2B SaaS companies cannot afford $1.3 million per month in token costs.
But here are three actionable strategies:
1. Optimize Token Usage for Maximum Output
Not all token spend is equal. Steinberger is using tokens to experiment and iterate rapidly. Your team can do the same by focusing on high-leverage use cases: customer support chatbots, personalized email sequences, or predictive lead scoring. Prioritize tasks where AI’s ROI is highest.
2. Build Token Budgets into Your GTM Plan
If you’re launching a new AI-powered feature, account for token costs in your pricing. For example, you can offer tiered usage plans where heavy AI users pay more. Alternatively, include token credits as part of your enterprise plan. This turns a cost into a revenue driver.
3. Leverage Open-Source Models
Not every AI task requires OpenAI’s API. Open-source models like those from Mistral, Llama, or even smaller fine-tuned models can handle many B2B use cases at a fraction of the cost. Steinberger’s OpenClaw is open-source itself—why not follow that playbook?
The Future of Software Development: Tokens as the New Commodity
Steinberger’s experiment raises a fundamental question: what happens when tokens become so cheap (or free) that they no longer constrain development? In that world, building software becomes about creativity, not efficiency. Engineers can test hundreds of hypotheses per day, iterate in real-time, and ship products at unprecedented speed.
For B2B SaaS, this is both an opportunity and a threat. Your competitors might already be using token-maxxing to out-innovate you. If you’re not experimenting with heavy token usage, you’re falling behind.
But here’s the flip side: tokens are still expensive for most companies. The key is to find the sweet spot where token spend accelerates your growth without breaking your budget. Steinberger’s $1.3 million bill is a wild outlier, but it’s also a proof point: with enough compute, you can build world-class AI agents in record time.
What This Means for Your Revenue Team
As a revenue leader, you shouldn’t just watch this drama unfold—you should act. Here’s your playbook:
- Ask your engineering team: What would you build if tokens were free? Use that list to prioritize AI features that could drive revenue.
- Revisit your budget: Are you allocating enough to AI compute? Consider it a growth investment, not an operational cost.
- Monitor the talent market: If your top AI engineers are leaving for companies with tokenmaxxing perks, you need to match the offer—or risk losing your competitive edge.
Steinberger’s token bill is more than a spectacle. It’s a signal of where the industry is heading. The companies that embrace token-driven development will win. The ones that don’t will be left with yesterday’s software.
Final thought: The next time you hear about a $1.3 million token bill, don’t freak out. Ask yourself: how can we use tokens to build faster, ship better products, and win more customers? Because in the new AI economy, token spend is the new R&D budget—and the winners are already spending big.