A five-level ambition scale for AI foundation model labs

We're in a unique moment for AI companies building their own foundation models. Many industry veterans and legendary researchers are launching new labs. Some aim to become OpenAI-sized behemoths; others want to focus on research without worrying about commercialization. That mix makes it hard to tell who's trying to make money and who isn't.

AI Lab Ambition Scale: Who Wants to Commercialize?

To simplify the landscape, here's a five-level sliding scale for any organization building a foundation model. The scale measures ambition, not success — it asks whether a lab is actively trying to commercialize, not whether it has already done so.

The five levels (ambition, not revenue)

  • Level 5: We are already making millions every day. Full commercial focus and scale.
  • Level 4: We have a detailed multi-stage plan to capture major market share and generate enormous revenue.
  • Level 3: We have several promising product ideas and a vague roadmap. Productization is likely but not locked down.
  • Level 2: We have the outlines of a concept or plan. Research dominates and product intent is tentative.
  • Level 1: Commercialization is not the priority. Researchers value autonomy and scientific progress over revenue.
Measure labs by their stated ambitions, not by current cash flow.

Why this scale matters

Many investors and journalists treat every new AI lab as a potential unicorn. But the true source of tension in the industry is mismatch: when a lab's stated or expected ambition doesn't match what it actually does. For example:

  • OpenAI's shift from a research-focused mandate toward aggressive commercialization created confusion and debate. That felt like a jump from Level 1 to Level 5 for many observers.
  • Meta's early research posture arguably looked like Level 2, while the company's internal goals aligned more with Level 4 ambitions.

Because founders, researchers, and backers can often choose their path, many labs sit somewhere between levels. Right now there's so much capital in AI that investors will often back a lab even when product plans are fuzzy. That lets teams pick a level based on preference: those uninterested in becoming billionaires might prefer Level 2; those chasing market dominance aim for Level 4 or 5.

How four high-profile labs score on the scale

Humans—Level 3

Humans& was this week's big AI headline and inspired this scale. The founders emphasize a next generation of models that prioritize communication and coordination rather than pure scaling laws. That's an intriguing research pitch.

But the company has been coy about concrete monetizable products. The clearest signal: they say they plan an AI workplace tool that could replace or reinvent products like Slack, Jira, and Google Docs. The description hints at a commercial product but stops short of specifics.

Given the product hints and public positioning, Humans& best fits Level 3: tangible product intent, but not a fully detailed commercialization plan.

Thinking Machines Lab (TML)—borderline Level 3–4

TML started as a heavyweight bet: a former ChatGPT CTO and project lead raised a multibillion-dollar seed round. That suggested a clear path toward market leadership, aligning with Level 4 ambitions.

Recent executive departures complicate the view. CTO and co-founder departures, plus multiple other exits, signal internal disagreement about direction. Those moves raise the possibility that TML's roadmap was less solid than expected — perhaps closer to Level 2 or 3.

At the time of writing there's not enough public evidence to downgrade them definitively, but the situation makes their Level 4 placement fragile.

World Labs—Level 4 (potentially Level 5)

Led by Fei-Fei Li, World Labs began with strong academic credibility. After raising $230 million in 2024 for spatial and world-modeling work, the lab shipped both a world-generating model and a commercial product.

There's demonstrable demand from video games, visual effects, and simulation industries for world models. Major labs have yet to offer a direct competitor. Those signs point to a serious commercialization trajectory — Level 4, and possibly Level 5 if growth accelerates.

For readers following gaming news, World Labs' technology is especially relevant: procedural world generation and spatial AI can reshape game development pipelines, reduce art costs, and enable new immersive experiences.

Safe Superintelligence (SSI)—Level 1 (but watch for pivots)

Founded by former OpenAI chief scientist Ilya Sutskever, SSI is a research-first project. The team has insulated itself from commercial pressures, even declining acquisition overtures. There's no apparent product cycle; the focus is on advancing foundational science and building a superintelligent model.

SSI raised roughly $3 billion for this scientific mission. That funding level gives them latitude to remain non-commercial. Still, the lab's founder has publicly noted two scenarios that could prompt a pivot: very long timelines or the realization that the most powerful AI must interact with the world. Either could move SSI quickly up the scale.

Practical takeaways

  1. Use ambition, not balance sheets, to understand AI labs. Public claims, product milestones, and hiring/exits reveal intent.
  2. Expect some labs to change levels rapidly. Research breakthroughs, market demand (especially from sectors covered in gaming news), or executive churn can shift priorities.
  3. If you're an investor or partner, clarify whether a lab aims for research independence or commercialization. Misaligned expectations cause the most friction.

Conclusion

AI's current boom lets labs choose their path. This five-level ambition scale helps readers, investors, and partners judge a lab's likely trajectory. Watching hires, product signals, and customer demand — including interest from gaming and entertainment — will show whether a lab is serious about building profitable products or more interested in pure science.