Walid Boulanouar recently shared something that caught my attention because it captures the exact whiplash many people feel about AI right now. He wrote: "us: we are into an ai bubble.... funding will stop.." and then immediately added: "meanwhile: anthropic raised $30B in funding at a $380B post-money valuation."

That contrast is the story of this market in two lines. On one hand, we keep hearing the familiar warning: too much hype, too much capital, too many companies with thin differentiation. On the other hand, the biggest checks in venture history keep getting written, often at valuations that feel detached from typical metrics.

So are we in an AI bubble? And if we are, why are mega-rounds still happening?

The point Walid Boulanouar is making (and why it matters)

Walid is not doing a detailed fundraising breakdown. He is holding up a mirror to our collective narrative.

We often talk about "the bubble" as if funding will suddenly shut off everywhere, all at once. But the reality tends to be uneven. When risk appetite changes, capital usually does not disappear. It concentrates.

Walid Boulanouar is essentially pointing out a mismatch between what we say (funding will stop) and what the market is doing (funding is still flooding to the top).

That mismatch matters because founders, operators, and investors make decisions based on the story they believe. If you think money is about to vanish, you might under-invest, pause hiring, or abandon a product direction too early. If you think the boom will last forever, you might scale cost structure ahead of real demand.

The useful move is to separate three different questions that get blended into one:

Is there hype in AI?
Is there too much funding for many AI startups?
Is capital still available for a small set of AI leaders?

All three can be true at the same time.

Why mega-rounds can coexist with "bubble" talk

When people say "bubble," they usually mean prices are being set by expectation rather than by fundamentals. In AI, expectation is undeniably doing a lot of work. But mega-rounds can still be rational within a specific logic.

1) The market is treating frontier AI like infrastructure

The biggest AI labs are being valued less like typical SaaS businesses and more like infrastructure providers. Infrastructure markets often have:

Massive upfront costs (compute, data, talent, distribution)
Scale advantages (better models, better products, better unit economics over time)
Winner-take-most dynamics (or at least winner-take-a-lot)

If an investor believes a company is building a foundational layer for the next decade of software, then a very large round can be framed as buying a strategic position, not simply buying next year’s revenue multiple.

2) Capital concentrates when uncertainty rises

In periods of uncertainty, many investors reduce the number of bets they make. But the dollars that do get deployed often go toward perceived category leaders.

So the "funding will stop" narrative can be true for:

Pre-seed and seed experiments without clear differentiation
Me-too wrappers with no distribution edge
Teams that cannot explain why their model, data, or GTM is defensible

At the same time, funding can accelerate for a smaller set of companies that appear to have:

Technical moat (or at least credible frontier capability)
Distribution channels (enterprise relationships, consumer reach, partnerships)
Strategic importance (national competitiveness, platform power)

That is not a contradiction. It is a sorting mechanism.

3) Strategic buyers and ecosystems change the math

AI is not only a venture story. It is a platform story. When large tech companies, cloud providers, and governments get involved, the incentives shift.

A mega-round can be influenced by:

Long-term compute commitments
Distribution agreements
Ecosystem control (who sets standards, APIs, and default tools)
Defensive motives (preventing rivals from owning the next platform)

In other words, some funding rounds are not purely financial bets. They are strategic positioning.

What "AI bubble" really means at the company level

Bubble talk becomes actionable when you translate it into operational risk.

The real risk: too many companies with the same pitch

AI lowers the cost of building certain features. That is good for builders, but it also creates a flood of similar products.

If your company can be copied in a weekend using the same foundation models and the same design patterns, investors will eventually price that in.

The bubble is often not "AI" broadly. It is the layer where:

Differentiation is shallow
Switching costs are low
Distribution is rented (ads, marketplaces, SEO with no brand)

The second risk: cost structure tied to external pricing

Many AI products have a margin profile that depends on model pricing, inference costs, or cloud credits. When usage grows, costs can rise faster than revenue if you are not careful.

If the market turns and investors demand discipline, companies with fragile unit economics get squeezed first.

The third risk: expectations outrun adoption speed

Even if AI will transform workflows, adoption can be slower than hype suggests. Security reviews, procurement cycles, compliance constraints, and change management all slow down real deployment.

So yes, you can have long-term truth and short-term froth at the same time.

A practical way to read the current funding signals

Walid Boulanouar’s two-line post is a reminder to look at the funding market in segments, not as a single temperature reading.

Here is a simple framework:

Segment A: Frontier model builders

Capital intensity: extremely high
Narrative: infrastructure and platform
Likely outcome: a small number of winners, massive stakes

Segment B: Vertical AI with distribution

Capital intensity: moderate
Narrative: workflow ownership in a domain (legal, healthcare, sales, finance)
Likely outcome: durable companies if they own data, outcomes, and channels

Segment C: Thin wrappers and undifferentiated copilots

Capital intensity: low to moderate
Narrative: speed and novelty
Likely outcome: fast churn, pricing pressure, consolidation

When people say "funding will stop," they are usually talking about Segment C. When you see headline rounds, they are usually Segment A.

What founders should do with this insight

If you are building in AI, Walid’s contrast suggests a clear takeaway: plan for a bifurcated market.

1) Stop relying on the macro story

Whether the headline is "bubble" or "boom," your job is to build:

Clear differentiation (data advantage, workflow integration, outcome guarantees)
Defensible distribution (partnerships, community, brand, embedded channels)
Healthy unit economics (measured, monitored, and improving)

2) Raise like capital is selective, because it is

Even if mega-rounds are happening, assume your round will be judged on:

Why you win in your niche
Proof of willingness to pay
Evidence you can scale without exploding costs

3) Speak to the new investor question: "Why you, not a model update?"

Every foundation model release compresses feature advantages. Investors know that.

If your pitch can be erased by a model improvement, you need a new pitch.

Closing thought

Walid Boulanouar’s post is short, but it lands because it is honest about the confusion: people warn about an AI bubble while the market keeps producing astonishing fundraising headlines. The best interpretation is not that one side is lying. It is that the market is splitting into tiers.

Capital is not uniformly risk-on or risk-off. It is increasingly concentrated, increasingly strategic, and increasingly impatient with undifferentiated products. If you internalize that, you can build and fundraise with clearer eyes.

This blog post expands on a viral LinkedIn post by Walid Boulanouar. View the original LinkedIn post →