Loading prices…
STKR NewsSTKR News0 of 3 free this month
Startups

Is AI in a bubble? Menlo Ventures’ Deedy Das on Anthropic’s IPO and Europe

Venture capital is shifting from growth-at-all-costs to a focus on sustainable unit economics as the AI landscape matures beyond the initial hype cycle.

Originally on Sifted →
AB

Adrian Boysel

Contributor

Nov 19, 2026

5 min read

Photo illustration / STKR News

We are currently standing at a strange crossroads in the technology sector. On one hand, you have the massive, gravity-defying valuations of the foundational model companies. On the other, you have a growing chorus of skeptics asking where the actual revenue is. Deedy Das from Menlo Ventures recently shared some perspectives on this dynamic, particularly regarding Anthropic and the broader European ecosystem, that deserve a closer look from the builder's perspective.

The Valuation Disconnect

For founders, the numbers we see in the headlines often feel disconnected from the reality of building a product. When a company like Anthropic is discussed in the context of an IPO, the conversation usually centers on their massive compute costs and their relationship with cloud providers like Amazon and Google. But the real story for the rest of us is about the trickle-down effect of these valuations.

High valuations for the 'god-mode' AI companies create a massive vacuum. They suck up the available capital, making it harder for early-stage application layer startups to get a look in unless they can prove immediate utility. Das points out that while the bubble talk is loud, the underlying utility of these models is real. This isn't 2021 where we were trading digital monkey pictures; we are trading tokens that actually perform labor. The question isn't whether the tech works, it's whether the business models around the tech can survive the cost of goods sold.

The European Arbitrage

One of the more interesting takeaways from the current VC sentiment is the renewed focus on Europe. For a long time, the Silicon Valley consensus was that Europe was great for engineering talent but terrible for building massive companies due to regulation and a lack of late-stage capital. That narrative is shifting.

Builders in Europe often operate with a level of capital efficiency that American founders have forgotten. When capital is scarce, you don't hire a hundred engineers to solve a ten-engineer problem. You focus on the product. As Menlo and other major US firms look toward Europe, they aren't just looking for cheaper labor; they are looking for founders who have been forced to build sustainable businesses from day one because they didn't have the luxury of a 100x revenue multiple valuation.

The Application Layer Struggle

If you are building in the AI space right now, you are likely feeling the 'middle-child' syndrome. You aren't a foundational model, and you aren't a legacy SaaS company adding a wrapper. You are trying to find a new way to deliver value. The skepticism voiced by seasoned investors like Das suggests that the 'wrapper' era is officially over. Investors are no longer funding companies that simply provide a better UI for GPT-4.

What builders need to realize is that the value has moved to the data flywheels. If your application doesn't generate a proprietary data loop that makes the model better or the workflow more integrated, you don't have a moat. You have a feature that OpenAI or Anthropic will eventually release for free. The focus is shifting toward vertical AI—tools built for specific industries like law, medicine, or specialized engineering—where the general-purpose models lack the nuance to compete.

The IPO Horizon and Exit Liquidity

The talk of an Anthropic IPO is a signal to the market. We need an exit. The VC ecosystem has been an entry-only party for the last few years, and the pressure is mounting for these massive paper valuations to turn into liquid cash. However, for the average founder, an Anthropic IPO is a double-edged sword. While it validates the sector, it also sets a benchmark. If Anthropic goes public and the stock stalls because the margins are too thin, the 'AI premium' for every other startup will vanish overnight.

We have to be honest about the margins. Running these models is expensive. If you are a builder, your primary goal right now should be reducing your inference costs. Optimization is the new innovation. The founders who win the next two years won't be the ones with the flashiest demos; they will be the ones who figured out how to provide the same value using a smaller, cheaper, fine-tuned model instead of a massive frontier model.

A Founder's Reality Check

It is easy to get caught up in the macro-talk of bubbles and venture cycles. But the ground-level reality is that we are in a period of intense 'weeding out.' The tourists are leaving. The people who were here for the quick flip are realizing that AI is actually very hard to build and even harder to sell.

  • Focus on Unit Economics: If you lose money on every API call, you don't have a business; you have an expensive hobby.
  • Geography is Secondary to Talent: Whether you are in Berlin or San Francisco, the ability to recruit engineers who understand latency and model optimization is your biggest asset.
  • Don't Wait for the IPO: Don't build your roadmap based on when the 'big guys' go public. Build for the customers who are willing to pay you today.
The real risk isn't that AI is a bubble; the risk is that you spend two years building something that is a line of code in someone else's update.

The skepticism we are seeing from firms like Menlo is healthy. It's a return to fundamentals. They are looking for companies that can survive a 'nuclear winter' where capital isn't free. As a founder, that should be your North Star. If you can build a profitable, useful tool in this environment, you'll be unstoppable when the cycle turns back to growth.

What This Means for the Future

We are likely to see a bifurcation in the market. There will be the 'Utility AI' sector—companies solving real problems with boring, efficient tech—and the 'Frontier AI' sector—the high-stakes gamble on AGI. Most builders belong in the former. It might not get you the same headlines as a billion-dollar round, but it will get you a sustainable company that actually owns its market.

The takeaway here is simple: Ignore the noise about IPOs and bubbles. Focus on the cost per token and the value per customer. If those two numbers move in the right direction, the rest of the venture capital drama doesn't matter.


Read the original at Sifted →

The Brief

Stay Updated on Cutting-Edge Tech

A six-minute morning dispatch on the markets and the technology shaping them.

Free. No spam. Unsubscribe anytime.

Write for STKR

Become a Contributor

Earn $STKR for published stories on markets, protocols, and culture.

  • Earn $STKR for every published piece
  • Editorial support from the STKR desk
  • Byline visibility across the network
  • First look at the upcoming creator program
Apply to Write

Keep reading

All stories

Comments

24 reader responses