Menlo Ventures’ Matt Murphy Says AI Startups Need a New Playbook as Anthropic Surges

Menlo Ventures partner Matt Murphy says the AI startup market is moving so quickly that the old rules of venture investing no longer apply. In a recent TechCrunch Equity podcast discussion, Murphy pointed to Anthropic’s explosive rise — including a reported $47 billion revenue run rate by May, up from $9 billion in 2025 — as evidence that the fastest-growing companies in the sector are being built on a very different model than the last generation of software startups.

Murphy, who led Menlo’s investment in Anthropic’s $500 million Series D, argued that the company’s success was not simply about having a strong model. Instead, he said the bigger advantage came from turning that model into a broader platform through products such as Claude Code, MCP and Claude Skills. He also said startup founders now need to think differently about speed, product scope and what actually creates a durable business.

The comments matter because Anthropic has become one of the clearest examples of how artificial intelligence is compressing startup timelines, reshaping valuations and changing the definition of a competitive moat. They also offer a rare look at how a major venture firm decided to back the company before many investors were ready to do so.

Why Menlo backed Anthropic so early

Murphy said Menlo’s initial decision to support Anthropic came before the company had launched publicly and before it generated revenue, a stage when many investors were still trying to understand whether frontier model companies could become enduring businesses. Menlo ultimately led Anthropic’s $500 million Series D after first betting on the company at a $4 billion pre-revenue valuation.

That early commitment was not easy to fit into Menlo’s traditional fund structure, Murphy explained. Even so, the firm saw enough promise in the team, the technical direction and the opportunity ahead to move forward. In his view, the participation of Google and Amazon as investors was an important signal that the company had begun to attract heavyweight validation.

Murphy described the involvement of major strategic investors as an early sign that Anthropic was gaining real momentum, not just attention from the startup world.

The investment also illustrates how frontier AI changed venture norms. In earlier eras, backers often waited for evidence of product-market fit, revenue and distribution. With Anthropic, Menlo appears to have wagered that the race to build foundational AI systems would reward speed, technical depth and strategic positioning long before a conventional business model was obvious.

What makes Anthropic different from earlier startup waves?

Anthropic’s trajectory stands out because the company appears to have grown at a pace that Murphy says he has never seen in 25 years of investing. He compared its acceleration with the internet boom, the mobile era and the early cloud wave, saying none of those periods produced a comparable ramp.

The company’s growth is important for two reasons. First, it suggests that AI adoption can scale faster than previous software categories because the product is immediately useful to both consumers and businesses. Second, it shows that the market may now reward companies that pair model quality with tightly integrated tools, workflows and developer ecosystems.

Murphy’s core argument is that a great model is not enough on its own. As more companies gain access to increasingly capable models, differentiation has shifted toward productization, distribution and how effectively a startup can become part of a customer’s daily workflow.

The platform shift

Anthropic’s newer products help explain why investors may now see the company as more than a model lab. Claude Code gives developers a coding-focused interface, MCP helps connect systems and tools, and Claude Skills adds task-specific functionality that broadens what the platform can do.

In Murphy’s telling, those layers matter because they transform Anthropic from a company that simply provides a model into one that offers an integrated stack. That stack can create stickiness, expand usage and make the company harder to displace than a standalone model provider.

This is a meaningful distinction for founders elsewhere in AI. Building a model may still be a technical achievement, but converting that model into an ecosystem is increasingly what separates a breakout business from a commodity layer.

How fast are today’s AI startups growing?

Today’s AI startups are growing faster than Murphy says he has seen at any point in his career. He specifically cited companies such as Lovable and Legora as examples of startups moving at a pace that exceeds even the most aggressive growth he has witnessed over 25 years.

That speed is not just a bragging point. It changes everything about company building, from hiring and product releases to fundraising and competitive defense. In markets where adoption can compound in months rather than years, founders have less time to experiment, less room for product delays and more pressure to lock in users before rivals catch up.

The result is a new kind of venture environment, one where the winners can look obvious very quickly but where the distance between first move and lasting dominance may still be uncertain. The speed of AI growth can create outsize valuations, but it can also expose weak execution just as quickly.

Why the model is no longer the only moat

For years, AI startups often pitched the quality of their model as their primary advantage. Murphy’s view suggests that this is no longer enough, especially in a market where foundational capabilities are improving rapidly across the industry.

As more model providers close the gap on raw performance, startups need another layer of differentiation. That can come from developer tools, proprietary workflows, domain-specific products or integrations that make the technology operationally indispensable.

Anthropic’s push into coding and enterprise-adjacent tools shows how that strategy works in practice. The model may attract attention, but the surrounding product ecosystem is what can drive usage, retention and revenue.

This logic also helps explain why investors increasingly talk about application layers, agentic workflows and enterprise systems rather than model benchmarks alone. The question is not just which lab has the smartest model, but which company can turn intelligence into recurring business value.

How founders should think about competition

Founders in AI need to assume that competitors will move quickly, and that users will compare products on workflow utility rather than technical elegance alone. Murphy’s comments imply that speed, distribution and product depth are now central to defensibility.

• Build products that solve a real workflow problem, not just a demo-worthy feature.
• Assume model quality will converge over time and plan for a broader moat.
• Design for usage depth, not only early adoption.
• Move quickly enough to capture market attention before the category fragments.

That does not mean model research is irrelevant. It means research leadership must be paired with product strategy and customer value creation if a startup wants to sustain momentum.

What was the backlash to Mythos?

Murphy also addressed criticism around Anthropic’s Mythos rollout, pushing back on the idea that the launch was primarily a branding exercise dressed up as safety. In the podcast discussion, he defended the broader intent behind the rollout and framed the response as part of the ongoing tension between AI safety messaging and commercial positioning.

The debate is familiar in the AI industry, where nearly every major model company is trying to balance trust, caution and market expansion. Safety language can be genuine, but it can also be viewed skeptically when tied to product announcements and competitive moves.

Murphy’s view suggests that he sees the Mythos discussion as evidence of how closely observers scrutinize frontier AI companies. As these firms become more powerful and more visible, their launches are judged not only on technical merit but also on whether the messaging feels authentic.

Murphy argued that the criticism misses the point if it assumes safety framing and product strategy must be mutually exclusive.

Why this interview matters for the broader AI market

The conversation matters because it captures a shift in how sophisticated investors are assessing AI businesses. The most important question is no longer whether AI can produce a breakthrough product. The question is how fast that product can turn into a platform with real distribution power.

Anthropic is an unusually useful case study because it sits at the center of several major trends at once: frontier model competition, enterprise adoption, investor appetite for pre-revenue companies and the race to build tooling around generative AI.

It also reflects a larger truth about the current cycle. AI startups are no longer only selling possibility. The strongest ones are already showing extreme growth, which forces both founders and investors to think in a compressed timeframe.

That creates opportunity, but it also raises the stakes. When companies can rise from near-zero to multi-billion-dollar scale in a short period, they may also face more scrutiny around execution, safety, monetization and long-term durability.

What founders can learn from Murphy’s view

Founders looking at the Anthropic story may draw several lessons from Murphy’s comments. The first is that the frontier model itself is only the starting point. The second is that investors increasingly want to see product layers that make the technology useful in specific workflows. The third is that speed is now a strategic asset, not merely an operational one.

Another lesson is that companies can win early credibility by showing enough momentum to attract major backers or strategic partners. In Anthropic’s case, the involvement of Google and Amazon helped signal that the company was not just another experimental startup.

Finally, Murphy’s perspective suggests that founders should not assume the market will reward technical excellence in isolation. The winners will likely be the companies that pair strong foundations with clear distribution, practical use cases and a path to sustained customer dependence.

Timeline of Anthropic’s rise

Anthropic’s growth path has moved quickly from early-stage experiment to one of the most closely watched AI companies in the world.

The bigger picture: AI investing after the first wave

Murphy’s comments point to an AI market that has entered a new phase. The first wave was about proving that generative AI could work. The current wave is about deciding which companies can turn that capability into persistent economic value.

That distinction matters for venture capital because it changes how investors evaluate risk. A startup no longer needs to merely promise the future; in some cases, it must show that the future is arriving so quickly that growth itself becomes a competitive weapon.

Anthropic, in Murphy’s view, is one of the clearest examples of that shift. Its pace, platform expansion and investor support make it a benchmark for what the next generation of AI startups may look like — and for how much faster they may move than any previous software company.

For founders, the message is blunt: in AI, the bar is rising fast. It is no longer enough to build a good model and wait for the market to catch up. The companies that stand out will be the ones that turn intelligence into a full product experience, move decisively and scale before the window closes.