Claude in Excel Is a Declaration of War on AI+X Startups

This week, Anthropic released Claude in Excel - and it sent a chill down my spine. Not because of what it does, but because of how it does it. The gap between Microsoft’s approach and Anthropic’s approach to AI in spreadsheets is not incremental. It is architectural.

Microsoft uses AI to execute Excel functions. Claude uses Excel as a presentation layer.

The result? Anthropic wins on both speed and quality. And that distinction - between bolting AI onto existing software and building AI-native from the ground up - is creating a gap far wider than most people realize.

”SaaS Is Dead” Is Becoming Reality

The phrase has been floating around for a year. Now it has teeth. Simply improving an existing service is no longer enough to survive. To have any shot in this market, you need to dominate at least one of three layers:

• Framework dominance: AI operating tools like Opencode and Clawdbot that define how agents work
• Infrastructure dominance: Foundational services like Vercel, Supabase, Cloudflare, and Stripe that everything else is built on
• Full-stack integration: Platforms like Base44, Replit, Lovable, Cursor, Comet, and Manus that combine the entire chain

If you don’t own any of these layers, your only remaining lever is price. And a price war in AI products is not sustainable - costs scale with usage. More users mean higher fixed costs, not lower marginal ones. This is not a game you can win by being cheaper.

Companies With Strong AI Models Grow at a Different Speed

Kuaishou, the company behind Kling, is generating $20 million in monthly recurring revenue. Google’s market share surged after releasing Nano Banana Pro and Gemini-3. These are not incremental gains. They are step-function jumps driven by model capability.

Then there is Grok. Largely overlooked in some markets, Grok’s penetration speed actually outpaces Google’s. The reason is uncomfortable: xAI released world-class image and video generation models with virtually no restrictions - an aggressive play that traded safety guardrails for explosive adoption.

The common thread is clear. The companies growing fastest are the ones that own the model. Most startups do not, and building one is out of reach for nearly all of them.

Why You Need to Burn More Tokens - And Why It Matters

“Why are we spending so much on tokens?” If you manage a team, you have heard this question. Honestly, until early last year, I did not have a great answer either.

Then Claude Code changed everything. Watching it work - faster than most humans, producing output that rivals experienced engineers - made the value proposition undeniable. The reason to encourage token usage became clear: you cannot understand what AI-native means unless you use AI intensively yourself.

That lived experience is exactly where early-stage AI startups should be looking for direction:

• Deep enterprise engagement: Deliver CS-level support quality that locks in business customers
• Relationship data collection: Gather the web and offline relationship data that is invisible at first glance, using AI to structure and prepare it
• Niche dominance: Find a narrow vertical with proven demand and grow fast within it
• AI-native product design: Build something where AI is not a feature - it is the architecture

2026 Is Running Out of Time for Early-Stage Startups

One and a half months after this time last year, Manus launched. That company was later acquired by Meta for roughly $3 billion. This year, something even more disruptive is coming. Anthropic, Kuaishou, Vercel, Supabase, and Cloudflare are already preparing their next moves.

The winners of the AI era will not be the companies that build better features. They will be the companies that redefine how AI itself operates.

What AI-native service will emerge in 2026? And what are you building to be ready for it?