You know something fundamental is shifting in tech when the drama between startups and their AI providers becomes more entertaining than reality television. That's exactly where we found ourselves in this conversation with Foundation Capital partners Ashu Garg and Jaya Gupta.

"This is better than Love Island. I love this shit. Like the gossip and the intricacies of like the people itself. It's amazing."

Manny wasn't wrong. The Cursor pricing saga - where a startup's infrastructure provider started building a competing product, leading to poached PMs, revoked discounts, and user revolts - is just one symptom of a much larger transformation happening in software.

The Death of SaaS (As We Know It)

The conversation started with what might be the most sobering statistic for anyone running a SaaS company today. According to Ashu, the middle market is getting absolutely crushed:

"Mid-sized SaaS companies are struggling. Nine out of 10 are seeing some churn, but the churn isn't dramatic yet. They're seeing employee attrition. They're fighting a war of feature by feature. They're trying to add AI pixie dust here and there. But net-net, they're all struggling."

The companies he's talking about - those between $100M and $1B in revenue - find themselves in an impossible position. They're too small to acquire their way out of trouble like the giants can, but too big and established to pivot quickly like startups.

Ashu pointed to Outreach as a prime example: "Incredible company for a decade. Look at the numbers today. It's flat, maybe marginally declining."

Meanwhile, tiny AI-native startups are "growing like crazy" from the bottom up. The bit hasn't flipped yet, but Ashu thinks we're not far from a tipping point where customers abandon their incumbent platforms en masse.

The $20 Million First Customer

Perhaps nothing illustrates the speed of this transformation better than Ashu's revelation about deal sizes in the AI era:

"I funded a company earlier this year. Their first deal, which hasn't been signed yet - knock on wood - but the first customer is likely to give them a $20 million plus TCV deal. When's the last time you saw a seed stage company get a first customer at $20 million plus?"

This isn't normal SaaS growth. This is a company jumping from seed stage to Series E valuations in a single deal. The company in question? They're migrating legacy SAP and Oracle code using AI. When you're solving billion-dollar problems with AI, apparently the old rules about gradual revenue growth simply don't apply.

Why Experience Became a Liability

One of the most controversial takes came from Jaya, who argued that in AI, youth beats experience:

"AI is new for everyone. Like, if anyone can predict what's happening in six months, I would call that bullshit. No one knows what's happening. You are seeing in this market a ton of younger founders even outpace second time, third time founders that have built unicorn companies."

The logic is simple but profound: everyone started learning AI at roughly the same time. But younger founders have less to unlearn, move faster, and are using AI itself to build their companies more efficiently. As Jaya put it, "Knowledge has been quickly democratized."

This might explain why Foundation Capital has such an unusual approach to evaluating companies...

"If You Have Revenue, Don't Call Me"

In perhaps the most counterintuitive investment philosophy you'll hear, Ashu actively avoids companies with revenue:

"Even though we don't really invest in companies with revenues, in fact, I always tell people, if you have revenues, don't call me. I'd rather not deal with messy revenues. I want to deal with big ideas."

This isn't just contrarianism. Ashu argues that early revenue often constrains vision and forces founders to serve existing customers rather than reimagining entire categories. Even without revenue, he looks for other forms of traction: Who are the early customers you're talking to? Which engineers are you recruiting? If you're building in sales tech and haven't talked to Manny, "I'm not funding you."

The 500 Agent Future

The partners saved their boldest prediction for last. Forget building another point solution or feature company. The future belongs to companies that think bigger:

"The world of AI apps and AI agents is about 500 agents from one company replacing 500 feature companies. You've got to think broad, you've got to think big, and you've got to execute like crazy to see which agent works, because very often one or two agents doesn't solve anything for a customer. They need enough of these agents to really move the meter."

This runs counter to everything VCs have preached about focus for the last decade. But in Ashu's view, the narrow AI startup is already dead. Customers don't want to manage hundreds of point solutions anymore - they want comprehensive agent armies that actually move the needle.

The Plot Twists Keep Coming

Beyond these major themes, the conversation was peppered with surprising predictions and hot takes:

On the AI giants: Despite their massive valuations and growth, Ashu is "very skeptical on both OpenAI and Anthropic." He believes that "when the dust settles, it's not clear that the winners in that category will be either of the two companies."

On pricing models: While everyone talks about outcome-based pricing, Jaya thinks we'll see an evolution through usage-based and workflow-based models first. True outcome-based pricing remains elusive because, as she notes, "the outcome is actually just a function of the customer's product as well, not just your software."

On commitment issues: Both partners openly admitted to having "commitment issues" when it comes to investing, preferring to "date" founders for four to five months while gathering what Jaya calls "observability data" on how they think and learn.

What This Means for Founders

If you're building in AI right now, the message is clear but daunting. The playbook that worked for SaaS won't work here. As Ashu put it:

"A lot of the lessons that you and I learned over the last few decades of software apply, but a lot more don't. Knowing when to break the mold and reinvent and reimagine how you do things, I think is a big part of winning in the AI space today."

For Manny, who built Outreach into a unicorn, this resonates deeply. He's now tackling the problem of monetization and margin management for AI agents - the very issue that was "the bane of his existence" at Outreach. Sometimes you do "irrational things at irrational times," as he puts it.

But in a world where the drama is better than Love Island and first customers write $20 million checks, maybe irrational is exactly what we need.

Companies & Products Mentioned

Foundation Capital
OpenAI
Anthropic
Cursor
Windsurf
Claude (and Claude Code)
11X
Harvey
Scribe
Lovable
Tenor
Fulcrum
Outreach
Salesloft
Databricks
Salesforce
Oracle
SAP
Excel
Wix
McKinsey

See Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.

S2E16: AI is better than Love Island | Ashu Garg and Jaya Gupta (Foundation Capital)

Get Paid with Manny Medina

S2E16: AI is better than Love Island | Ashu Garg and Jaya Gupta (Foundation Capital)