Tessera's 6 Engineers Just Replaced 60 Consultants. Nova Hired 40.

Tessera Labs needed just six engineers to replace 60 SAP consultants at their first enterprise client. Nova Intelligence hired 40 people before landing their first customer.

Both companies raised massive rounds in May 2026 to tackle the same $89 billion SAP modernization market. Both have ex-Google AI founders. Both claim their AI agents will transform enterprise software. But their team structures could not be more different.

The numbers tell the story. Tessera raised $60 million from Andreessen Horowitz with a team that fits in a minivan. Nova raised $31.5 million (and another round we're tracking) with a headcount approaching a traditional consultancy.

The Tessera Model: Ship First, Hire Later

Kabir Nagrecha runs Tessera like a Special Forces unit. His six-person AI team operates on a simple principle: every engineer owns a full vertical of the product. No product managers. No scrum masters. No layers.

"We hired people who've built entire products solo," Nagrecha told investors during the raise. The approach paid off immediately. Their multi-agent system replaced a 60-person SAP implementation team at their pilot customer, cutting project timelines in half.

The team structure breaks every conventional rule:

• Engineers talk directly to customers
• Everyone deploys to production
• No meetings longer than 15 minutes
• Code reviews happen in pairs, not committees

Andreessen Horowitz partner Seema Amble backed the approach with conviction. She's seen hundreds of AI startups hire aggressively post-funding only to burn through runway building infrastructure instead of shipping product.

The Nova Model: Research Lab Meets Consultancy

Emma Qian took the opposite approach at Nova Intelligence. The former Google DeepMind researcher assembled what looks more like an academic department than a startup. Forty employees before product-market fit.

Nova's hiring spree reflects a different philosophy. They believe SAP modernization requires deep domain expertise across industries. Their team includes:

• 12 PhD researchers from top AI labs
• 8 former SAP consultants from Big Four firms
• 15 engineers building the platform
• 5 business development executives

The pitch resonated with investors. Accel led their round, betting that enterprise AI needs subject matter experts, not just great engineers.

The $89 Billion Reality Check

Fortune reports that 77% of the world's transactions touch an SAP system. The modernization opportunity is massive. But which team structure actually works?

Early signals favor Tessera's approach. They've shipped production systems to three Fortune 500 companies since January. Their lean team means they can undercut traditional consulting fees by 80% while maintaining 70% gross margins.

Nova has published impressive research papers and built partnerships with systems integrators. But they haven't announced a single production deployment. Their burn rate exceeds $2 million monthly.

The Pattern Emerging Across AI

This isn't just about two companies. It's about two competing theories of how to build AI businesses.

The Tessera model mirrors what we're seeing at companies like Perplexity (12 engineers competing with Google) and Midjourney (11 people generating $200M in revenue). Small teams with extreme ownership shipping fast.

The Nova model follows the Anthropic and Cohere playbook. Hire top researchers, build foundational models, worry about customers later.

For operators building AI teams, the lesson is stark. In consumer AI, research depth might matter. In enterprise AI, shipping velocity wins. SAP customers don't care about your Nature papers. They care whether your agents work on Monday morning.

The irony? Both companies are trying to replace armies of consultants. Tessera figured out how to avoid becoming one.