Beacon acquires AI safety startup Hayes Labs, plans to close a new business every 10 days to make the real economy AI-native

Beacon Software is a holding company that acquires established software businesses serving what it calls the real economy — campgrounds, labor unions, youth sports leagues, construction firms, oil rigs — and rebuilds them as AI-native platforms. The acquisition pace has accelerated: Nilam Ganenthiran says Beacon is now closing a new business every ten days, up from roughly one every three weeks a year ago. The company currently serves 22,000 customers.

“At that time, you know, we were acquiring a business every three weeks or so. Today, we're closing on a new business every ten days. ... I'm Leonard, cofounder and CEO here at Haze Labs. Hayes was an AI safety company focused on red teaming, guardrails observability, and evals, both for the Frontier Labs, including OpenAI and Anthropic, who are our first customers. ... The ramp data from about a quarter ago showed that SMBs are only 20% adopting AI versus, like, 85% for the Fortune 500.”

Hayes Labs acquisition

The headline move is Beacon's acquisition of Hayes Labs, an AI safety company founded by Leonard Tang, who joins as VP of AI Research. Hayes built red teaming, guardrails, observability, and evaluation capabilities for frontier labs — with OpenAI and Anthropic as early customers — and then expanded into large enterprises including banks, Fortune 500s, insurance companies, and healthcare firms.

The strategic logic is that trust, not access, is the primary reason AI adoption among SMBs remains low. Ramp data cited by Ganenthiran puts SMB AI adoption at roughly 20%, versus 85% for Fortune 500 companies. Beacon's existing businesses already have the customer relationships; Hayes brings the safety infrastructure to make AI deployment credible to those customers.

Tang frames it as an alignment argument too. Rather than letting frontier labs define what values AI models should be optimized for, Beacon and Hayes want to align those values to the businesses that have historically been underserved by Silicon Valley — and to their customers.

What AI is actually doing in these businesses

Ganenthiran gives a concrete illustration from the campground vertical, where private owners across North America reportedly spend 40% of their time on payroll, compliance, and manual spot assignment. Beacon has built tools inside existing software to match guests to campground spots, run dynamic pricing by season, and load-balance staffing against occupancy. A Fortune 500 construction firm (unnamed) is saving millions on tool supply after Beacon deployed AI matching and GPS tracking to route industrial tools to the right job sites, replacing a culture of over-ordering.

Ganenthiran pushes back on the cost-cutting framing explicitly: across Beacon's portfolio, headcount is up, not down, as AI-driven growth has expanded what these businesses can take on.

What comes next

Tang points to two distinct near-term bets. The first is AI employees — SDRs, support engineers, and similar roles dispatched on demand to lever up existing owner-operator teams. The second is what he calls company world models: models built on a business's own data that help an owner anticipate what happens next and make better strategic decisions, not just execute existing ones. Both are framed as products Beacon is positioned to build because of its depth inside these businesses, and the robotics question — whether physical hardware eventually gets layered in on top of software — is left open as a five-year possibility rather than a current initiative.