AI models change every 42 days. Buyers cannot assemble a best-of-breed stack at that pace, and the vendors know it. Harvey expanded from law firms into professional services, co-building a Tax AI model with PwC across 25 jurisdictions. Glean moved from enterprise search to vertical solutions for healthcare, financial services, and government. ElevenLabs went from text-to-speech to full voice agents for customer service and audiobooks. The SaaS playbook rewarded owning one pain point. The AI playbook is rewarding something else entirely.
Foundation model companies are executing the same pivot. OpenAI launched a dedicated Healthcare and Life Sciences vertical with industry-specific sales teams and solutions engineers. Anthropic built an Industries organization with account executives covering healthcare, insurance, and federal markets. Neither company is selling APIs anymore. They are selling three-to-five year platform commitments to buyers who cannot afford to keep re-integrating.
The structural reason to read Tunguz in full is the compounding logic he surfaces: once an AI system is integrated, it sees how teams operate, captures workflows, and builds more systems on top of them. Falling software development costs mean trusted platforms with broad adoption can expand faster than any point solution. Bundling is not nostalgia for the pre-SaaS era. It is the logical endgame of a market where switching costs are measured in organizational memory, not just contract value.
[READ ORIGINAL →]