16 Changes to the Way Enterprises Are Building and Buying Generative AI
?? Luis Herrera ??
The problem with computers is all they can do is provide answers (Picasso)
Andreessen Horowitz just laid out a vision for the future of generative AI in the enterprise, and it’s clear: the game is changing fast. Here are some of the big surprises:
1. Budgets Are Ballooning
Companies spent about $7 million each on generative AI in 2023. Now? They’re doubling, even quintupling that for 2024. Why? Because they’re betting big on AI moving from experimental to essential. Production-ready. Ready to deliver real results.
2. Recurring Budgets Are Here
No more siphoning funds from “innovation” budgets. Generative AI is graduating into core software spend. Companies see it as a strategic pillar now, not a side project. For some, it’s about automating parts of the business (think: customer service) to cut costs while aiming for scale.
3. Open Source Wins Favor
Closed systems? No longer the default. Nearly half of enterprises now prefer open-source models, shifting away from last year’s proprietary-heavy mindset. Why? Control, customization, and security. Open source gives them the flexibility to shape AI to fit their unique needs.
4. Multi-Model Strategies
Enterprises aren’t picking just one AI model and hoping it fits all. They’re getting smart—choosing multiple models to balance performance, cost, and adaptability. It’s about agility, avoiding lock-in, and staying ready for whatever advancements come next. In a fast-moving field, flexibility wins.
5. Control and Customization: Non-Negotiable
Enterprises want control over their models—not just to ensure security, but to truly understand what their AI is doing. Fine-tuning instead of building from scratch keeps things manageable while letting them adapt AI to their specific needs.
In short, this isn’t a casual dip into AI anymore; it’s a leap. Enterprises are shifting to more integrated, controlled, and customized AI strategies that don’t just keep up with change but shape it.