Anthropic, Google, Microsoft, OpenAI, xAI, and others are making trillion-dollar bets on AI infrastructure. They promise artificial general intelligence (AGI) delivered from the cloud, which they describe as “a country of geniuses in a datacenter”. Tokens are the “currency of intelligence.” Unlimited intelligence — coming soon.

But here’s what that narrative obscures for enterprise leaders: Your ROI from AI depends almost entirely on something you’re not investing in — your organization’s capacity to reinvent itself. Worse, the AGI story is actively dangerous. It suggests your employees are the problem to be automated away, when in fact they’re the only solution you have.

The AI infrastructure bubble is beside the point — there’s a deeper issue. The AI industry narrative doesn’t just inflate valuations; it distracts, devalues, and distorts the human-centered transformation work that enterprises actually need to do.

The Distraction

Generative AI (genAI) and agentic AI are powerful technologies — they’re delivering unprecedented value at the individual task level and represent the fastest technology adoption in human history. But scaling from task-level value to enterprise value requires workflow and business process reinvention. That’s hard, messy, and human work, which involves challenging comfortable processes, experimenting and tinkering, and building new muscle for continuous change.

The AI industry narrative says, “Wait. More intelligence is coming. Bigger models. Smarter agents. Just hold on.” The truth is that you don’t need to wait for some future version of AI to transform. You need courage, focus, and people willing to reinvent how work gets done. The noise around AI creates an excuse for inaction: “Let’s wait and see what’s possible” becomes the enemy of “Let’s transform what we have right now.” The trillion-dollar infrastructure race, artificial superintelligence timelines, and existential debates all obscure the real work.

The Devaluation

Here’s the script from Silicon Valley: “Support us. Invest trillions in infrastructure. We’re building intelligence that will replace your workforce.”

This is not just wrong — it’s insane. All that infrastructure rests on a foundation that doesn’t exist yet — your organization’s capability to reinvent itself and your employees’ willingness to drive that reinvention. Without those, the trillion-dollar datacenters are built on sand. Your employees aren’t a problem to be automated away. They’re the reinventors your organization desperately needs. They understand which workflows are broken, where processes create friction, what customers need, and how work really gets done vs. how the org chart says it gets done.

The companies winning with AI today aren’t planning for employee replacement — they’re empowering employees to reinvent. DBS Bank built a customer-centric innovation culture to drive digital transformation a decade ago. Westpac runs an “AI Shark Tank” program today, recognizing teams that foster innovation and creativity, particularly in the use of AI to improve customer and employee experiences. These organizations anchor and incentivize employees to challenge the status quo with these technologies.

Here’s the irony: The grand vision of cheap, unlimited intelligence promises to replace the very employees who are essential to making AI valuable. The AGI narrative devalues the people who hold the key to ROI. It positions them as costs to eliminate rather than capabilities to unleash.

The Distortion

Do frontier AI capabilities need continued investment and development? Yes. Do they require hundreds of billions or trillions in infrastructure investments to chase AGI dreams? That’s a separate conversation, driven by nationalistic competition, economic posturing, and Silicon Valley boardroom ambitions. It’s theater on a massive scale. For enterprises, this creates unnecessary noise: confusion about what’s possible now vs. what might be possible later, paralysis around whether to wait for smarter models, and misaligned expectations that AI will solve problems automatically.

Don’t lose sight of the fact that today’s genAI and agentic AI capabilities are already powerful enough to fundamentally alter workflows, so long as organizations are willing invest in the human side of transformation. It’s inevitable that the technology will get better and cheaper — cloud computing followed the same path. Your organization developing the muscle to reinvent itself continuously isn’t inevitable.

What Should You Do? 

GenAI and agentic AI can improve workflows today. Focus on what’s real, deployed, and improvable — don’t hold off for the distant possibility of AGI before taking action. Start by:

  • Investing in your reinventors. In every transformation, technology is the tool and your people are the advantage. Build authority structures that enable experimentation; create recognition systems for workflow transformation, not just efficiency; and protect and celebrate (responsible) tinkering.
  • Building organizational muscle for continuous reinvention. This is digital transformation 2.0. Most capabilities are now digitized, so it’s time to reinvent workflows. Start with controllable processes where experimentation is safe; scale transformation capabilities before increasing technology spend; and develop cautious urgency — move on culture now and scale tech as it matures.
  • Tuning out the AGI noise. AGI timelines don’t matter for your 2026–2027 strategy — the infrastructure race is someone else’s game. Your competitive advantage comes from transformation capability, not compute capacity.

The question isn’t whether/when AGI will be realized — it’s whether your organization will be ready to reinvent itself with the technologies we already have. Start building your transformation capability, because all the intelligence in the world means nothing without reinventors willing to use it.