The Race to AGI: What It Actually Means for Your Business

Almost every week, a CEO asks me some version of the same question: "If OpenAI, Google, and the others reach AGI in a few years, should we even bother investing in AI now?"

It's a reasonable question. The race to build artificial general intelligence is the most expensive technology contest in history, and the headlines make it sound like everything will change overnight. But for most businesses, I think the race matters for different reasons than they expect.

The scale of the race

The numbers are hard to grasp. Amazon, Microsoft, Alphabet, and Meta together plan to spend roughly $700 billion or more on capital expenditure in 2026, most of it on AI data centres and chips. That's up about 77% from 2025, and analysts expect it to pass $1 trillion in 2027.

OpenAI's Stargate programme, announced in January 2025 as a $500 billion venture, has grown into the umbrella for all of its data centre plans. OpenAI says it has secured around 10 gigawatts of computing capacity, although much of that is contracted rather than running yet.

My observation: Whatever you think about AGI, this level of investment guarantees one thing: the AI tools available to your business will keep getting more capable, quickly.

Nobody agrees on what AGI is

OpenAI's charter defines AGI as highly autonomous systems that outperform humans at most economically valuable work. But even the companies closest to it don't share a definition. When Microsoft and OpenAI restructured their partnership in October 2025, they agreed an independent expert panel would verify any AGI claim, without publicly explaining how. By April 2026, reports suggested the AGI clause in their contract had effectively been written out.

My observation: AGI is as much a business and contractual term as a technical one. I'd be cautious about planning your strategy around a finish line that the runners themselves can't define.

Waiting for AGI is the wrong strategy

The models available today can already do far more than most companies use them for. In my work, the bottleneck is almost never model capability. It's data, workflows, skills, and change management.

My observation: Companies that wait for a more powerful model will find the same bottlenecks waiting for them. Those that build the foundations now will be ready to use each new capability as it arrives.

What the race actually means for your business

1. Capabilities will jump, often. New models arrive every few months with better reasoning, coding, and ability to act. Systems should be designed so the underlying model can be swapped without rebuilding everything.

2. Prices will move in both directions. Last year's capability gets cheaper. The newest models often cost more. Rising chip and memory costs are already showing up in the providers' own spending.

3. Power is concentrating. A handful of companies control the leading models, the chips, and the data centres. That makes vendor dependency a real strategic risk.

4. Your advantage is what they can't copy. Labs racing towards AGI will make general intelligence cheaper for everyone. What they can't commoditise is your proprietary data, your customer relationships, your processes, and your industry knowledge.

My observation: When intelligence becomes a commodity, the companies that win are the ones that know exactly where to apply it.

The risks worth watching

This race also carries risks for businesses that depend on it. Spending is running far ahead of AI revenue, which has led to real debate about a bubble. Some providers are heavily dependent on one another financially. Energy and chip supply are genuine constraints. And regulation is still catching up with what the technology can do.

My observation: I advise clients to treat their AI vendors the way they'd treat any critical supplier: check their financial health, keep contracts flexible, and always have a fallback.

What I've seen work

1. Don't wait. Build on today's capabilities and redesign workflows now.
2. Stay model-agnostic so you can switch as better or cheaper options appear.
3. Avoid single-vendor lock-in for anything critical.
4. Invest in what's yours: data, processes, and people.
5. Review your AI roadmap regularly, because what's possible will keep changing.
6. Watch your vendors as closely as they watch the market.

The race to AGI is being run by a few companies with very deep pockets. You don't need to join it to benefit from it. You need to be ready to use what it produces.