“It’s venture money, not adventure money.” This insightful comment from a venture capitalist (VC) captures a critical truth: in the midst of a technology hype cycle, caution often gives way to urgency. VCs must deploy their capital, and the fear of missing out on significant opportunities overshadows the risks of potential failures, especially when competitors are also eager to invest.
This dynamic is mirrored in many companies today, particularly with the current excitement around AI. Whether it’s large language models (LLMs) or machine learning (ML), any project can be quickly rebranded as AI to secure funding—even those sidelined in previous years.
Are You Prepared for AI?
As billions are set to flow into AI over the next decade, caution is warranted. History shows us that while technologies like search, social media, and mobile have had a profound and lasting impact, others, like virtual reality (VR) and cryptocurrency, have faced limitations.
Five years ago, headlines buzzed with predictions about AI's potential. Today, companies are scrambling to showcase their AI expenditures and claims of transformative capabilities. This scattergun investment approach produces a mix of substantial successes and numerous failures. Similarly, company leadership may endorse AI initiatives driven by misplaced optimism rather than genuine potential.
Despite this landscape, there’s no denying the revolutionary role of LLMs, as illustrated by ChatGPT reaching 100 million users faster than many transformative technologies.
So, how do you prioritize your investments in AI to maximize returns and minimize waste? Focus on these three critical questions to eliminate 80% of unnecessary spending:
1. Understand Total Cost Over Time
Before approving new AI projects, assess the initial and ongoing resource costs. For every 10 hours of your data science team’s work, there can be five times as much hidden engineering, DevOps, and support time. Many promising initiatives falter due to a lack of ongoing investment. While it’s tempting to say yes to every AI project, overcommitting can drain resources that could fund genuinely impactful initiatives.
Furthermore, consider the increasing marginal costs associated with AI. Large models demand significant investment to train and maintain, and failure to deliver promised functionalities can lead to customer dissatisfaction and reputational damage—experiences seen with Google and IBM’s early AI endeavors.
2. Ask Why Can't Anyone Else Do This?
Remember that textbook lessons about commoditization are invaluable. I learned this firsthand while designing memory chips at Micron, where brand recognition mattered less than price—an essential trait of commodity products. The tech industry often operates on two levels: monopolies and commodities. Before embracing an AI initiative, ask, “What unique advantage do we have?” Projects likely to become commoditized without a scale advantage are risky. Instead, focus on initiatives that build defensive advantages, whether through data exclusivity, proprietary insights, or strong network effects.
3. Make a Few Bets You’re Willing to Pursue
The best investment opportunities often enhance your existing business model. As the BASF slogan suggests, “We don’t make the things you buy; we make the things you buy better.” If AI can improve your current products, that’s a compelling investment. The next best opportunities involve expanding your offerings into new areas or adapting to shifts in the value chain.
However, the most crucial bets challenge you to innovate at the risk of disrupting your current business model—if you don’t tackle this challenge, your competitors will. Focus on a small number of strategic initiatives and commit to their success, leaving the rest to VCs and startups.
While the excitement around AI is warranted, history teaches that hype can lead to both promising discoveries and significant waste. By following these guidelines, you can position your investments for optimal returns and meaningful impact.
Ranking
