First, my thanks to Michael Burnett from Fuel Ventures for flagging Mary Meeker’s latest AI Trends report—a 340-page juggernaut that’s been shaping my thinking about the next decade of automotive investing.

As I dug into it, I was reminded of my caution in “AI won’t solve everything in automotive” (Speculative AI blog). Meeker’s report is bullish on AI’s potential but hints at the messy, complex path ahead, particularly for sectors like automotive.

10-Year Predictions: What’s Coming?

Meeker’s report paints a picture of AI as a compounding force, predicting that:

• Compute costs will keep rising exponentially, with a 360% annual growth rate in training compute for key AI models over the past 15 years. The next decade will continuously increase computing needs, pushing hardware and software innovation.
• Data volumes will dwarf anything we’ve seen before—AI datasets have grown 260% annually for 15 years, and there’s no sign of slowing. This means a relentless push for richer data, more sensors, and deeper analytics for automotive.
• AGI is no longer science fiction—Meeker includes insights from OpenAI and others suggesting AGI (Artificial General Intelligence) could be achievable within a decade, although with massive caveats and societal implications.
• Physical world AI ramps are fast and data-driven, with self-driving vehicles (Waymo, Tesla), automated trucking, and intelligent vehicle platforms like Applied Intuition proving commercial viability in select markets.

These forecasts are staggering, but as I argued in my previous blog, “Speculative AI: DeepSeek vs Silicon Valley,” not everything can—or should—be solved by AI. The automotive sector still hinges on physical infrastructure, regulatory hurdles, and human expertise. Even the best AI will need to coexist with the real-world messiness of roads, drivers, and evolving consumer expectations.

Implications for Automotive Investors

What does this mean for automotive investing over the next decade? Here’s my take:

1. The data premium – As AI’s hunger for data grows, automotive businesses that own, clean, and deploy data well will hold an outsized advantage. Data isn’t just exhaust—it’s an asset class in itself.
2. Compute pressure – Expect higher capex needs for AI-ready hardware, from vehicle sensors to cloud platforms. This could stretch balance sheets and challenge traditional cost structures for OEMs and mobility startups.
3. Talent as the new oil—Meeker echoes Steve Ballmer’s rallying cry for developers, developers, developers. In automotive, that means prioritising AI-native teams rather than just bolting AI on as a feature.
4. AGI’s shadow – While AGI is still a moving target, investors must be mindful of how it could disrupt value chains. For now, focus on tangible AI deployments (autonomy, predictive maintenance, personalised experiences) but be ready to pivot if AGI’s promises become reality.
5. Not everything is an AI problem – As I’ve said before, automotive’s challenges—like supply chain volatility, geopolitics, and shifting consumer trust—often require real-world fixes, not algorithms. The best companies will know when to use AI as a tool, not a panacea.

Final Word

Mary Meeker’s report is a clarion call to stay alert. The next decade of AI will be bigger, faster, and potentially more disruptive than the last. For automotive investors like me, it’s about finding the sweet spot: where AI augments the business but doesn’t become the business.

I’ll leave you with this: AI won’t solve everything in automotive, but will shape everything.

Have a great week!