The author of this article argues that the AI companies that prove to be the most successful won't necessarily be the newest, but those pursuing an enduring need.

The following article comes from Liz Harrow (pictured below), who is a partner at Shakti VC, an inception stage venture capital firm investing in the next generation of AI and space commerce businesses. The editors are pleased to share this content about topics that are, as we have seen with the SpaceX record-breaking IPO and the ascent of major AI breakthroughs such as ChatGPT and its rivals, very much in the public eye.

Liz Harrow

According to the Organisation for Economic Cooperation and Development in 2025, VC investments in AI firms globally made up more than half (61 per cent, $258.7 billion of the total), doubling its 2022 share from 30 per cent. “Mega deals” – those valued at more than $100 – account for 73 per cent of all transactions. VC had languished in the aftermath of Covid when global interest rates spiked to curb inflation pressures. The rise of AI as a dominant tech investment theme was a reason, among others, for a subsequent rebound in VC fundraising and activity. (See a related article here.)

The standard editorial disclaimers apply to views of guest contributors. To comment and provide ideas, email tom.burroughes@wealthbriefing.com and amanda.cheesley@clearviewpublishing.com

We are lucky to speak with many smart, thoughtful LPs, and recently the same three questions have come up from investors, both family offices and institutions. 

They are good questions, and are likely to be on the minds of others too. So we wanted to share how we think about them.

1. Did we miss AI?

With all the hype around the first expected LLM IPOs, it is a fair question. And honestly, some days it can be hard to imagine what else is left to build.

But we have seen this pattern before, and Amazon’s IPO in 1997 is a case in point.

When Amazon went public in 1997, many believed that e-commerce would be a winner-takes-all market. But Amazon was just the first wave. It helped build the infrastructure and consumer behaviour that unlocked an entire generation of e-commerce winners, including eBay in 1998, Overstock in 2002, Blue Nile in 2004, and later Shopify and Etsy, both in 2015.

Even Jeff Bezos later reflected, “In retrospect, we significantly underestimated how much time would be available to enter these categories.”

The good news is that the large LLMs have laid the railroad tracks. They have built much of the foundational infrastructure for AI. Now the trains can run, and AI can begin to reimagine every large industry.

So no, we do not believe that you missed the boat. We are certainly not closing up shop at SHAKTI. In fact, we believe this is just the beginning, and we are ready for the next wave of category leaders.

For more on why we are excited about AI 3.0, check out our post on the Imagination Era.

 

2. Won’t Anthropic and OpenAI just do everything?

Also a very good question, especially given how deep the pockets of these companies are.

My partner Keval Desai has a particularly useful perspective on this from his time at Google in the years before and after its IPO. Google tried to enter many categories, and rightfully so. Some of those efforts included Google Health, Google Answers, a “knowledge marketplace,” and Google Finance. None of them became category leaders.

The lesson is that entering a new business line is not as easy as adding a new product. It often requires different salespeople, product teams, expertise, and operating models. Eventually, the experiment may no longer be worth the effort.

Second, customers do not necessarily want everything to come from the same company. In fact, they often prefer to minimise risk by working with several suppliers across their technology stack. That naturally creates room for more vendors.

The takeaway: owning the horizontal infrastructure layer does not automatically give a company the domain expertise, workflow empathy, or user trust required to win within a specific vertical.

Will the big companies try to do everything? Probably. But if history rhymes, there will still be plenty of opportunity for application layer companies to build enduring businesses.
 

3. Can a company founded before ChatGPT still win in AI?

Companies that are not “AI native” can absolutely win in AI. But we believe they need two ingredients.

The first is an enduring use case. We call this a “toothbrush use case,” meaning something ubiquitous and frequent. Without the right use case, the technology itself is irrelevant.

The second is the founder. Is that person able to reimagine the company using the latest technology in pursuit of that enduring use case?

Two examples from within our own portfolio illustrate this well.

Canva, which was founded in 2013, nearly a decade before ChatGPT. The company was recently ranked fourth among gen AI mobile apps by monthly active users, according to Andreessen Horowitz. Yet its mission to “empower the whole world to design” has not changed.

As Melanie Perkins, our Shakti Titan, has said, “AI has been a really important part of that for us for many years.”

The second example is Gatik AI, a leader in autonomous trucking in North America. We first invested in Gatik in 2019, several years before ChatGPT. Its core use case has remained constant: moving freight autonomously and at scale. Gatik was founded before the current AI wave, but it has continued to evolve its technology stack into a scalable and interpretable AI system that is purpose-built for safe freight movement. And recently, Gatik was featured in The Wall Street Journal for its milestone partnership with Pepsi.

The companies that win will not necessarily be the ones founded most recently. They will be the ones pursuing an enduring need, led by founders who continue to reimagine how that need can be solved.