Episode summary
Everyone Is Still Undersizing the AI Market | Eric Vishria
Invest Like the Best
· · 1 hr 18 min · Hosts: Patrick O'Shaughnessy · Guests: Eric Vishria
Summary
Eric Vishria argues that AI is not a commodity market where one provider will eat everything, but a much larger market that will likely produce an oligopoly plus several giant smaller winners. He compares the current moment to cloud, where many smart investors initially missed AWS and later overcorrected, and says the same kind of undersizing is happening in AI now. The real bottleneck, in his view, is not whether demand exists but whether infrastructure can scale fast enough across power, chips, memory, and software.
A major theme is that AI changes the competitive frontier for software. Vishria says the old SaaS playbook breaks when hitting plan can destroy value rather than create it, because the market moves so fast that the best companies must constantly reinvent themselves. He says leaders now need to understand the jagged edge of model capability, not just customer needs, and that the best sellers are often founders who can bridge those two worlds. Companies like Sierra and Cursor succeed, he argues, because they treat product development as sand castles that must be rebuilt as models improve.
He is also bullish on hardware and robotics, but only where there is a clear new workload or a viable data flywheel. He tells the Cerebras story as an example of productive naivete, then says robotics will likely emerge first in controlled or repeatable settings before generalizing into household tasks. Throughout, he emphasizes that great founders, technical depth, and real partnership matter more than rigid role definitions.
Key takeaways
- Vishria thinks AI will produce an oligopoly of winners plus multiple $100 billion-plus companies, not a single dominant vendor.
- He says SaaS winners now must reorient around a shifting competitive frontier, because “every single day that you are hitting your plan, you are destroying equity value.”
- He believes the best AI companies are “AI sherpas” that help enterprises bridge the gap between existing workflows and model capability.
- On the demand side, he says enterprise AI adoption is already more eager and threat-aware than cloud adoption was in its early years.
- He is most worried about energy, arguing that compute converts energy into intelligence and that energy supply may limit future AI scale.
- He thinks robotics will advance by bootstrapping high-value data and then layering post-training, similar to how LLMs were built from internet data.
Notable quotes
“every single day that you are hitting your plan, you are destroying equity value”
“I really believe they will be these like hundred billion dollar crazy smaller winners”
“let's be their AI sherpa”
“They understand that jagged edge of capability and they build around it”
How the episode opens
I I'm very dismissive of this idea that like people are going to vibe code their own like and whatever like it that that isn't the issue. The issue with SAS companies is every single day that you are hitting your plan, you are destroying equity value. Like think about that.
Our whole careers we learned you lay out a plan, you execute against it like relentlessly and violently. You hit your plan or you exceed your plan and you keep on building that and that's how you build equity value. And now you're going in here just like every day you hit that plan.
You're up. — You're up. You're destroying value. I love asking you and all your partners this every time we hang out, which is, okay, you've got these singular investments. You don't do that many of investments each per year, and then the ones that go on to work, like so far Sierra and Fireworks certainly are. You get to learn so much about the world through the lens of the company.
Yeah. So, I'd actually love to do both of those. Maybe starting with fireworks. What What do you know or what have you learned about the world and how it's reordering itself based on watching the world through the lens of fireworks that would maybe be surprising or interesting?
These are two trillion, three trillion, four trillion parameter models. — Well, it turns out running those models is damn hard.
And running them efficiently is like super hard. And the way to see this and everybody can see this which is everybody from AWS to Azure to GCP to the Neoclouds to the Fireworksbased 10 togethers of the world like all of them they all run these stock open-source models that are available in their developer pools and everything else and and like that's what it is and the performance difference for a fireworks versus like a cloud provider is like 5x
And that is just the speed performance.
From the episode’s captions: roughly the first two minutes, lightly cleaned. Keep watching on YouTube →
Topics
- ai market sizing
- saas competitive frontier
- energy bottleneck for ai
- ai chips
- venture investing strategy
Companies and tickers mentioned
- Benchmark
- Fireworks
- Sierra
- Cerebras
Watch the episode
Watch “Everyone Is Still Undersizing the AI Market | Eric Vishria” on YouTube
Summary, takeaways and quote selection by Octagon. Quotes are short excerpts; the full conversation is in the episode.
More from Invest Like the Best
Invest Like the Best, this episode and its artwork belong to Colossus | Investing & Business Podcasts. Octagon is not affiliated with the show; this page is an independent summary.
