There are investment ideas you never pursue because doing the research would take too much effort.
A company you keep hearing about. A connection you suspect matters. An assumption behind a stock you own that you haven’t quite tested.
You save the article. Open another tab. Tell yourself you’ll come back into it.
And sometimes you never do.
I know that feeling too well.
I’ve spent years doing investment research. And one thing I've learned is that having an investment idea is one thing. Doing the work to know whether the evidence actually supports it is another.
Digging through years of filings. Reconstructing what management said 6 quarters ago. Connecting developments across companies. Figuring out what actually changed.
And just when you've pieced everything together, another quarter comes around. The work starts again.
An Olympic athlete has coaches and doctors behind them. A large investment fund has teams of analysts.
But when you invest for yourself, as I was doing, you're often the entire research department.
So a year ago, I kept coming back to the question:
Why should having a world-class research team behind you be a privilege reserved for the largest funds?
That's why I started Tessara.
We're a small team with a huge ambition: to build the world's best AI research team for public-market investors.
Your investment ideas. A research team behind them.
Today, we’re opening up that work to you, where you can access your own AI research team.
We're starting where our research runs deepest: memory and semiconductors.
Over the past 12 months, we've built the foundations to make this possible.
We've mapped the AI supply chain and the critical constraints shaping it, developed deep domain knowledge, and built company research records that we continuously maintain.
On top of that foundation, we've developed specialized AI agents that investigate your questions and connect evidence across sources.
So when you bring Tessara a question, you're not starting from scratch. There's already a substantial body of research to build on.
Here’s an example.
What if the opportunity is in the GPUs already installed?
We asked Tessara: “Who can earn more from an existing AI cluster without adding another GPU?”
It surfaced an interesting disclosure.
In its September earnings call, Oracle revealed that GPU capacity coming up for renewal had been renewed or resold at a 20% premium to previous contracts.
The surprising part? Most of those GPUs were already four years old or older.
Tessara connected that with something CoreWeave had discussed.
CoreWeave explained that its initial customer contracts are designed to repay the asset-level financing used to build GPU clusters. Once those contracts end, the infrastructure can be contracted again, potentially generating additional returns on top of those already earned.
Two different companies. Two separate disclosures. Together, they raise a question that doesn't get nearly as much attention as the next generation of GPUs:
How much of an AI cluster's value comes after its first contract?
We're spending billions building new AI infrastructure. But what if the infrastructure already installed has more earning power left than investors expect?
Of course, Oracle's disclosure covers a specific renewal cohort, not its entire fleet. It doesn't establish how profitable those renewals are, or whether the same economics will persist.
But it opens up a question worth investigating: How do the returns on an AI cluster change once its initial financing has been repaid?
That's what excites me about Tessara.
A question leads to evidence. Evidence reveals a connection. And that connection opens up a different way to think about an investment.
Not just finding answers. Finding better questions worth investigating.
And you can inspect the original sources, challenge the conclusions and take the research further yourself.
What would you put your research team to work on?
Think about a company you own, or one you're considering investing in.
What's the one thing you'd want to understand before putting more capital behind your view?
Have an investment question that needs investigating? Ask.
Bring Tessara the question you'd ask a great research analyst if you had one working alongside you. Let its AI agents work through the evidence and test the assumptions behind your thesis.
Heard a claim you want to check? Use Tessara Search
Find the original passage in an earnings call, filing or research source. See what was said, when it was said, in what context. Instantly.
Whether you're investigating an idea or checking a critical disclosure, the goal is the same: spend less time assembling the facts and more time deciding what they mean.
You keep the judgment. You get help with the work
Try Tessara Pro free for 7 days. Then $150/month unless you cancel before the trial ends.
If you're at a fund and interested in using Tessara with your team, reply to this email. I'd love to hear from you.
There's much more to come, but I'm excited to finally put Tessara in your hands.
If you're unsure where to begin, just let me know the company you're following and the question on your mind. I'll help you find a good starting point.
Teng