Ron

Quiet observer of the markets. Learning as I go.
Here for the signal, not the noise. 📊
Finance | Macro | Markets

[Ron @RonAMacro]

The smartest people on X aren’t the loudest — they’re the ones who actually think for themselves.

[Ron @RonAMacro]

@HenryInvests Good luck and great job!

[Ron @RonAMacro]

Paul, the market still thinks you’re a fintech. You’re not. You’re the only AI company that owns the full credit feedback loop of origination to repayment to default across many millions of borrowers over a decade. That data moat doesn’t exist anywhere else. Anthropic and OpenAI can’t replicate it with a plugin. Hammer that message HARD!

[Ron @RonAMacro]

Technically not that hard. But credit decisioning isn’t a plugin problem, it’s a data problem.

Anthropic and OpenAI don’t have 12+years of loan performance data, repayment behavior, and default outcomes across millions of borrowers. Upstart does. That’s the moat. The model only gets smarter because it owns the feedback loop which is origination to repayment to default, end to end.

A credit plugin from a foundation model would be a generic wrapper. Upstart’s model is trained on outcomes. Big difference.

[Ron @RonAMacro]

@HenryInvests @Upstart With $2.4B in committed institutional capital secured in roughly 60 days, can you speak to the institutional appetite you’re seeing now vs. 12 months ago and what’s the gating factor to scaling committed capital well beyond $5B?

[Ron @RonAMacro]

@davegirouard You can't buy 14 years of repayment outcomes on the open internet.

[Ron @RonAMacro]

@goddreng10 Yes.

[Ron @RonAMacro]

@HenryInvests @goddreng10 BULLISH AND LONG.

[Ron @RonAMacro]

@HenryInvests Whoa. Q1 already printed 61% YoY origination growth. The UTPI was at roughly 30-35 during Q1. It's now at 43 and going vertical.

[Ron @RonAMacro]

@TedPillows Agree 100%.

[Ron @RonAMacro]

@EndWokeness Pier Village in Long Branch is a beautiful area. This is sad to see.

[Ron @RonAMacro]

This is one of the simplest, most impactful things that could happen for working Americans. Zero taxes on the bottom half costs almost nothing to the government and means everything to those families. Jeff, please bring this directly to the President. You have his ear and the credibility to make it happen. Push.

[Ron @RonAMacro]

@ZynxBTC Agree.

[Ron @RonAMacro]

@HenryInvests It is recurring. 👀

[Ron @RonAMacro]

Excellent analysis.

An 87% increase in GAAP net income on a stock trading at 2x forward revenue. With a PEG ratio that would be approximately 0.15 at current prices against that earnings growth. That is not an expensive stock. That is one of the cheapest growth stories in the entire market if the guidance proves accurate.

[Ron @RonAMacro]

@HenryInvests Winning super prime business requires pricing concessions that compress near-term margins. The long term payoff is a larger, more diversified, more defensible platform. The short term cost is exactly what Q1 showed.

[Ron @RonAMacro]

@HenryInvests This directly contradicts the private credit stress narrative. If private credit was genuinely deteriorating the assessed value would be falling relative to cost. The orange line would be declining. It’s doing the opposite.

[Ron @RonAMacro]

Paul, you built the AI. Dave built the company around it. Same foundation, new chapter.
What does Paul Gu’s Upstart look like differently than Dave Girouard’s Upstart? And what’s the one thing you’ll do that finally closes the gap between business performance and shareholder returns?

[Ron @RonAMacro]

The EBITDA margin dropped from 20% to 13% because Upstart is deploying capital into co-investments that generate future returns rather than immediate fee income. It’s not margin erosion, it’s capital deployment into a growing, performing portfolio that pays back over 12-24 months.

This is the Amazon web services moment. AWS looked like margin compression to the income statement for years before it became the most profitable division in the company.

[Ron @RonAMacro]

This is exactly why UPST is different from every other AI story. LLMs get commoditized because the models are available to everyone. GPT. Claude. Gemini. No moat.
Upstart's models can't be commoditized because the data can't be replicated. 104 million real world repayment events accumulated over 14 years. That's not a prompt you feed into ChatGPT. That's irreplaceable truth about human credit behavior that lives exclusively inside Upstart's training dataset. When LLMs commoditize the app layer every fintech built on top of OpenAI gets hurt bad. Upstart gets stronger because the scarce asset was never the software. It was always the data.