Sai Arora retweeted
The core confusion here is that Katie thinks of credit as a sourcing problem when it's an underwriting problem. Debt doesn't work for pre-revenue startups because they have no cash flows to service the debt and often no collateral worth recovering.
Debt is available where those things exist: venture debt complements later stage equity rounds, and equipment and ABS cover hard assets.
That said, debt is much different from equity and a VC's intuitions don't map well to it at all. If you lend to people you're trying to be friends with ("founder friendly" in the Silicon Valley argot), you'll be too soft on covenants, you'll hesitate to enforce those covenants, and you'll find out what the collateral is worth at the worst time.
The crux of the issue here is that AI is properly understood as being commodities and infrastructure with a thin software layer on top. That doesn't map to VC intuition or their understanding of AI, which is, of course, software-first.
And this explains why VCs keep reaching for debt. Venture math works when marginal cost is near zero and gross margins run at 80%+, so one winner pays for the whole fund.
AI inverts that. Inference has real COGS, labs are tied to capex cycles, and much of the value goes to whoever owns the scarce inputs: power, chips, interconnect and land. That's a credit and commodities game, where returns come from cost of capital, utilization and duration matching. Picking power law winners doesn't help here.
Further, the software layer is thinner than it looks, because it rents its margin from the layer below. If model prices fall, app layer companies get squeezed by the next feature release. If compute stays tight, they get squeezed on input costs. In both cases the moat is below them.
can one of the private credit guys pls go rip a fund focused on 20-smthg sf founders?
there's a ton of places to get debt but i stg none of them are legible + accessible to the lil zucks running around sf. much needed in the age of asset heavy businesses accruing the most value
this would look like front office of a vc, back office of a trad credit firm
that means:
•on the ground in sf. talent mapping. events. hire a sourcer corgi girlie. be a chiller founders wanna hang with. this is the core differentiation from trad credit firms founders don't know personally
•legit financial dd, not vibes investing
•educate the lil zucks on asset classes and forms of financing. lack of financial literacy across the valley is astounding. we need a new form of education / accelerator, focused on financing for founders
•new form of network value-add for early stage founders. this neofirm would spike on PE style search algo connections, not ex. recruiting engineers
hmu if you do this. i have a ton of inflecting founder friends with "companies that touch atoms" who need you asap
Recurring theme in convos with Physical AI companies is how antiquated and slow it is to deal with private credit people if you’re not what they’d consider a “big deal” (e.g. $5M+ opp)
can one of the private credit guys pls go rip a fund focused on 20-smthg sf founders?
there's a ton of places to get debt but i stg none of them are legible + accessible to the lil zucks running around sf. much needed in the age of asset heavy businesses accruing the most value
this would look like front office of a vc, back office of a trad credit firm
that means:
•on the ground in sf. talent mapping. events. hire a sourcer corgi girlie. be a chiller founders wanna hang with. this is the core differentiation from trad credit firms founders don't know personally
•legit financial dd, not vibes investing
•educate the lil zucks on asset classes and forms of financing. lack of financial literacy across the valley is astounding. we need a new form of education / accelerator, focused on financing for founders
•new form of network value-add for early stage founders. this neofirm would spike on PE style search algo connections, not ex. recruiting engineers
hmu if you do this. i have a ton of inflecting founder friends with "companies that touch atoms" who need you asap
91
Sai Arora retweeted
we have a winner for the most absurd AI startup ever
this Peter Thiel backed company uses AI robots to stab fish in the brain so they taste better and last up to 3x longer
it's called Shinkei and it modernizes a centuries-old Japanese technique
dozens of Michelin-star restaurants are already lining up to buy the fish
here's how it works:
> the problem is when a fish gets caught, it usually just suffocates on the boat
> that stress floods the meat with hormones and lactic acid, which hurts the flavor and texture
> the old Japanese fix for this was spiking the fish through the brain right after the catch, which stops all that
> but doing it by hand is impractical for most American fishing crews
> so Shinkei built a fridge-sized robot that sits right on the boat's deck
> the crew feeds in a fish and AI identifies the exact species, so it knows exactly where the brain and gills are
> the robot then drives the spike through the brain and cuts near the gills in about 7 seconds
> the fish goes into an ice slurry so the blood drains out
> so the fish doesn't sit there stressing out, and that tasty flavor and texture is preserved
example number 5,392,417 of how AI can be applied to literally anything
Sai Arora retweeted
My friends in France often ask: what's so great about building a startup in San Francisco? Isn't the city terrible? Isn't it only tech bros?
So we decided to film our experience building a startup in SF.
Episode 1 is live here - and there's no sleeping in the office, tenderloin controversies or permanent underclass
A machine should know whether it made a good part.
At twentyfour26, we’re building toward that: AI machinists that plan the work, run CNC machines, check the results, and use that feedback to correct the process.
We’re starting with on-demand CNC machining for hardware teams in the Bay Area, with turnaround times as fast as 48 hours.
Our bigger ambition is to make manufacturing expertise repeatable, so every machine we add inherits more of the knowledge needed to make the next part.
Hardware teams should be able to turn a new design into a physical part without weeks of waiting.
That’s why we’re building twentyfour26.
Sai Arora retweeted
The next 12–15 months are going to be game-changing for robotics
We’ll see way more real-world deployments, new infra getting built to manage them at scale, simulation getting much better, and models improving fast.
Sai Arora retweeted
Maybe naive of me but in the purchasing of our new laser found out something unintuitive about financing. Financing / underwriting / deal making on equipment is more flexible than I thought
we tried getting financing from one manufacturer but they ultimately told us we’d have to put a number down which was higher than we preferred. I took that offer at face value. Assuming they would present their best offer when asked
then when we got our approval from mits, the first mfg all of a sudden found a way to offer the same terms
the difference was hundreds of thousands of dollars between the first offer and what they ultimately said they could do
Sai Arora retweeted
the only way to succeed as a robotics company, is to provide value without needing any robots in the first place.
Sai Arora retweeted
Deploying robots is hard and expensive
Robotics founders need to burn thoughtfully.
Venture funding in Robotics has grown by over 400% over the last 2 years.
Actual installations of robotics is up <25% YoY.
Robotics is not AI. The diffusion will be slower for a while and your underlying metrics will need to catch up to the valuation at some point.
A reorder is such a strong signal. One motion tells you the following:
1. The company worked out almost EVERY single kink that came up during pilot
2. The robot exceeded expectations around the value it provided
3. Customers see the proof and want to double down now.
The best robotics demo is a reorder.
@DynaRobotics says Din Tai Fung signed a bigger contract after deploying napkin-folding robots across locations.
Show me sites live, repeat orders and margins after support.
Pilot count is messaging market fit.
Growing the total number of robots deployed at the individual location level and the total number of locations is product-market-fit.
Good work by Dyna.
75
This is so true.
Lots of companies can get pilots.
But converting those pilots into paid expansion is much rarer.
Companies that can get a customer to deploy more robots, across more sites, after seeing the first one work are showing something much closer to real PMF.
The best robotics demo is a reorder.
@DynaRobotics says Din Tai Fung signed a bigger contract after deploying napkin-folding robots across locations.
Show me sites live, repeat orders and margins after support.
Pilot count is messaging market fit.
Growing the total number of robots deployed at the individual location level and the total number of locations is product-market-fit.
Good work by Dyna.
💯
“The line I keep thinking about came from a robotics founder: the hardest problems in AI are deployment problems now, not research problems.”
Post pilot deployment problems like maintenance, servicing and financing in full force all at once.
I spent 48 hours in SF hanging out with billionaires, data labs and AI founders and execs running $1B to $100B companies.
My observations:
1. Sooo MANY stories of data brokers making like $2M a WEEK selling company data to labs. I can't tell if it's the beginning or a bubble, and I kept wondering which niches are left. Digging into it actively.
2. At dinner, someone asked the table, "If a new model drops next month, does your company get better or worse?" HALF the table went quiet! I think that's the best question to ask about any AI startup right now.
3. The bar people talk about now is hitting $1B in revenue in 12 months (I guess that's the new unicorn?). What I didn't hear much about is retention. AI revenue can come in fast and leave fast, and I think we're going to see a lot of stories of AI companies with huge revenue and terrible retention.
4. It's pretty clear we're living through the reinvention of the computer. The model is the CPU, the harness is the operating system, and skills and plugins are the apps. Thinking about it that way makes it easier to see where the whitespace is.
5. We're maybe 12 months from an agent that can watch you work for a week and then do your job without any instructions. The screen recording plus agent memory plus local model combination makes this possible right now.
6. Agent harnesses feel like the new GPT wrappers, in a good way, especially vertical ones. The best teams I talked to treat the model as swappable. One founder walked me through routing different models to different jobs: one for taste and visuals, a cheaper one for the everyday stuff, another for text.
They use prompts they can move across models over fine-tuning, and A/B test the output on what users actually do with it: do they edit it, export it, share it? A vertical harness is basically four things: the connectors for that industry's data, evals written by people who know the work, rules for what the agent can and can't touch, and a clean screen for a human to review. The model is maybe 20% of it.
7. There's a lot of reverse engineering going on/cloning going on more than I've ever seen. People are decompiling software and rebuilding it with agents, and someone showed a suite of like 10 apps that are worth $5B+ each all that he cloned open source. Me and everyone are still thinking through what it means when software gets this easy to copy.
8. My guess is the value moves to data. If the software can be rebuilt quickly, what's worth paying for in a legacy SaaS company is the years of customer data and workflows inside it. Some SaaS companies are secretly data companies.
Gets me thinking about how do you truly value data for this saas companies? Like if we were buying these dying saas companies at 2x ARR before (as example), maybe there is a better metric. Idk, just thinking out loud here.
9. It feels like a good chunk of legacy software won't make it to the agentic era, especially anything built around people clicking through a lot of screens.
10. Every company I talked to wants to pipe its internal data into one place and expose it to agents through MCP. Then they look at the data and realize it's a mess: five versions of the same customer, old Slack channels, PDFs nobody can search. Cleaning that up feels like a big consulting opportunity!
11. A lot of founders are building the basic building blocks for agents: APIs, data layers and tools. Some of the companies I saw barely have a UI because the main user is an agent. I really believe the startups that build primitives to this agentic era will do extremely well. Building blocks is the best way to think about it.
12. Labs can get plenty of training data, but it's harder to get real examples of agents failing inside companies. Whoever has those logs has something valuable.
13. Someone running a big consumer company told me the next fight is over memory. The model that remembers the most about you wins, and the smartest model matters less than people think.
14. Physical AI came up a lot. Astra I think was the first model that got people into AI can actually do CAD, and I think a lot of people are thinking what does vibe manufacturing really look like. The line I keep thinking about came from a robotics founder: the hardest problems in AI are deployment problems now, not research problems. Their robots work on their own, call a human when they get stuck, and every one of those calls becomes training data.
15. Cybersecurity companies are having a moment and are actually cool companies to build now, especially the ones that clean up after a breach. Every agent with access to your systems is a new door into your company, and a lot of those doors were left open.
16. Some startups are worth more dead than alive now. A shutdown company with 5 years of Slack, email and code can get paid for licensing all of it. I wouldn't be surprised if founders start factoring that into when they shut down.
17. People are starting to talk about monetizing the agent web. It feels really early, but when the most popular user of the internet is an agent, what startups should you be building?
18. I think agent insurance is coming. Once an agent makes an expensive mistake, insurers will want audits before they cover anyone.
19. Several people said the hire they want most is someone with taste who can write evals. The models are getting good at generating ideas, so the hard part is judging which ones are good.
20. New trend I observed: startup raises VC, uses venture dollars to become a creator, startup fails, founder falls back on being a full-time creator and incubates other products bootstrapped.
21. Getting SIGNIFICANT distribution is HARD right now. Founders are sponsoring buses, buying billboards like crazy and throwing parties with famous DJs to get noticed. Kinda feels dotcom era in that sense.
Attention is what's scarce! The growth guy going viral sharing his UGC strategy from Wispr Flow is a good example. It's the sauce everyone wants, and I truly believe there are probably only 50 people on the planet who know how to scale these types of businesses efficiently in this era.
Always good to get that dose of SF. I felt sooo lucky to have such an incredible trip during such an incredible time in building businesses. Surreal in so many ways.
Happy to be home and locked in.
Until next time!
1
6
296
Mercoa just launched on @ycombinator's Launch YC!
Mercoa CardPay: AI Agents for B2B Virtual Cards.
Check it out: ycombinator.com/launches/OA1…
3
243
@united what’s worse is that you think sending me a $15 meal voucher (via a virtual card that expires tomorrow) 2 min before I board a nonstop flight to my destination is the way to remedy the situation.
1
88
Sai Arora retweeted
To @kvlly and anyone in a similar situation, we'll buy out a part of your ZD contract with a transition to @usepylon. Migration can be done in 1-2 weeks.
happy to chat cal.com/sersic/demo
One of the best parts of building @MercoaFinance has been learning about all of the different verticals our customers operate in. Excited to share more here about our perspective on the state of #verticalsaas opportunities in the weeks to come
1
161
Sai Arora retweeted
There isn't a single book that reads
"Hey, I had this idea. I started it. Everything went great. The end."




















