Alan Grosheider - MKS Ventures

09/03/2026

How Do You Get Rich?

I hate these cheesy fake videos, but the point of this one is dead on.

If you want to make a lot of money building a company...

If you want to make a lot of money investing in companies...

Build to Exit!

Plan your exit from the very beginning. Build to sell.

That's our game plan at MKS Ventures.

https://www.instagram.com/reel/Dc0-efJCEfx/?utm_source=ig_web_copy_link&igsi=MzRlODBiNWFlZA==

09/02/2026

Where's your moat? You hear that a lot.

If you're focused only on building great AI functionality, you probably have zero moat.

The AI controller layer is the most important part, and most people have no idea they even need it.

Three big things happened over the last couple of weeks that really reinforce this concept.

Full article in bio.

The most defensible part of an AI company isn't the model — it's the layer that decides what the model is allowed to do. Why the control layer is the real moat.

08/19/2026

Think about what your team did by hand this week.

Pulled comps. Chased documents. Read through a contract looking for one clause. Coordinated vendors by email. Followed up on the follow-up.

At CBRE, software is doing pieces of all of that right now.

I went through their own published materials this week instead of the coverage about them. What's already deployed:

→ Contract and lease terms pulled out automatically instead of read line by line
→ AI that ranks likely buyers before anyone picks up the phone
→ Building systems AI across a billion square feet, cutting repeat maintenance alarms by 98%
→ Procurement contracts you can search by asking a question

Meanwhile, the average real estate firm is using AI to write listing descriptions.

Both of those get called "AI adoption." They are not the same thing, and the difference is going to show up in your business before it shows up in a headline.

Here's what I'd want to hear if I were running a firm right now: you are not going to build what CBRE built, and you don't need to. They have a dedicated AI organization and a five-year head start.

What you can do is stop buying features and start buying and building finished workflows.

A tool that drafts an email saves you minutes. A system that completes an entire process end-to-end saves you hours — and hours are the only thing that actually changes your economics.

Pick the one task in your operation that eats the most time and produces nothing but phone calls. Start there.

Full breakdown in this week's newsletter, including the three mistakes I see firms make in year one.

CBRE has AI running across 20,000 sites and inside its core workflows. What's actually deployed, and what a real estate firm should do about it.

08/16/2026

This is me over 30 years ago.

I had just secured a $3 million VC investment in my company.

Man, I thought I had it all figured out. 😅

If I only knew then what I know now.

1. VC investment comes with a lot of strings attached.

2. VC investment comes with unbelievable stress.

3. Don't raise VC money unless you absolutely need a lot of capital to build your business.

4. Only raise what you need to build what you really want to build.

5. Fun, play, and creativity are more important than you think.

6. You lose most of the fun, play, and creativity with unnecessary pressure.

7. If you're not having fun, the money is not worth it.

These days, I help people learn how to raise only the capital they need and to have fun with their business.

08/13/2026

Do you really need to raise venture capital?

Most likely, you don't.

In fact, it may be the worst thing you could do for your company.

Here's my experience and why most companies should raise capital differently.

08/13/2026

He used to have sixty people writing code.

Now he has two.

I had lunch yesterday with a friend I've known for about ten years. He's an IT project manager who built a real company — sixty-plus offshore resources delivering IT services, a management layer on top, the whole structure.

Over the last few months, he built a Claude-based automation engine.

Same code output. Two people.

That's not even the part that got me. The part that got me: work that used to take six weeks now takes days.

He didn't hire better engineers. He didn't buy a platform. He sat down, mapped what his people actually did all day, and rebuilt the workflow, so the machine handles the connective tissue and the humans handle the judgment.

I left that lunch thinking about my own company. Honest conclusion: we're not moving fast enough either.

Almost nobody is.

Ask a company if they're using AI, and they'll say yes. Ask what for, and you get a chatbot. FAQ responses. Faster email drafts. Summarizing a document someone still has to read.

That's surface AI. It speeds up the work. It doesn't remove the work.

Meanwhile, the real money is sitting in the boring layer — the people whose full-time job is moving information between systems that don't talk to each other.

NAR puts transaction coordination at roughly 14 hours per real estate closing. Zoom out, and it's about 45 hours of combined effort per transaction, 30 of them on paperwork and coordination.

Thirty hours per deal of work requiring almost no judgment.

And the engineering firms performing those inspections? Some of them do $500M a year and run entirely on email. Junior staff sending requests, catching bids, scheduling, tracking it in spreadsheets, routing reports up a chain.

Nine-figure companies running a workflow you could explain to a sixth grader.

I see the same thing in healthcare. One rehab facility needs three people to produce a single daily report on who came in, who left, and where each patient is.

Nobody's even scoping it.

Here's why: stopping to rebuild feels expensive right now. Falling behind feels theoretical. So everybody waits.

Does automating this reduce headcount? Yes. Say it out loud instead of dodging it.

But the alternative isn't preserving those jobs. It's a competitor rebuilding the workflow, pricing at a level you can't match, and taking the book of business — which costs every job in the building.

If the boring work in your company can be replaced with AI, you'd better be working on it right now.

Someone else already is.

FULL ARTICLE BELOW

AI foot-dragging is adopting AI for chatbots while your real workflows stay manual. Why companies stall on automation — and what it costs when a competitor doesn't.

08/11/2026

Airtable just gave us a fascinating lesson in startup economics.

The company built an incredible business with roughly $480M in ARR. But along the way, it raised about $1.4 billion and reached a valuation of nearly $12 billion.

Now it's being acquired at an implied equity value of roughly $2.25 billion.

Think about that: a $2.25 BILLION exit, and some later-stage investors are reportedly still losing money.

That's why I'm increasingly convinced that raising more capital and chasing bigger valuations aren't necessarily signs of success.

Every round raises the bar for the eventual exit.

There's nothing wrong with raising capital. Raise what you need to build and accelerate the business. But capital efficiency, founder ownership and a realistic path to an exit matter.

I'd rather see a company raise $3M and sell for $50M than raise $300M and need a multi-billion-dollar exit for everyone to win.

That's a big part of the thinking behind our Built to Exit thesis at MKS Ventures.

08/09/2026

Everyone still talks about building the next IPO.

The data suggests that's becoming the exception, not the norm.

The chart below (shared by Nasdaq in a recent article by Phil Mackintosh, with contributions from Nicole Torskiy) tracks venture-backed companies seven years after their first financing round.

A few things jump off the page:

• In the mid-1990s, more than 25% of venture-backed companies reached an IPO within seven years.

• By companies first funded in 2009, that number had fallen to roughly 2%.

• Meanwhile, acquisition rates stayed surprisingly consistent.

• And a growing percentage of companies simply remained private much longer.

That's one of the reasons our philosophy at MKS Ventures is different.

We don't build companies hoping Wall Street eventually notices.

We build companies that customers want, acquirers value, and founders can actually own.

Our target isn't a 12-20 year journey to an uncertain IPO.

It's to build capital-efficient companies designed to become highly attractive acquisition targets in about four years.

It's to deliver fantastic returns to our investors in a shorter time.

That's what we call Built to Exit.

Ironically, focusing on developing an exceptional acquisition candidate often leads to a healthier company, regardless of how it ultimately exits.

The best exit strategy isn't chasing the IPO.

It's building a business that someone can't afford not to buy.

Credit to Nasdaq's Phil Mackintosh (with Nicole Torskiy) for highlighting this data, based on research by Michael Ewens and Joan Farre-Mensa.

08/07/2026

Most companies are racing to build AI agents.

I think many are building the wrong thing.

An AI agent by itself isn't much of a competitive advantage anymore.
Anyone can connect an LLM to a workflow.

The real moat is everything around the agent.

The graphic below is a simplified view of the AI Controller architecture we're building at Blue222. It's not just about connecting to an LLM. It's about orchestrating an entire business.

• Authentication
• Security and permissions
• Business rules
• Data orchestration
• API integrations
• Human approvals
• Audit trails
• Logging
• Memory
• Cost management
• Reliability

The AI agent becomes just one component inside a much larger system that understands your business, enforces your policies, and coordinates dozens of specialized agents behind the scenes.

That's why I think the next generation of enterprise AI won't be won by whoever builds the smartest agent.

It will be won by whoever builds the best orchestration layer.

The agent is becoming the employee.

The controller is becoming the company.

That's where the defensibility is.

08/06/2026

Everyone is asking the wrong question about AI and real estate.

It isn't whether AI replaces brokers. Probably not.

It's that AI will replace much of what brokers spend their day doing — and that's the bigger disruption.

Spreadsheets didn't eliminate accountants.
CAD didn't eliminate engineers.
Word processors didn't eliminate lawyers.

They collapsed the time cost of the work and quietly raised the standard for what "good" means.

AI is doing that to real estate right now. Listing descriptions. Follow-up. Scheduling. Contract prep. Market analysis. Inspection coordination. Due diligence.

Instead of hiring another assistant, many brokers will simply add another AI.

But here's what doesn't go away:

Negotiating when the other side is unreasonable. Reading the client who says yes and means no. Calming a seller at 10 p.m. Solving the problem the file didn't anticipate.

Those skills don't get less valuable. They become more valuable — because they account for a larger share of what the client is actually paying for.

The disruption isn't the broker disappearing.

It's the broker's calendar disappearing.

New article on where this actually lands — including why the real pressure is on brokerage economics, not on agents.

The question isn't whether AI will replace real estate brokers.

The question is whether another broker using AI will replace you.

https://blue222.com/the-truth-about-ai-and-real-estate-brokers/

Will AI replace real estate brokers? No — but it will replace most of what brokers do all day. What that means for agents, brokerages, and splits.