The Three Postures of an Angel Investor

There's a version of angel investing that looks like this: write the check, read the quarterly update, vote yes on the board resolution, wait. Disciplined, systematic, and almost entirely passive. It's a legitimate way to run a portfolio. It is also the version that adds the least value both to the founder and, over a long enough horizon, to the investor's own return.

The more useful frame is that angel investing isn't one job. It's a sequence of different jobs, and the job changes on a schedule the angel doesn't control. It is set partly by how much capital the angel has deployed, and partly by the governance event most angels underweight: the point at which the company gets a board.

Three ways to show up after the check clears

Once capital has been committed, angels tend to settle into one of three postures.

The Analytical Monitor behaves like a mini-VC: reads the financials, exercises information rights, votes on resolutions, waits for the outcome. Systematic, low-effort, low-friction — and largely detached from the thing that actually determines whether the investment works.

The Strategic Connector doesn't do the day-to-day work but compounds a career's worth of relationships on the founder's behalf: introductions to enterprise buyers, key hires, and the introductions to Seed and Series A firms that will decide whether this company gets to exist in eighteen months.

The Operational Co-Pilot goes further: a fractional executive who helps build the first sales playbook, pressure-tests the financial model, or shapes the early engineering org design.

None of these is wrong or right. The case for leaning toward Connector and Co-Pilot, deliberately, is a de-risking argument. Early-stage startups don’t fail because of macro cycles, they fail because of an unwillingness to learn and pivot and of execution mistakes: the wrong first sales hire, a pricing model that doesn't hold, a go-to-market motion targeting the wrong buyer. An angel who actively coaches a founder through those specific decisions is actively helping reduce the 50+% failure rate to the next round. 

Why the early window runs on relationships, not rights

What makes the active posture possible in the first place, and that most first-time angels miss, is that there usually is no board at pre-seed.

Most early checks go in via SAFEs, a security explicitly designed to be fast and light on governance. There is no board seat that comes with it. No formal voting rights, no contractual claim on the founder's time, no table to sit at. Whatever influence an angel has in this window is not codified on paper. It's earned, offered, or simply taken, through the relationship.

It means the "value-add" angel isn't leaning on some formal governance lever. It means access is a function of how useful, low-friction, and trustworthy the angel actually is to the founder without any fiduciary structure enforcing the relationship. And it means the angel who wants to be more than a name on the cap table has to design that role on purpose, because nothing about the SAFE requires the founder to give it to them.

A few ways angels who take this seriously can structure it before it becomes a governance question

1. Audit your superpower and say so explicitly
   Identify the one thing you're good at, ideally better than most people in the founder's network: enterprise sales motion, pricing, a regulatory domain, hiring a first VP. Tell the founder plainly not to just send general updates but to call when they hit a wall in this specific area.
   
   

2. Agree on the cadence instead of leaving it open-ended
   A recurring 30 - 45 minute monthly call beats being on call for ad hoc late night texts. It's more sustainable for the angel and more predictable for the founder.
   
   

3. Define the triggers that warrant an unscheduled call
   Name the handful of situations that justify going off-cadence, so both sides know what "urgent" means: A competitor move, a hiring decision in your domain, a fundraising strategy question.

The key takeaway from all of this is that "angel investor" describes a capital commitment, not a fixed job description. The job is Analytical Monitor, Connector, or Co-Pilot from pre-seed onwards, operating on relationships because there's no board to formalize anything. 

The practical implication: whatever role an angel wants to play -  Connector, Co-Pilot, or some mix - the window to play it actively is genuinely short relative to the decade-plus the capital stays locked up.

What Four Investments Taught Me About Defense Tech Before I Knew I Was Learning It

Last week I was asked to co-chair the Defense, Intelligence and Aerospace Special Interest Group at the Band of Angels, alongside Rick Lu of Pacific Defense.

How did I get here?

One day you're an angel investor writing about industrial software, robotics and startup mechanics, the next you're co-chairing a group that vets seed-stage defense deals.

Aside from occansional writings about military doctrine and satellites, it was only in going back through my portfolio to think about what I'd bring to the DIA SIG that I noticed four investments over the last several years that, at the time, I would not have filed under "defense tech." I filed them under "great technical team, differentiated tech, real customer pain." All four shared a second property: every one of them builds something a commercial customer needs and that a military customer needs for the same underlying reason.

Geosite built geospatial intelligence and property-risk analytics, the kind of thing an insurer needs to price risk on a piece of land, and the kind of thing an intelligence analyst needs to understand what's happening on that same piece of land. Descartes Labs, which does a substantial amount of work with the geospatial intelligence community, acquired the company in April 2024, and both were in turn acquired by Earth Daily Analytics (EDA).

Metalware is building security for firmware for the operators who build and run it, which turns out to be exactly what the Department of Defense and its suppliers need when they think about their own systems infrastructure. The company has been doing work for government primes on its way into direct DoD relationships.

Okapi Orbits is the clearest version of the pattern. As the number of satellites in orbit keeps climbing, someone has to manage the traffic and avoid the collisions. This is a space domain awareness problem for every commercial satellite operator and for every government that operates satellites of its own. The company raised a 13M seed round in 2025 to build that infrastructure.

Nobility Space builds air-breathing electric propulsion for satellites operating in Very Low Earth Orbit (VLEO), altitudes low enough for most spacecraft to sustain orbit. The promise of VLEO is enticing: sharper imagery, faster communications links, more frequent revisit rates, and a harder target to track. VLEO is naturally attractive for persistent intelligence, surveillance, and reconnaissance (ISR) and space domain awareness called out in the U.S. Space Force doctrine in April 2025. 

These are companies where the underlying technical problem is the same problem, viewed from two markets with different sales cycles, procurement processes, and risk tolerances. 

It is also the underlying theme of most of what I've written about military doctrine, innovation, and space over the years, including a piece I wrote in December on the Pentagon's Portfolio Acquisition Executive (PAE) model, where individual PAEs are given combined financial and product authority over a mission portfolio and are judged, as I put it then, not by compliance metrics but by "how quickly solutions are delivered, how relevant they are."

The Long Middle of the Founder’s Journey: Board Alignment, Trade-Offs, and Realism

Nobody talks about the long middle of the founder's journey. The years after conviction, but before clarity. It is the time where most startup stories are decided and characters are shaped.

I have watched a version of this play out with an industrial SaaS founder I am privileged to advise. From the onset everything looked great. The company solved a hard problem uniquely, had real customers, recurring revenue, credible investors and a friendly board.

⁓⁓⁓⁓⁓⁓⁓⁓⁓⁓⁓⁓⁓⁓⁓⁓⁓⁓⁓⁓⁓⁓⁓⁓⁓⁓⁓

The friction showed up as the company scaled: the expectation an industrial company should grow like a SaaS unicorn, timelines not matching reality, a bulging capital structure. Not every good business can be stretched into someone else's success story. Founder and board constantly negotiated about control vs. trust, founder ambition vs. realism, speed vs. durability. Board meetings started optimizing for a clean narrative instead of hard trade-offs. Fundraising logic started dictating operating decisions. 

The founder was working harder and controlling less and the runway no longer was just financial, it became psychological.

Industrial software startups very rarely IPO. They are acquired by large industrial companies if they manage to become really large. If they are on the smaller side they get bought by private equity platforms that price EBITDA over growth.

Once those choices become explicit, everything come into focus: product calls get easier, conversations with acquirers become deliberate instead of reactive, and an exit stops feeling like failure and starts looking like completion. 

⁓⁓⁓⁓⁓⁓⁓⁓⁓⁓⁓⁓⁓⁓⁓⁓⁓⁓⁓⁓⁓⁓⁓⁓⁓⁓⁓

This particularly ending did not produce a 10x outcome for the investors. 

It forged a founder who deeply understood the system, the trade-offs it demanded, and the cost of pretending otherwise.

It also delivered a very lucrative exit for a founder already focused on his next venture.

VCs are paid to hunt for 10X outliers. Founders are building for life-changing wealth.

Venture capitalists are structurally incentivized to hunt for rare outlier startups with greater than 10x outcomes that can return the entire fund.

Most early exits do not produce these >10x returns, but exits with lower multiples can be life changing for founders.

This leads to a major structural misalignment because first-time founders frequently lack the market timing and M&A experience necessary to recognize the optimal window for an exit.

Every Stage Has a 50% Conversion Rate

The path for startup founders with VC funding seems obvious: keep raising, always. The rationale: later-stage exits are worth dramatically more since exit values climb steeply at every stage, and dilution doesn't climb nearly as fast.

But that thinking ignores the odds of getting there. Roughly half of all companies don't make it to the next round, and the odds get longer at every stage and only a small fraction of seed-stage companies ever reach the later rounds at all. Ilya Strebulaev of Stanford GSB recently published his findings on nearly 53,000 companies that raised their first round in 2014, tracking their fate through 2023: only 4% had an IPO exit, and 20% exited via M&A.

Founders Should Ask Themselves This Every Six Months

So the math doesn't give a clean "always keep raising" answer. It comes down to one question a founder needs to ask at every stage:

Are my odds of reaching the next round better than the average company sitting where I am right now, or am I just hoping?

If yes, raising again beats cashing out, assuming the value climbs faster than the dilution and the risk. The dilution is the easy part to model. The risk is much harder to assess.

At later stages, top line growth is a simple proxy for risk. Rory O'Driscoll of Scale Venture Partners documented this in his seminal analysis termed the 'SaaS Mendoza Line of growth': once private companies fell below the expected growth rate, only 35% saw even one year of growth re-acceleration, and fewer than 10% managed two or more years of it. In other words, once you fall below the growth rate expected by your next round's investors, your odds are very slim.

If There is Growth Momentum: You Live to  Raise Again … And Die Another Day

The key metric is the momentum of that growth, which shows whether the growth rate is accelerating or decelerating.

If the growth rate is rising, and rising faster each quarter, it's worth to keep on pushing.

Yet even then, founders face risks they do not control: market windows opening and closing in times and ways no one can predict, competitor gaining market share, key people leaving.

SaaStr's Jason Lemkin put it bluntly in a 20VC podcast: 'When you start to get into nosebleed territory it has to be worth 10X to go for it. Building something generational - you kind of know as a founder when you're on that path.'

He added that 3X over a three-to-four-year horizon isn't worth it: "There is way too much risk for not enough money."

If you can't see a path to 10X, you should seriously consider selling.

If The Growth Rate Stalls: Head for the Exit

The easy case, and the one rarely talked about, is the M&A exit.

If growth is still rising but the increments are shrinking, the founder and investors should push for an orderly exit.

If growth is already falling, the window closed a few quarters ago — and they're usually the last to know.

Fixing the SAFE Gap: How Founders Can Rebuild Governance

Most seed-stage companies never had a board.

At pre-seed, SAFEs account for roughly 90% of U.S. deals. At seed, where a priced round is a real alternative, roughly 64% of rounds still get raised on SAFEs and another 10% on convertible notes. Only about 27% of rounds are raised as priced equity, the one structure that seats an investor-director and creates a fiduciary board.

The SAFE instrument was built to close rounds in days instead of months, argue less about valuation, and keep the cap table simple. But the few weeks of speed come with a cost, and that cost is the loss of governance.

SAFE holders are contract counterparties, not stockholders. They have no fiduciary duties and no board seats. The result: companies now raise tens of millions of dollars and operate for three or more years without a board ever being constituted.

When there is no priced round, no one has the right and the standing to ask the hard questions. There is no structured review of what's working and what isn't. There is no forum where outside perspectives are exchanged without management in the room. There is no mechanism to replace a CEO or key executives who stop performing.

A company with no outside check during that period of highest uncertainty has no check on the founder's blind spots: The VP hire nobody vetted before the offer went out, the churn number nobody tracked consistently, the work the founder was still doing two quarters after it should have been delegated.

The probability to graduate from Series Seed to Series A has within two years after the Serie Seed has plummeted since SAFEs have become the dominant fundraising instrument, caused by a mix of macroeconomic shifts and cap table mechanics. Multi SAFE stacking and inflated valuation caps have all led to increased VC selectivity with higher expectations. 

The inflation of SAFEs has removed the automatic mechanism for governance. Until a priced round seats a real board there is a void for someone else to purposefully fill that gap. It is where founders should leverage informal allies to replicate boardroom oversight without the administrative burden.

Enter the experienced angel advisor.

Experienced angel don't need a board seat to help close that gap. They can provide real oversight, with no formal structure and no extra dilution attached to it. They substitute organizational power with personal power. 

Experienced angels will help establish clear operating rules.

• They set the ground rules early. They agree, in the first real advisor conversation, on what gets reviewed and how often.

• They introduce a cadence. A structured check-in every month or every other does the job a CEO evaluation and an executive session would otherwise do.
  
  

• They will track the same signals a Series A board would before the company is forced to: Non-financial North Star metrics tied to customer value, revenue, cash burn and runway.

• They review the first VP hires before the offer goes out. One bad key hire costs a company more than any single missed sales quarter.

• They help pick the next round's investors, not just the price. Capital comes with governance and named partners.

Founders should get experienced angel advisors on board at the pre-seed stage and agree on their explicit roles. Then and only then the angel advisors will deliver value all the way to the first priced round, and sometimes even beyond.

This post was inspired by The Cure for Bad Boards Isn't No Boards by Pascal Levensohn, Private Company Director, July 27, 2026.

Disposable UIs and Bedrock Foundations: The new architecture of B2B SaaS.

The hot take in B2B SW right now: You don't need enterprise software, you just let agents access the database and let them do the work. 

But several signals that are pointing to something different and more nuanced.

Jason Lemkin's team runs SaaStr on 3 humans and 20+ AI agents. His conclusion isn't that CRM is dead. It is the opposite: 20 agents are writing directly to a database and produce 20 different definitions of a qualified lead. No shared forecasting logic, no audit trail. It turns out the agents need the shared system of record as much as the humans do.
Salesforce's answer is to go headless where CRM acts as an API/MPC substrate, with Slack, voice, and custom UIs as interchangeable surfaces on top. 

Andre Wenz of SAP Signavio pushes the argument one layer up. He observes that today's application screens of pipelines, stage views, and approval forms aren't the process. Instead, they are cloud era artifacts, born when software was too expensive to rebuild per task. With AI, the interface itself is generated per objective and is disposable. The surface is fluid but the foundation is orderly.The canonical records, permissions, policies and controls are the bedrock underneath. 

And then there is Celonis. They are already living the 'foundation, not the app' thesis on top of  any single ERP stack. Their Process Intelligence Graph mines the actual processes across SAP, Oracle, Salesforce and other B2B apps and feed that context to agents via MCP. 

Put together, here is my read on where this goes: 

Generated, fluid, disposable UIs will become the norm; fixed screens will be limited to very few use cases and demos. 

The system of record and governance layers become more valuable because ungoverned agents are more dangerous than ungoverned humans. But rebuilding a system of record is hard and will take to time get deployed and distributed. 

The real battleground is the governance layer. 

SpaceX wants 100,000 satellites in VLEO, making it the default orbit for the space industry

On July 6 2026, SpaceX asked the FCC for authority to fly Gen3 Starlink 100,000 satellites spread across two Very Low Earth Orbit (VLEO) shells at roughly 325 km and 475 km. This is a more than sixfold increase over the 15,000 satellites the FCC has approved for Starlink so far and in orbits far lower than Starlink's original constellation. 

While Starlink built its original business at 550 km and higher, Starlink already runs more than 650 satellites at 340 to 370 km to support direct to cell service. It moved lower when a real business opportunity emerged in the form of direct to device connectivity, and Gen3 is the next step and a full commitment to VLEO as the default orbit for the next generation of the constellation.

Below 450 km, residual atmosphere drags an unpowered satellite out of the sky within weeks. For sustained operations below, significant material science and physics problems need to be solved. Two key solutions are in the early stages of validation:

The first is propulsion. Air breathing electric propulsion (ABEP) ingests the residual atmosphere itself, mostly atomic oxygen, as reaction mass converts the abrasive oxygen into propellant. ESA's DISCOVERER program and NATO backed work at Kreios Space are pushing this from research into flight hardware. Once ABEP is flight qualified, a VLEO satellite trades a lifespan measured in weeks and months for one primarily limited only by the fuel it carries. 

The second is autonomy. A constellation spread across two tightly stacked shells at 325 km and 475 km, with satellites moving between altitudes and inclinations as SpaceX's own filing describes, cannot run on ground computed maneuvers alone. Spacecraft in VLEO should hold station, avoid collisions, and execute rendezvous and proximity operations with minimal supervision from the ground. Pair that with ABEP and a satellite gains the ability to not just survive at 325 km, but to actively move, between shells, between inclinations, even between orbits, in response to demand.

Propulsion solves how long a satellite can stay in VLEO. Autonomy solves how it moves once it gets there. Together they turn a fixed shell into a fluid, maneuverable layer of orbit, exactly what SpaceX's own request for flexible altitude and inclination assignment assumes will exist.

For investors, the signal is clear. The world’s largest satellite operator just announced to make VLEO its primary orbit at a scale that dwarfs every LEO constellation flown or proposed before it. That is a strong vote that the drag problem gets solved in time to matter, and it will pull capital toward the propulsion and autonomy companies that make it possible.

For founders building in VLEO, the filing is validation and warning at the same time. Validation, because the largest player in the industry just confirmed where the puck is going. Warning, because it will  be hard to compete with an even partially built 100,000 satellite constellation for orbital slots, spectrum, and coordination. 

Everyone still talks about LEO as the orbit that matters. As of July 6, VLEO is the New New Thing.