In March, in the same building as my new office in Bozeman, the Pentagon opened a front door.

It's called a Defense Innovation Unit OnRamp Hub, and it sits on the Montana State Innovation Campus. There’s a good reason they’re here. I advise startups at MSU and around the country, and the flood of defense and dual-use startups is real. That’s why this building is buzzing with lidar, photonics, and drone startups.

The narrative shift around defense tech startups has been dramatic, and the lesson in it is relevant for all startups: If the world is full of pigeons, make sure you get the best hole.

Only 8 years ago, Google employees protested by the thousands when they discovered their company was working with the military. Some resigned. Google dropped the contract and published a set of AI Principles listing the things it would never build, weapons among them.

In February 2025, Google deleted that section.

What changed? Nobody sat the protesters down and out-debated them. Instead, the industry just changed the words.

The key pivot came from Andreessen Horowitz, which around 2022 announced an investment thesis it called "American Dynamism."

"Military-industrial complex" carries decades of baggage. "American Dynamism" carries a flag. The former was foisted upon the industry by a presidential critic; the latter invites you to a cause. Katherine Boyle, who co-founded the practice at A16Z, framed the whole shift in 2023 as taking the "moral courage to say the word America out loud."

That is not an argument. It's a renaming. And it buys permission to feel good about something previously suspect.

The fusion energy industry ran the same play for a different reason. For fifty years the industry sold breakthrough physics, and its critics had a ready punchline: always thirty years away. Then, somewhere around 2024, the language migrated. Fusion stopped being clean energy and started being compute fuel, the power source for AI data centers. Microsoft and Google signed huge offtake agreements, years before the tech is economically viable.

In the military’s case, the reframe bought moral permission. Here, it bought customers.

The GLP-1 revolution depends on another reframe, this time for bureaucratic acceptance. Medicare's Part D won’t cover weight loss drugs. So pharma looked around for all the conditions that weight loss would help. The word they needed decided which study they paid for. Novo Nordisk got Wegovy approved for cardiovascular risk reduction; Lilly got Zepbound approved for sleep apnea. Presto, covered.

To be clear, the reframing didn’t cause the shifts. They capitalized on demand (for defense innovation, data center power, weight loss drugs) and removed the asterisk. They let people embrace their products.

Each of those companies got a different prize. Permission, customers, a budget line. Here's how to find out which one is available to you.

Which budget do I fall into?

That’s the Wegovy move. If your name maps to no existing line, you're asking a buyer to invent one, which is the hardest sale in the enterprise.

Conventional: "We're an autonomous inspection platform, and we'd love to run a pilot." That's innovation budget. Small, discretionary, and the first thing cut when things get bumpy.

Better: "We cut unplanned downtime on your critical assets." That's maintenance budget. Already funded, already someone's problem, already has a number attached.

Who am I associated with?

Your category name (say, vendor to the military-industrial complex) picks your comparable set, and your comparable set decides what math people run on you. Last year investors in defense startups started hyping young defense companies like Anduril as “neoprimes” to associate them with the big, reliably profitable defense behemoths like Lockheed while still making them sound more agile. Some of the companies tagged as neoprimes have indeed become prime contractors to the military, but not all. Which of course is the risk: if your positioning gets too far ahead of reality, you open a new line of attack for critics.

Conventional: "We're a diagnostics company." You’re destined for reimbursement fights and LabCorp comps.

Better: "We're the patient-selection layer for oncology trials." You’re highlighting the value you create in the pharma pipeline.

What word am I spending energy denying?

Channel the fusion industry. "We're not a services business." "This isn't a science project." "We're not another AI copilot." It’s like saying, don’t think of a pink elephant.

Conventional: "We're not a services business. We're a product company that happens to do some implementation." Yep, you’re a services business.

Better: "We're the calibration layer every one of these instruments runs on." No denial. Nothing to argue with.

So look at your framing. Is it a loaded word your critics would choose? Or is it a name that hands your audience a frame they'd be glad to stand inside?

If you're still using their word, you're still arguing on their turf. Find your "America." Then say it out loud.