Ian Tien (Mattermost) on Sovereign AI, Defense Tech, and Building in National Security

Mattermost founder Ian Tien shares lessons on sovereign AI, defense tech, enterprise sales, and building an open-source company in national security.

In this episode of the Founders in Arms podcast, we sit down with Ian Tien, founder and CEO of Mattermost, to explore one of the most important emerging categories in tech: sovereign AI and defense software.

Ian shares how Mattermost evolved from an open-source Slack alternative into a mission-critical collaboration platform used in national security environments—and what founders need to understand about building in this space.

This conversation dives deep into:

  • What “sovereign AI” actually means

  • Why governments care about owning their tech stack

  • How open source enabled bottom-up adoption in defense

  • Lessons from selling into government and enterprise

  • The realities of defense tech vs Silicon Valley startups

  • How AI is reshaping national security

  • Why supply chain security is the new frontier

  • What founders get wrong about defense tech

In this episode, we cover:

(00:00) The rise of sovereign AI

Governments around the world are investing heavily in sovereign AI.

The goal:

  • Reduce dependence on external systems

  • Control their own infrastructure

  • Strengthen national resilience

This shift is becoming a core part of national security strategy.

(04:08) What Mattermost actually does

Mattermost started as an open-source alternative to Slack.

Today, it has evolved into:

  • A self-hosted collaboration platform

  • Workflow and automation tools

  • Secure communication infrastructure

It’s designed for environments where security and control are critical.

(05:33) From gaming startup to enterprise software

Mattermost’s origin is unexpected:

  • Started as a gaming company

  • Built internal tools for reliability

  • Turned those tools into a product

This mirrors companies like Slack and Discord, which also emerged from gaming roots.

(07:54) The concept of “sovereign collaboration”

If governments want sovereign AI, they also need:

Sovereign collaboration.

That means:

  • Self-hosted systems

  • No external dependencies

  • Secure, multi-user workflows

This is where Mattermost fits.

(08:11) Bottom-up adoption in national security

One of the most surprising insights:

Mattermost spread inside defense environments organically.

Why?

  • Better UX than existing tools

  • Easy to deploy

  • Open source accessibility

Teams adopted it before leadership formally approved it.

(09:21) Open source as a distribution strategy

Open source enabled:

  • Rapid experimentation

  • Trust through transparency

  • Bottom-up enterprise adoption

Eventually:

  • Organizations upgrade to enterprise versions

  • Revenue follows usage

(10:27) The reality of VC money

Ian shares a candid lesson:

Taking VC isn’t the problem—spending it poorly is.

Examples:

  • Overspending on non-essential items

  • Hiring before understanding the market

  • Scaling without clear product-market fit

Profitability forced better discipline.

(13:28) Why capital alone doesn’t solve growth

Money only works if:

  • You understand your customer

  • You have the right go-to-market strategy

  • You invest in the right areas

Otherwise, it accelerates mistakes.

(14:18) Learning an entirely new industry

Moving into defense required learning:

  • Government procurement

  • Military terminology

  • Security classifications

  • Enterprise sales cycles

It’s essentially learning a new language.

(14:25) The “acronym problem” in government

Selling into defense means navigating:

  • DoD terminology

  • Security levels (IL4, IL5, etc.)

  • Complex procurement processes

Understanding the language is critical to credibility.

(17:23) How to learn a new domain quickly

Ian’s approach:

  • Attend industry conferences

  • Spend time in-person with customers

  • Immerse yourself in the ecosystem

There’s no shortcut—exposure compounds.

(19:00) The CEO job: hardest problems only

The CEO’s role is simple (but hard):

Work on the hardest problems in the company.

That includes:

  • Biggest risks

  • Biggest opportunities

  • Cross-functional challenges

(24:49) Hiring executives like co-founders

Ian emphasizes:

Executives should feel like partners, not employees.

Key traits:

  • Long-term alignment

  • Trust and ownership

  • Ability to scale with the company

(26:34) What makes a great executive

Look for:

  • Experience at different company stages

  • Both success and failure (“scar tissue”)

  • Ability to operate under uncertainty

These traits matter more than pedigree.

(30:50) AI + defense = massive global shift

Two major investment areas in 2025:

  • AI

  • Defense

Governments are increasing defense spending dramatically, creating massive opportunities for startups.

(31:00) Why sovereign AI matters globally

Countries want:

  • Control over their AI systems

  • Visibility into supply chains

  • Independence from foreign influence

This is driving demand for local models and infrastructure.

(34:28) The importance of supply chain security

AI is now part of the software supply chain.

Risks include:

  • Compromised open-source components

  • Foreign influence in models

  • Hidden vulnerabilities

Security is becoming a top priority.

(38:52) Advice for defense tech founders

The biggest mistake:

Staying inside the building.

Instead:

  • Meet customers in person

  • Attend defense conferences

  • Understand real-world problems

This space rewards proximity to users.

(42:07) Is defense tech really slow?

Common belief:
Government sales take years.

Reality:

  • It’s similar to enterprise sales

  • Slow if you don’t know the system

  • Fast if you understand the process

(47:02) Understanding government buying models

There are three main approaches:

  1. Butts-in-seats contracting (services-heavy)

  2. Commercial off-the-shelf software (COTS)

  3. Hybrid platforms (e.g. Palantir-style)

Each has different incentives and tradeoffs.

(50:00) Why software procurement is broken

Government systems are optimized for hardware, not software.

Result:

  • Slow procurement

  • Inefficient processes

  • Difficulty adopting modern SaaS

This is slowly changing.

Key Takeaways for Founders

Sovereign AI is a massive emerging category

Governments want control over their AI and infrastructure.

Open source can unlock unexpected markets

Bottom-up adoption can reach even national security environments.

Capital without clarity is dangerous

Money amplifies both good and bad decisions.

Defense tech is learnable—but requires immersion

You must understand the language, culture, and systems.

Enterprise and government sales are similar

Both require trust, relationships, and patience.

AI introduces new supply chain risks

Security and governance will define the next wave of winners.

About the Guest

About Ian Tien

Ian Tien is the founder and CEO of Mattermost, an open-source collaboration platform used by enterprises and governments worldwide.

He previously worked at Microsoft and Trilogy, and has spent over a decade building software for secure, mission-critical environments.

Listen to Founders in Arms

Founders in Arms is a podcast for builders—covering startups, AI, enterprise software, and the realities of company building.

🎙 Subscribe to Founders in Arms Podcast on your favorite platform
💬 Join the conversation at TribeChat.com
🚀 Discover more founder insights

Previous
Previous

Nicholas Rudder (Sphere) on Pivoting, Selling Before Building, and Global Tax Compliance

Next
Next

Noah Glass (Olo) on 20 Years to Product-Market Fit, Enterprise SaaS, and “Embrace the Suck”