Nick | Compute Labs

A couple things kept us steady:
• We focused on what we could control: the quality of our messaging, how fast we followed up, and how well we listened
• We didn’t take silence personally. Momentum requires some detachment

Q1 was about proving to ourselves that we could build from scratch.

Onto Q2

We kept showing up anyway.

We knew clarity would come from doing, not from waiting for permission.

We sent cold emails. Got on imperfect calls. Rewrote the narrative as we learned what resonated. Got ghosted more times than I can count. Followed up anyway.

Building a business pipeline from zero isn’t glamorous.

In Q1 2025, we lacked brand recognition. No inbound leads. Few warm introductions. Just a constantly evolving deck, a developing product, and a vision we believed in: providing investors a way to tap into the AI boom through structured compute

We kept revisiting the same questions:

• Who exactly are we for?
• What specific problem are we solving?
• Why now?

Once we nailed these, the same outreach started hitting.

Lesson: Don’t underestimate how much clarity drives performance.

Before trying to scale your message, sharpen it.

The biggest unlock in early sales? Clarity.

We didn’t land better meetings because we got louder.

We landed better meetings because we got clearer.

At first, our messaging felt off. Calls didn’t stick.

It didn’t feel like a big breakthrough at the time, but it's what got the ball rolling.

Channel sales might be unsexy, but they scale faster than you’d expect.

We’re a long way from having it nailed. But 2025 Q1 taught us that value can come from unexpected places as long as you’re open to asking.

We used to think to ourselves, "We don't really have a network."

Turns out, we just weren’t asking.

Once we started reaching out, people showed up:

• An investor intro’d us to an LP
• That LP connected us to a financing partner
• Which led to a lead… who intro’d us to two more

Building pipeline early is messy but it pays off if you keep refining as you go.

• Cold > Warm (sometimes)
Cold leads worked better when our message was sharp and personal. Effort earns attention.

• Momentum is silent.
You don’t feel it daily. But 30, 60, 90 days in, someone will reply. Once they do, the flywheel starts turning.

Some lessons that stuck with us:

• It’s a maze, not a funnel: Timelines will stretch and conversations will restart months later.

• Your network works only if you work it: Intros came from friends-of-friends and advisors once we asked.

Everyone talks about the sales funnel.

Outreach → Meeting → Close

But when you’re building a pipeline from zero, it’s more like:

Outreach → Maybe → Ghosted → No → Delay → Yes

• POV > Pitch: Instead of leading with “here’s what we do,” we started with “here’s what we’re seeing” and "what's your perspective?"

We didn’t crack some secret code — we just tried to be useful, respectful of people’s time, and willing to get better each week.

• Precision over spray-and-pray: Our best cold messages were short and showed we did homework. Don't be an LLM.

Here’s what actually worked for us:

• Shortlists over lead lists: We grouped targets by segment to learn what messaging worked best per vertical

Compute Labs doesn't have a known logo yet.

But we still got into some great conversations and eventually built a real pipeline.

Feeling honored after an intense two weeks with #HumanX, #GDC, #BlockTalk, and #GTC.

Highlights:
• Reuniting with friends
• Engaging with partners & customers
• Visiting partners on-site
• Speaking at BlockTalk keynotes & panels
• Welcoming our Head of Growth to the B2B world

Now, onto follow-ups!

Why was #HumanX worth attending?

Nothing replaces in-person conversations.

Pitches land differently when you can read body language, adjust in real-time, and get unfiltered feedback. The ability to adapt your pitch is immediate, as opposed to doing the same pitch 5x a week over calls.

HumanX was such a blast. Learned a ton, made new friends, and got to see first-hand where the trends in AI are going.

I'm looking forward to debriefing with my team, and more importantly can't wait to go home to my wife & cats.

If we just missed each other, please reach out!

Here's my conference booking link:
app.usemotion.com/meet/nikola...

Excited to attend both #HumanX and #GTC this month. If y'all are attending, let's chat!

I'm looking for data-centers and cloud providers that want to scale their operations, but aren't interested in the capital risk & depreciation associated with owning hardware.

With a sea of suppliers, it's the small teams that often surprise you, boasting stronger OEM/manufacturer ties than some industry giants.

Let's not underestimate the Davids amongst a bunch of Goliaths.

Fair point. The Anora crew probably had a secret stash of H100s they couldn't disclose.

Seriously though. For those of you thinking you can sacrifice sleep - sure. Go ahead. Just know that good sleep is actually better for your career & business, but more importantly your overall happiness.

Turns out Bryan Johnson was right.

The older I get (yes, I know I’m still young), the more obvious it is: bad sleep = bad work.

Unfortunately, I’m not waking up at 4 AM to chug 80 supplements and bathe in red light anytime soon.

A sad day for my nostalgia playing MMOs over Skype.

😂😅

Can't comment on OpenAI specifically, but besides that this is generally true. Yeah.
Nvidia kept a very low supply to start and that inflated the market.

It's coming down right now at a steady pace.

Getting 5090s right now isn’t about wanting them—it’s about who actually has them:
🔒 OEMs lock in priority deals.
👥 Large buyers scoop early allocations.
💸 Some resellers simply markup to something ridiculous.

Had to dig through overlooked channels to lock in our batch. Hardware markets are weird.

Hardware sourcing = patience + persistence + some negotiating.

5090s don't grow on trees, but after weeks of chasing suppliers and resellers, we locked in a solid supply.

Now onto the next problem.

Step 1: Ask a supplier for 5090s.

Step 2: Get laughed at.

Step 3: Find a less obvious supplier, time the conversation right, and actually close the deal.

Turns out persistence + knowing where to look = a working strategy.