The Quietest Story in the AI Boom Is About Who's Building It
Black founders just posted their best quarter since 2022. Three deals carried most of the weight. The structural problem hasn't moved.
FYI - this is the newsletter post that fully expands on the tek INTEL briefing that I send out on Friday afternoon. Let me know if this hits or misses the mark with you.
In Q1 of 2026, Black-founded US startups raised $643 million. That’s the strongest quarterly performance since Q2 2022, when founders pulled in $653 million during the immediate post-George Floyd capital push.
Crunchbase published the data this week. Every outlet that covers diversity in venture ran with the same frame: progress. Recovery. Resurgence.
The numbers, looked at honestly, tell a different story. Let me walk you through it.
The legitimate good news first
Thirty-four deals closed for Black-founded startups in Q1. After eight to nine quarters of a broader venture funding downturn that hit diverse founders disproportionately hard, that’s real movement. The companies that closed those rounds are doing serious work — particularly in AI infrastructure, where capital is currently concentrated at historic levels.
So yes — $643 million is up. That deserves acknowledgment.
Now let’s look at what those 34 deals are actually made of.
Three checks did most of the work
Of that $643 million, the lion’s share came from just three transactions:
SambaNova Systems raised a $350 million Series E for its AI hardware platform. Intel reportedly increased its stake to 8.2% in the round.
Noviq, a sports prediction startup, closed a $75 million Series B.
Harper, a YC-backed AI insurance platform, raised $47 million.
That’s $472 million across three deals. Roughly 73% of the quarterly total.
Subtract those, and the other 31 Black-founded startups combined raised about $171 million in Q1 — an average of around $5.5 million per deal across three months in a $252 billion venture market.
The “record quarter” headline is real. The distribution is not what most readers would assume from it.
What 0.26% actually means
Here’s the framing that didn’t make most of the headlines:
$643 million is 0.26% of the $252 billion raised by US startups in Q1 2026.
For comparison, Black Americans are roughly 14% of the US population.
The gap between the 0.26% capital share and the 14% population share hasn’t narrowed during the AI boom. It’s widened.
Total US startup funding hit $290 billion in 2025. Black-founded companies captured $942 million of that — 0.32%. Down from the 2021 peak of $5.2 billion, when post-2020 corporate commitments were still flowing.
The trajectory since 2021 has been downward, not upward. A record quarter sitting inside a downward trajectory is still a downward trajectory.
The structural read
I want to quote Crunchbase’s head of research, Gené Teare, directly here because her framing is sharper than most of the analysis circulating this week:
“We are eight to nine quarters into a venture funding downturn, but Crunchbase data has shown a persistent decline in funding to Black-founded companies that outpaces the overall decline in startup funding. One has to wonder if the abundance of caution that’s now prevalent in the industry has prevented investors from taking chances on first-time founders who are more likely to be diverse.”
The mechanism she names — “access to networks, relationships, and early introductions” — is the same one researchers have been documenting for fifteen years.
It hasn’t budged. It might have hardened.
In a market where the bar for first-round funding now includes “warm introduction from an existing portfolio founder,” and where 75% of venture capital is concentrating into AI infrastructure plays anchored by a handful of established firms, the relationship-based gatekeeping problem isn’t abstract. It’s mechanical.
What this tells us about who AI is being built for
This is the part the headlines don’t talk about, but it’s worth saying plainly.
The AI infrastructure layer being constructed right now — the foundation models, the agent platforms, the data center buildouts, the tooling around them — is being capitalized at a scale we haven’t seen in any prior tech wave.
This week alone:
Anthropic filed for IPO at a $47 billion revenue run-rate
SoftBank committed €75 billion to French AI data centers
SpaceX filed a $75 billion IPO specifically for space-based AI compute
Apollo and Blackstone are syndicating a $36 billion chip-financing debt deal for Anthropic alone
When the capital pools that decide what gets built are this concentrated, and when 99.74% of that capital is going to companies whose founders aren’t Black, the products being shipped will reflect the priors of the people who funded them.
The training data choices. The deployment defaults. The use cases that get optimized for. The communities whose feedback gets taken seriously and whose feedback gets ignored.
This isn’t a values argument. It’s a product argument.
AI tools built without diverse capital tables ship with predictable blind spots — we’ve watched this movie now in facial recognition, in hiring algorithms, in healthcare risk models, in lending decisions. Each generation we’ve expressed surprise. We should be past surprise.
What to do with this information
I’m not going to wrap this in optimism that isn’t there. The structural problem is structural. It doesn’t get solved by individual founders working harder or by an essay on Saturday morning.
But there are practical things worth doing this week:
If you’re a founder: The “warm intro” requirement is real and it’s not going away. Build your network now, before you need the round. The community we built at Digital Collective exists specifically to address the relationships-and-early-introductions gap that Crunchbase named. That’s not marketing copy — that’s literally why we built it.
If you’re an investor or LP: The data argument for inclusive deployment of capital has been published for years. Read the HBCUvc report. Read the Crunchbase Diversity Spotlight quarterly. Pattern-match against your own portfolio. Then act.
If you’re an operator or creator: Pay attention to which AI tools you build your business on and whose products you amplify. The dollars you spend are signal too. Distribution is leverage. Use it.
If you write about this stuff: Stop running the celebration headline without the denominator. $643 million is meaningless without the 0.26%. And 0.26% is meaningless without the trajectory it sits inside.
The bottom line
A record quarter for Black founders inside a $252 billion venture market that gave them 0.26% of the pool isn’t a recovery story. It’s a concentration story.
The AI boom is producing the largest capital flows in the history of technology. The gap between who’s allocating that capital and who’s receiving it is widening, not closing. And the tools being built on top of that capital will inherit the blind spots of the people writing the checks.
That’s the headline.
Everything else is comfortable noise.
Sources: Crunchbase News (Diversity Spotlight Q1 2026 report), TechCrunch (May 31, 2026), Black Enterprise (June 1, 2026), Build Fast with AI (June 5, 2026).
If this resonated, share it with a founder, operator, or investor who needs to read it. And if you’re ready to build inside a community designed to solve the network-and-introductions problem this piece names, Digital Collective is open.
— James
"..taking chances on first-time founders who are more likely to be diverse.” sadly is the same story, different day. This will likely get worse. I've talked about who's building these models - the blind spots are dangerous. Call them what they are, Tech Bros. (I think they like that). But women, and especially women of color, are being shut out and it shows. Change is needed quickly, but times are making that increasingly difficult.