There you have it! The Pentagon is now in talks to hand a single AI startup a $5 billion loan — w…
September 14, 2026 · 0 likes · 0 comments
China Threat AI Defense
There you have it! The Pentagon is now in talks to hand a single AI startup a $5 billion loan — while private venture capital is shoving money into defense tech faster than anyone can count it.
Let me get the numbers straight, because they matter here.
VCs put $14.6 billion into defense tech in just the first five months of this year. That's about nine times the $1.6 billion they invested in all of 2020. Anduril raised $5 billion in May at a $61 billion valuation and is reportedly chasing a new round at $100 billion — that would put a nine-year-old company above Northrop Grumman on paper. Four startups alone (Anduril, Helsing, ShieldAI, Saronic) have pulled in $21.5 billion.
So the capital is there. It is abundant. It is fighting to get in.
Which is exactly why the next part gives me pause.
The Pentagon's Office of Strategic Capital wants to lend Fluidstack, an AI cloud-computing startup, roughly $5 billion. Not $5 million. $5 billion. To one company. And this is the part nobody has answered: the office hasn't disclosed what happens if Fluidstack can't pay it back.
I want us to win the AI race against China. I have said it a hundred times. The strategic case is real. But there's a difference between the government buying the outcome and the government becoming the lender.
When private money is this eager, why is the taxpayer the one underwriting single-company credit risk at $5 billion? That is not procurement. That is venture betting with a public balance sheet — and taxpayers eat the loss if the bet goes bad.
To be fair, there's a case on the other side. Mark Esper argues we need to move faster and field cutting-edge tech before the next fight, not after. Fixed-price contracts from these firms could genuinely cut the waste the primes got fat on for decades. Speed matters. Fielding matters.
But speed and a $5 billion loan to an unproven startup are not the same thing. Buy the capability. Award the contract. Set the milestones. Let the people with billions already lined up carry the downside.
If Fluidstack is a great bet, the market will fund it — the market is clearly desperate to. If it's too risky for private capital to touch, that tells you something the taxpayer should hear before we sign, not after.
Ask the question early. Not after the next funding round.
By the way — this story surfaced on UnbiasedHeadlines.com, my news platform built entirely by AI agents. No spin, no editors picking sides, just the facts across every political lean. Full article: https://lnkd.in/eGZTQycv
And if you want to see how the machine that found this story actually works — I'm doing a live walkthrough of the full agent stack Sept 22, 1:30PM ET. Come see it run.
Would you hand $5 billion of public money to a startup the private market is already lining up to fund?
Let me get the numbers straight, because they matter here.
VCs put $14.6 billion into defense tech in just the first five months of this year. That's about nine times the $1.6 billion they invested in all of 2020. Anduril raised $5 billion in May at a $61 billion valuation and is reportedly chasing a new round at $100 billion — that would put a nine-year-old company above Northrop Grumman on paper. Four startups alone (Anduril, Helsing, ShieldAI, Saronic) have pulled in $21.5 billion.
So the capital is there. It is abundant. It is fighting to get in.
Which is exactly why the next part gives me pause.
The Pentagon's Office of Strategic Capital wants to lend Fluidstack, an AI cloud-computing startup, roughly $5 billion. Not $5 million. $5 billion. To one company. And this is the part nobody has answered: the office hasn't disclosed what happens if Fluidstack can't pay it back.
I want us to win the AI race against China. I have said it a hundred times. The strategic case is real. But there's a difference between the government buying the outcome and the government becoming the lender.
When private money is this eager, why is the taxpayer the one underwriting single-company credit risk at $5 billion? That is not procurement. That is venture betting with a public balance sheet — and taxpayers eat the loss if the bet goes bad.
To be fair, there's a case on the other side. Mark Esper argues we need to move faster and field cutting-edge tech before the next fight, not after. Fixed-price contracts from these firms could genuinely cut the waste the primes got fat on for decades. Speed matters. Fielding matters.
But speed and a $5 billion loan to an unproven startup are not the same thing. Buy the capability. Award the contract. Set the milestones. Let the people with billions already lined up carry the downside.
If Fluidstack is a great bet, the market will fund it — the market is clearly desperate to. If it's too risky for private capital to touch, that tells you something the taxpayer should hear before we sign, not after.
Ask the question early. Not after the next funding round.
By the way — this story surfaced on UnbiasedHeadlines.com, my news platform built entirely by AI agents. No spin, no editors picking sides, just the facts across every political lean. Full article: https://lnkd.in/eGZTQycv
And if you want to see how the machine that found this story actually works — I'm doing a live walkthrough of the full agent stack Sept 22, 1:30PM ET. Come see it run.
Would you hand $5 billion of public money to a startup the private market is already lining up to fund?