Investors retreat as Big Tech piles AI debt on top of massive new AI equity sales

(Originally published June 24 in “What in the World“) The geeks borrowed too greedily and too deep.

That was the market’s reaction to SpaceX’s decision to follow its record, $86 billion sale of new shares this month with the sale Monday of $25 billion in bonds. The bonds are senior, unsecured notes, meaning they have no collateral to protect their holders, but would in the event of SpaceX’s liquidation be repaid before shareholders get a dime. Their sale appears at first glance to make SpaceX’s shares immediately less valuable. It’s perhaps no surprise, therefore, that the company’s newly listed shares fell 16% Monday and are now trading below their IPO price.

SpaceX is offering investors who buy its bonds up to 1.75 percentage points over U.S. Treasuries of equivalent maturity, a premium to what bonds of similar ratings offer. Specifically, SpaceX is selling $7.0 billion in notes paying 5.35% due in 2031, $6.0 billion at 5.65% due 2033, $6.0 billion at 5.875% due 2036, $2.5 billion at 6.6% due 2046, and $3.5 billion of 6.65% Senior Notes due 2056. It’s using the proceeds to pay off a variable-rate bridge loan due next year, meaning it’s spreading out its debt repayments and locking in its repayment costs. Investors may thus be freaking out for no reason: the new bonds don’t boost SpaceX’s overall debt, they spread it out over time, ostensibly reducing the company’s risk of default and putting it on firmer financial footing—especially with $86 billion of new equity to buttress the denominator in its debt-equity ratio.

But SpaceX isn’t the only one hitting investors up for loans. Tech companies have borrowed more than $150 billion so far this year, more than double what they did over the same period last year and have issued more debt in 2026 than in any other year since 1999. That’s a lot of moolah to owe.

Those kinds of concerns were behind Tuesday’s rout in tech stocks, which pulled the tech-heavy Nasdaq index down 2.2% and helped drag the benchmark S&P 500 1.4% lower. But it was chipmakers, including Micron, Nvidia, Qualcomm, Sandisk, and Western Digital, that led the decline.

Analysts have long fretted that Big Tech’s so-called “Magnificent Seven”—Alphabet, Amazon, Apple, Meta, Microsoft, Nvidia and Tesla—had grown to dominate the broader market. Their combined market capitalization now constitutes almost a third of the entire stock market’s value. That means worries about AI-related debt can easily move the entire market.

But what if AI is out to get us? Such a scenario might be just months away. AI from the West’s enemies may soon be unleashed against the biggest economic institutions we have—notably governments and companies. That’s according to a new warning from cybersecurity officials from the U.S.-led “Five Eyes” intelligence partnership, which includes Australia, Canada, New Zealand, and the United Kingdom.

In a rare, joint communiqué, the group warned that “frontier AI models” may be able to catch up with the West’s lead in AI to the extent that they can launch effective AI-powered cybersecurity attacks. “Frontier AI models are anticipated to exceed current industry expectations, fundamentally transforming both offensive and defensive cyber capabilities,” the group said.

AI-powered attacks have already been launched, and cybersecurity experts say they’ve been getting increasingly sophisticated. In May, Google said it blocked an attack from what was believed to be Russia-backed groups that would adapt automatically to locate and exploit security vulnerabilities. It said it had foiled similar attacks from China and North Korea.

The Five Eyes cybersecurity group warned companies to integrate AI into their own security measures. “Those who delay will face growing and avoidable risk,” it said.

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes:

<a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>