AI summary
- On 5 November the Philadelphia Semiconductor Index fell 12.25%, its worst day since March 2020, and one report's headline puts the loss across AI infrastructure stocks at over $1 trillion.
- Investor Bruce Liu says the trigger was Texas, where Gina Hinojosa beat Greg Abbott, and argues the market confused construction delays with a collapse in demand.
- CIBC says that if regulation tightens, it would cut its exposure to AI infrastructure suppliers and favour large cloud platforms and software companies.
- The SI label rests on an executive order that only binds federal agencies, and Democrats now control Congress.
- A stock rout isn't a reason to buy a .si: only buy one when the .com and .ai are taken or expensive, the words are generic and people search for them.
What happened to AI stocks on 5 November
On November 5, the Dow Jones Industrial Average fell 1,833 points in a single session and the Philadelphia Semiconductor Index dropped 12.25%. According to BigGo Finance, that was the index's worst one-day performance since March 2020. NVIDIA fell more than 9%, CoreWeave dropped 14.6% and Nebius fell 18%. The article's headline puts the loss across AI infrastructure stocks at over $1 trillion.
I've already written about the midterms and data centres, so this post isn't a rerun. It's about how the market reacted, and what that reaction does and doesn't tell someone deciding on a .si name.
Why Texas, not Congress, set it off
Most people assumed the Democrats' win in both the Senate and the House caused the drop. The investor Bruce Liu, speaking on Silicon Valley 101, said prediction platforms had already priced that outcome in. In his view the real trigger was the Texas governor's race. Democratic state Representative Gina Hinojosa beat the Republican incumbent, Greg Abbott, after prediction markets had held her odds below 40%. Liu called the selloff "the release of Gina risk."
Texas matters because it already has the second-most data centres in the US, as the Key West Citizen reports. The same piece quotes the analyst Nic Puckrin: "Gov. Gina Hinojosa wants an immediate, statewide moratorium on new data center approvals." Yet when CNN's Jake Tapper interviewed her after the election, she said she doesn't oppose data centres but wants more regulation so they benefit the people. Liu thinks she's more likely to take a middle path.
Delay or collapse? Two guests push back on the panic
The core argument in the BigGo piece is that the market mistook "construction delays" for "demand collapse". The energy consultant Ethan Xu predicts a severe US data centre shortage in 2027 and 2028. He says the real risk lies in earnings downgrades at upstream hardware makers, such as GPU and optical module suppliers. Liu points to Amazon's guidance of 40% growth over two years, its first ever, as the test. If the US can't build new data centres, he says, that target would fall to about 12% a year.
Sam Tabar, CEO of WhiteFiber, told the Key West Citizen that "The change in governor will not stop Texas from being a major data center market." Still, the opposition is real. In exit polls, 76% of voters said they oppose data centres being built in their local communities. Interconnected Capital counts 372 data centre moratoriums and 43 permanent bans across the US.
Where the money is moving: software over concrete
According to Yahoo Finance, CIBC Capital Markets built two stock baskets before the vote. Its Democratic basket holds 36 stocks. Its Republican basket holds 45 names and leans towards AI. Christopher Harvey, CIBC's head of equity and portfolio strategy, wrote that if regulation tightens, the bank would cut its exposure to the suppliers building AI infrastructure and favour large cloud platforms and software companies. He also warned against writing off the data centre boom.
The same article reports that Mark Zuckerberg said Meta's quarter delivered major milestones toward "personal superintelligence for everyone", including the launch of Muse and Muse Wonder. So bosses are still talking about superintelligence, even while hardware stocks fall.
Here's my read. The pressure is on power, permits, land and chips. Domain names sit on the other side, with software and products, where brands and search terms live. That doesn't make any domain safer, but it does mean a chip selloff tells you nothing direct about what names are worth.
What this means for the SI label itself
I'd watch this part more closely than the share prices. An essay on note published on 11 November points out that the executive order only applies to federal government agencies. It can't legally force the media, companies or other countries to say "SI". The Wall Street Journal still uses the old term, and even Newsmax still calls it "AI". The essay says MIT Technology Review reports that federal websites are gradually switching to "SI".
Democrats now control Congress, and BigGo notes that Bernie Sanders has proposed S.5493 to restrict superintelligence, with penalties for individuals of up to 20 years in prison. None of my sources say the new Congress plans to touch the SI wording. But the label rests on a single executive order, and in a September Quinnipiac poll 58% of respondents said they don't support Trump's handling of AI. I wouldn't bet much money on that label alone.
Should you buy a .si name because of this?
No, not because of a stock rout. Most .si names are speculation and the .si aftermarket is small, so don't assume you'll resell quickly. If the .com for your words is available at a fair price, buy the .com. It's the better buy for most businesses.
A .si only makes sense when the .com and .ai for the same plain words are taken or expensive, and the words are generic rather than a copy of one company's brand. Put every name through three questions:
- Are the .com and .ai taken? If either is free or cheap, buy that instead.
- Is it a generic name, not a brand? Plain words only. Never copy a company's name.
- Do people search for it? Real monthly searches mean real demand from end users.
Here's how a few names from my inventory do on that test. I've stuck to their data and added nothing:
- aimarket.si ($4,500): the .com is listed at $594,888 and the .ai at $47,995, with 12,100 searches a month.
- bankability.si ($2,500): the .com has been parked since 2002 and the .ai is listed at $49,800. Lenders and project-finance teams use the word every day, and a US click costs $15.26.
- aiprompts.si ($5,500): 201,000 searches a month worldwide, the .com has been held since 2018 and the .ai is listed at $44,500.
- webbrowser.si ($5,000): 165,000 searches a month, the .com has been held since 2002 and the .ai is listed at $48,900.
- futurehuman.si ($1,500): 5,400 searches a month, the .com has been held since 1999 and the .ai is listed at $290,000.
If a name fails any of the three questions, I'd walk away, whatever the market did this week.
Names from this store that pass the test
Questions
What caused the AI stock selloff on 5 November?
Investor Bruce Liu said the real trigger was the Texas governor's race, where Gina Hinojosa beat Greg Abbott, and he called it "the release of Gina risk." That day the Philadelphia Semiconductor Index fell 12.25%.
Does Gina Hinojosa oppose data centres?
Analyst Nic Puckrin said she wants an immediate, statewide moratorium on new data centre approvals. But when CNN's Jake Tapper interviewed her after the election, she said she doesn't oppose data centres and wants more regulation so they benefit the people.
Does the executive order force companies to say SI?
No. The executive order only applies to federal government agencies and can't legally force the media, companies or other countries to use "SI". The Wall Street Journal and Newsmax still say "AI".
Is a .si name a good buy after the selloff?
Not because of the selloff. Most .si names are speculation and the aftermarket is small, so only consider one when the .com and .ai are expensive. For example, aimarket.si is $4,500, while its .com is listed at $594,888 and its .ai at $47,995.
Sources
- BigGo Finance, 2026-11-12
- Yahoo Finance, 2026-11-12
- Key West Citizen, 2026-11-09
- note, 2026-11-11
This post was researched and drafted with AI help from the sources above, then checked by me before publishing. Facts are as reported on the dates shown.