Let me add my ‘Happy New Year’s Day’ Wishes to the flood in front of us all this Friday before a holiday weekend. As I outlined yesterday, this AI Tech Wave is lined up for a similar but larger 2026 for all things AI relative to 2025. Have gone through our 2025 look-back and 2026 look forward. And it feels like it’s going to be more of the same, just more. Like the Volume turned up to 11.
The big question all year of course has been the ‘AI Bubble’ one. And all things considered, pros and cons, we’re likely not quite there. Despite anecdotal data points around and beyond AI. But timing on these things is always hard.
“For an eventful year, 2025 is ending the same way it started. A year ago, Nvidia, OpenAI and others made it clear they would do everything necessary to win the race for artificial intelligence.”
“The year ended with the same companies doing the same thing. Nvidia capped off a year of dealmaking—mostly to cement its lead in AI chips by agreeing to pay $20 billion to license technology from Groq, one of its best-funded challengers. OpenAI is ending the year with a bang too, aiming to raise $100 billion in capital and in talks with Amazon for a $10 billion investment.”
“A lot did change in the last 12 months, and some of those changes could become challenges in 2026. The biggest shift is that AI went from interesting but hypothetical for most people into a reality of everyday life.”
“In U.S. financial markets, AI became a powerful force. It also fed inflation by pushing up electricity prices. On the ground, it generated more grumpiness than goodwill, with communities and local officials complaining about data centers and people blaming AI for the economy’s weak job growth. Warranted or not, the companies building AI will have to deal with the fallout in 2026.”
And the deals got more fantastic and frenetic well into the last days of 2025.
“Another thing that changed: The deals got more head-scratchier, as we’ve reported. OpenAI’s quest for $100 billion raises an obvious question: Where do you get $100 billion? Nvidia’s deal for Groq raises another one: Why pay $20 billion for a company that was valued at $6.9 billion just a few months ago, and not get the whole company?”
“As I’ve written, I don’t think we’re in an AI bubble right now. That’s largely because the main funders of the buildout are the richest, most valuable companies in the world. If the Groq deal amounts to nothing, it will be little more than a scratch on Nvidia’s finances and reputation.”
“And when those cash-rich giants are not standing behind a deal, investors are now behaving pretty rationally. CoreWeave, the cloud provider that is a leveraged bet on AI, had a spectacular IPO. But investors sobered up when it struggled with data center construction delays and when doubts grew about the scale of its borrowing. The stock is down 60% from its peak.”
“In retrospect, the peak of AI euphoria may have occurred on Sept. 10 when Oracle added nearly $250 billion in market capitalization after reporting a huge increase in future contract revenue, largely tied to AI. Since then, Oracle is down 40% and that $250 billion is long gone.”
“Now, the market is quick to deliver its harsh judgment. Fermi America, the young data center developer cofounded by former Texas Governor Rick Perry, had a successful IPO a few weeks after Oracle’s big day, when AI euphoria was still in the air. But Fermi said in mid-December it lost a data center tenant. Its shares fell 40%. Something pretty amazing will need to happen for investors to fall in love with CoreWeave, Oracle or Fermi again.”
“The worst investor behavior may be over, but you can bet that market strategists and financial advisers are still talking about an AI bubble with their clients. The advice is the same: diversify. The seven giant tech stocks account for 36% of the S&P 500’s value and 26% of its earnings, according to Goldman Sachs. What about the other 493 stocks, or the rest of the world for that matter?”
“If investors do diversify, it will mean selling tech stocks and buying everything else, including foreign stocks, which collectively trounced the tech giants in 2025.”
“Another headwind could be interest rates. Everyone in Silicon Valley remembers the wreckage wrought by the Federal Reserve when it raised interest rates in 2022. Now, the Fed is cutting short-term interest rates. But recently, long-term rates haven’t followed them down.”
Especially given the margin and balance sheet math concerns around these AI investments that continue and accelerate this year.
“If President Donald Trump’s new appointee to lead the Fed loses credibility with investors, they could start to worry about inflation and push up long-term rates. That could squeeze any remaining speculation out of the AI buildout.”
“The last potential shift is politics. The tech industry went into 2025 with a ton of political power in everything from AI to crypto to antitrust regulation. Now the industry faces fears of AI-driven job losses, higher electricity prices, and a boom and bust in speculative crypto that has cost investors—not the crypto industry—dearly. These are all shaping up to be issues in the midterm elections.”
“Maybe 2026 will start and end just like 2025 did. Probably not.”
Lots to ponder indeed going into the new year in this AI Tech Wave. Eyes wide open for us all starting monday. Stay tuned.
(NOTE: The discussions here are for information purposes only, and not meant as investment advice at any time. Thanks for joining us here)
Good afternoon. I get a lot of requests to share updates from/about our portfolio companies and SPV’s that we have led over the years. While we share them...
It’s now sunk in that Big Tech companies are spending billions to build AI infrastructure things at Scale. A trend that has no end point for a while....
