The United States is undertaking a historically unprecedented investment boom to build the computers, data centers, and other physical infrastructure needed to train and deploy Artificial Intelligence. Hundreds of billions of dollars have already been spent by hyperscalers racing to build smarter AI systems, and investment from major tech companies is set to shatter all previous records again this year. Amidst this boom, spending on data center construction has hit a new record high, now exceeding a $42B annualized pace, a more than 300% increase since the launch of ChatGPT in late 2022. While growth has slowed over the last six months, investment is still up more than 18% over the last year alone.

Yet that figure only reflects the costs to build data center facilities themselves, not the much larger costs of the expensive GPUs, TPUs, and other electronics housed within. Real US fixed investment in those computers and related peripheral equipment has surged to a record high of more than $270B annualized, up nearly 50% over the last year and up 77% since ChatGPT’s launch.

… America is making a globally and historically unprecedented bet on the success of Artificial Intelligence. As a share of the economy, that AI boom is already one of the largest investments in American history—dwarfing the peak of the broadband, electricity, or interstate highway buildouts and vastly exceeding the Manhattan or Apollo projects. And yet, US tech companies are doubling down, raising the stakes on their $1T gamble that AI models will continue their exponential capabilities growth and eventually become valuable enough to repay such a colossal investment. — Read More

The week AI killed software

Last Monday, $285 billion of market cap evaporated from software, financial services, and asset management stocks. Thomson Reuters lost $8.2 billion. In a single day. LegalZoom dropped 20%. India’s Nifty IT index posted its worst month since October 2008 — worse than the financial crisis.

he week AI killed softwarexxxxLast Monday, $285 billion of market cap evaporated from software, financial services, and asset management stocks. Thomson Reuters lost $8.2 billion. In a single day. LegalZoom dropped 20%. India’s Nifty IT index posted its worst month since October 2008 — worse than the financial crisis. — Read More

Years ago, some leading researchers told us that their objective was AGI. Eager to hear a coherent definition, we naively asked “how do you define AGI?”. They paused, looked at each other tentatively, and then offered up what’s since become something of a mantra in the field of AI: “well, we each kind of have our own definitions, but we’ll know it when we see it.”

This vignette typifies our quest for a concrete definition of AGI. It has proven elusive.

While the definition is elusive, the reality is not. AGI is here, now.

Coding agents are the first example. There are more on the way.

Long-horizon agents are functionally AGI, and 2026 will be their year. — Read More

#investing

The AI industry is in correction mode. Last week Nvidia reported its earnings and the world was holding its breath. If they miss, it is so over. If they crush it, we are so back. In the end, earnings beat expectations, but the stock slid anyway after an initial bump. Many things in the AI boom smell bad. The way money fuels the investment spree is quite questionable and it has become a meme, with the same $1T investment check moving hands among a small set of participants. We can call this “vendor financing”, but it is not a great look.

In my opinion the players more at risk here are the hyperscalers like Microsoft, Amazon and Oracle, and the neocloud players like Nebius and CoreWeave. They are in between the providers of chips like Nvidia and the buyers of compute like OpenAI. They really have no choice other than buying real chips from Nvidia, and hoping that there will be sustainable demand (read: revenue > COGS) from buyers of compute so that they can honor those commitments. If not, the buyers of compute will walk away, resizing their commitments (or going bankrupt), Nvidia already sold those GPUs, and the hyperscalers are left with billions and gigawatts of unused capacity that depreciate very quickly due to the short GPU lifespan. — Read More

The artificial intelligence revolution will be only three years old at the end of November. Think about that for a moment. In just 36 months AI has gone from great-new-toy, to global phenomenon, to where we are today – debating whether we are in one of the biggest technology bubbles or booms in modern times.

To us what’s happening is obvious. We both covered the internet bubble 25 years ago. We’ve been writing about – and in Om’s case investing in – technology since then. We can both say unequivocally that the conversations we are having now about the future of AI feel exactly like the conversations we had about the future of the internet in 1999.

We’re not only in a bubble but one that is arguably the biggest technology mania any of us have ever witnessed. — Read More

A month ago, I set out to answer a deceptively simple question: Is AI a bubble?

Since 2024, people have been asking me this as I’ve spoken at events around the world.

Even as Wall Street bankers largely see this as an investment boom, more people are asking the question in meeting rooms and conference halls in Europe and the US.

Some have made up their minds. 

Gary Marcus called it a “peak bubble.” The Atlantic warns that there is a “possibility that we’re currently experiencing an AI bubble, in which investor excitement has gotten too far ahead of the technology’s near-term productivity benefits. If that bubble bursts, it could put the dot-com crash to shame – and the tech giants and their Silicon Valley backers won’t be the only ones who suffer.” The Economist said that “the potential cost has risen alarmingly high.”

The best way to understand a question like this is to create a framework, one that you can update as new evidence emerges. Putting this together has taken dozens of hours of data analysis, modeling and numerous conversations with investors and executives.

This essay is that framework: five gauges to weigh genAI against history’s bubbles. — Read More

In just six months, the consumer AI landscape has been redrawn. Some products surged, others stalled, and a few unexpected players rewrote the leaderboard overnight. Deepseek rocketed from obscurity to a leading ChatGPT challenger. AI video models advanced from experimental to fairly dependable (at least for short clips!). And so-called “vibecoding” is changing who can create with AI, not just who can use it. The competition is tighter, the stakes are higher, and the winners aren’t just launching, they’re sticking.

We turned to the data to answer: Which AI apps are people actively using? What’s actually making money, beyond being popular? And which tools are moving beyond curiosity-driven dabbling to become daily staples?

This is the fourth installment of the Top 100 Gen AI Consumer Apps, our bi-annual ranking of the top 50 AI-first web products (by unique monthly visits, per Similarweb) and top 50 AI-first mobile apps (by monthly active users, per Sensor Tower). Since our last report in August 2024, 17 new companies have entered the rankings of top AI-first web products.  — Read More

There were a lot of speculation over the past years about what the next cycle of AI development could be. Agents? Reasoners? Actual multimodality?

I think it’s time to call it: the model is the product.

All current factors in research and market development push in this direction.

— Generalist scaling is stalling. 
— Opinionated training is working much better than expected. 
— Inference cost are in free fall. 

This is also an uncomfortable direction. All investors have been betting on the application layer. In the next stage of AI evolution, the application layer is likely to be the first to be automated and disrupted. — Read More