Bill Gurley, Artificial Intelligence, and the reset that could shake the entire market
Bill Gurley, partner at Benchmark and one of Silicon Valley’s best-known investors, is once again drawing attention to a topic that’s been keeping a lot of people in tech and finance up at night: the possibility of a bubble forming around Artificial Intelligence and software stocks.
In an interview on CNBC’s Money Movers, Gurley summed up the current moment well: a lot of people got rich very fast with the AI wave, which attracted even more investors and entrepreneurs trying to repeat that story. For him, this combo usually leads to the same place: the formation of bubbles and, inevitably, a market reset.
The core point is not denying the technological revolution of AI. On the contrary: Gurley reinforces that the wave is real, deep, and has massive economic impact. The problem lies in the excess of speculative capital piling on top of this trend, distorting prices, expectations, and company valuations until the market decides to recalibrate everything at once.
When too many people get rich quickly, the bubble risk goes up
According to Gurley, there is a pattern that repeats itself at different moments in economic history. As soon as a relatively small group makes a lot of money in a short time with a new technology, it triggers a psychological switch in the market: more people jump in, often without really understanding what they are buying, driven mainly by the fear of missing out.
That is exactly when bubbles tend to form, he explains. The logic stops being:
- How much value does this company actually create?
- What is the sustainable business model?
- What are the margins, risks, and competitors?
And turns into something much more superficial, almost like a mantra: AI is taking off, so any AI-related asset should go up. That kind of thinking paves the way for valuations to drift away from reality for quite a while, until some event – like weak results, higher interest rates, or a narrative shift – forces a reset.
Gurley is not speaking from intuition alone. He cites the work of economist Carlota Perez, author of Technological Revolutions and Financial Capital: The Dynamics of Bubbles and Golden Ages. Her thesis is that major technological revolutions always come with cycles of euphoria and correction.
One key point Gurley highlights when citing Perez is this: bubbles only exist when the underlying technology wave is real. In other words, the existence of excessive speculation does not mean the technology itself is a scam. What happens is that the market tries to price in decades of transformation in just a few years, creating distortions that later need to be corrected.
Receive the best innovation content in your email.
All the news, tips, trends, and resources you're looking for, delivered to your inbox.
By subscribing to the newsletter, you agree to receive communications from Método Viral. We are committed to always protecting and respecting your privacy.
Reset ahead: where software opportunities may emerge
When Gurley talks about a reset, he is not just predicting drops or corrections. He also points to a specific type of opportunity that could emerge in this scenario: software-as-a-service (SaaS) stocks that have been crushed in the market and are now too cheap relative to their business potential.
In his view, once the dust starts to settle, it is worth having a target price or a reasonable value range in mind for those stocks. The idea is simple: if the market overshoots on the downside, there may be a moment when companies with recurring revenue, solid products, and strong customer bases are traded as if they had no future.
In this context, Gurley suggests that more attentive investors might start “swallowing” these stocks, taking advantage of temporary pessimism. But of course, that depends on fundamentals analysis, not just chasing trends.
Software on the front line of the hit
While enthusiasm around AI grows, the traditional software sector has been taking a beating in the stock market. Gurley draws attention to this contrast: the technology that consumes the most software and infrastructure is at peak hype, but many publicly traded software companies are going through heavy repricing.
Some recent numbers help illustrate this scenario:
- Salesforce has already lost about 25% of its value in 2026.
- ServiceNow is also down around 25% for the year.
- The iShares Expanded Tech-Software Sector ETF (IGV), which tracks the software sector, is down roughly 20% in 2026.
In other words, while the AI narrative dominates conferences, keynotes, and timelines, a meaningful slice of the software companies that underpin this ecosystem is being aggressively repriced. This mismatch between narrative and price is exactly the kind of signal that usually shows up during cycle transition periods.
The invisible cost of the AI race
Another point Gurley emphasizes is the level of spending by big tech companies to sustain the current AI race. It is not just about hiring teams or launching new products: we are talking about colossal investments in infrastructure, chips, and memory.
According to recent projections, just Amazon, Meta, Google, and Microsoft together are expected to spend around 700 billion dollars on AI this year alone. That figure covers everything from data centers and cutting-edge GPUs to high-bandwidth memory, in a scenario where the costs of this kind of component have surged.
On top of that, the industry has been facing challenges with a shortage of memory for AI, especially in technologies like HBM (High Bandwidth Memory), used in boards and systems optimized for training and running large models. This combination of explosive demand, limited supply, and record investments squeezes margins, forces heavy strategic decisions, and increases the risk that some bets will take a very long time to pay off.
For outside observers, this level of spending may seem abstract, but it has very real effects:
- Companies need to justify internally why they keep pouring billions into infrastructure.
- Any sign of slowdown in AI-related revenue creates extra concern among shareholders.
- More fragile businesses, without a strong cash position, can fall by the wayside if monetization takes too long.
From Uber to AI: the view of someone who has dealt with extreme cash burn
Gurley talks about cash burn with the background of someone who has been through intense experiences. Benchmark was one of the early investors in Uber, and he played a relevant role in the process that led to the departure of then-CEO Travis Kalanick in 2017.
At the time, Uber had an annual burn rate of about 2 billion dollars. According to Gurley, that was already a constant source of anxiety. He described that period as “high anxiety”, both because of the size of the numbers and the level of uncertainty around when the company would start generating sustainable cash flow.
Now, looking at the leading companies in large AI models, such as Anthropic and OpenAI, he sees even more aggressive spending over very short time frames. And he drops a comment that neatly sums up how he feels about this model:
God bless those companies. It is a scary way to run a business.
It is not an empty criticism, but a warning. When fixed monthly costs are enormous and depend on external capital to keep going, any shift in funding market sentiment or in projected revenue can have dramatic effects. In boom times, that might seem manageable; in tighter periods, it becomes a systemic risk for startups and even for some already established companies.
Why Carlota Perez’s view helps make sense of the current moment
When Gurley turns to the work of Carlota Perez, he is basically proposing a historical framework for what is happening with AI today. The central idea is that technological revolutions tend to follow a similar script:
- A new technology emerges and starts to prove its practical value.
- Financial capital identifies the potential and pours in, accelerating expansion.
- Euphoria pushes prices up faster than concrete results.
- At some point, expectations and reality collide, triggering a correction phase.
- After the bubble bursts, the technology continues to be adopted in a more mature and widespread way.
In that context, Artificial Intelligence seems to be somewhere between the euphoria phase and the beginning of a cycle with more questions about return on investment, especially for companies that rely on very expensive infrastructure or on business models that are still largely unproven.
That does not mean AI will lose relevance. Quite the opposite: the trend is for it to spread even further, being integrated into cloud services, software platforms, consumer apps, productivity tools, and user interfaces in general. What is likely to change is how much the market is willing to pay upfront for that promise.
Real AI, financial hype, and the role of correction
Gurley’s message, in the end, has two sides. On the one hand, he reinforces that the AI wave is real and should continue to transform entire sectors of the economy – from customer support to software development, passing through design, marketing, healthcare, education, and beyond.
On the other hand, he warns that the current environment combines:
- Real cases of productivity gains and innovation;
- Overabundant capital betting on any AI-related narrative;
- Software and SaaS companies being aggressively repriced;
- Big Techs burning hundreds of billions of dollars to lead the race;
- Frontier-model startups running with frightening cost structures.
In this scenario, a reset is not just likely: it may be healthy to separate solid projects from purely speculative initiatives. Price corrections, even if painful in the short term, often create room for companies with stronger fundamentals to gain prominence, while those that depended solely on hype tend to fall behind.
For anyone following technology, especially AI, it is worth paying attention not only to announcements of new models and integrations, but also to the financial signals behind them: infrastructure costs, real adoption pace, impact on revenue, margins, and the ability to keep the engine running without eternally depending on cheap capital.
In the end, Gurley’s perspective – and that of scholars like Carlota Perez – helps us look at the current moment with a bit more calm. The Artificial Intelligence revolution is underway and is unlikely to move backward, but that does not prevent the market from having gone too far, too fast in pricing some of these stories. When the reset hits, those who have managed to distinguish real technology from financial delusion will probably have a much clearer view of where the long-term opportunities really are.
