What to expect from us in 2025

By Fred Vogelstein and Om Malik

This is a story about three dates: 1994, 2007, and 2022. And then three more dates: 1996, 2009 and 2025. 

1994 was the year Netscape unveiled the first mainstream internet browser. 2007 was the year Apple unveiled the iPhone. And 2022, a hair over 2 years ago, was the year Open AI unveiled ChatGPT. There are many ways to frame the digital revolution of the past 30 years. But to us, focusing on these three eras makes the most sense. 

Why the history lesson? Because we’re approaching some inflection points in the current innovation cycle, roughly two years in. We think it helps to think back to how we thought about the future two years into the 1994 and 2007 cycles to better understand where we are in the current one. 

In 1996, the internet bubble had barely inflated. Google hadn’t been founded yet. And the famous Microsoft antitrust trial was still three years away. In 2009 the app economy was less than a year old. Cloud computing was barely a thing. And few talked about the gig economy because companies like Uber and AirBnB were too small to have entered the zeitgeist.

But wow, the technology revolutions that followed five years after 1996 and 2007 reshaped Silicon Valley, along with business and society at large. We think we’re on the cusp of that happening again - a third technology innovation frenzy not seen in half a generation. We should expect new super profitable companies to emerge from this. We should also expect older, bigger companies like Google, Apple, Amazon, Meta, and Microsoft - the disruptors during previous eras - to be scrambling to adapt or die. 

Obvious? Maybe to some. But tracking the technology driving this revolution is why we started this newsletter five months ago. And since some of you have asked us what we’ve learned so far and for more detail about the lens we are using to think about the rest of 2025 and beyond, we thought we’d give you an answer.  

For starters, we think we’ve entered a new phase of the AI revolution in the past three weeks. The AI bubble has made global markets more reactive than they’ve been for a very long time. And it’s sent businesses and governments worldwide scrambling to stake out positions for fear of getting boxed out by rival businesses and nations. 

There isn’t even widespread agreement on how powerful DeepSeek’s innovations are. Demis Hassabis, the Nobel Prize chemist who runs Google’s DeepMind just said he thinks its expected impact is exaggerated.  Yet it remains a hot topic of conversation two weeks after version R1 was released. Even though US tech stocks have stabilized, the promise of Chinese AI dominance is driving a massive rally in Chinese stocks. 

Meanwhile, the Paris AI summit last week suggests that national and international politics will be an additional vector for investors and entrepreneurs to consider.  The browser and smartphone technology revolutions happened largely free of politics. The AI revolution already has begun to more resemble the space race between the US and Soviet Union in the 1960s. 

So we want to help you better understand which new technologies are likely to be the most disruptive - and why, in geeky detail. Good ideas and vision matter a lot in Silicon Valley. And during bubbles - we are in an AI bubble now - that’s often enough to launch a company. But what makes companies enduring when bubbles pop are those who also understand the underlying technology better than anyone else. 

We started our history lesson with the browser in 1994 because, notwithstanding the innovation before it,  it’s widely considered the starting point for the modern internet. It drove massive personal computer adoption, igniting our now insatiable demand for storage, bandwidth, and processing power. That drove a virtuous innovation cycle still with us today: Every couple of years the price per unit of storage, bandwidth and processing power gets cheaper while bandwidth and processing power get faster and manufacturers think of new ways to make the storage form factor smaller.

Without the leverage of the free open source software on generic x86 servers, you don’t get foundational hyperscale Silicon Valley companies like Google, Amazon, and Meta. Without cheap hard drives to go into iPods, and cheap broadband to download music, you don’t get Apple's resurgence. 

In 2007, the smartphone put all that cheap storage, abundant bandwidth, and ever increasing processing power in our pockets. The smartphone wasn't just a personal computer that fit in your pocket, it collapsed a dozen industries into one device - the book, the newspaper, the telephone, the radio, the camera, the video camera, the television, the VCR/DVD, the voice recorder, the music player, the video game, and the compass. 

The smartphone’s secondary and tertiary effects were also massive. Its convergence with cloud computing, and social networking helped create a thriving app economy. That in turn supercharged cloud computing and social networking. Its impact was so profound that eight years after the smartphone appeared Apple, Meta, Google and Amazon had joined Microsoft among the ten most valuable companies in the world.  

It created new industries like ride hailing, home rentals, influencer marketing, and mobile gaming. And it has permanently changed how we use money, travel, communicate, visit the doctor, shop and entertain ourselves.  Everything we used to do at home or at work, suddenly got wheels. It turbocharged the world, and also Silicon Valley. We are still feeling its after effects

We picked 2022 to start the most recent revolution because that's when ChatGPT from OpenAI was released. Insiders will tell you the AI revolution started with the publication of the transformers paper in 2017. But the rest of the world only started to appreciate the power and importance of AI when they used ChatGPT for the first time. And as we’ve said before, the anticipation around what AI might be able to do, has created a business valuation bubble not seen since the internet bubble in the late 1990s. 

We’re not going to try and predict how the daily gyrations in Washington DC will affect us long term. A lot of it is alarming. At the moment, it looks like a power grab hell bent on destruction but with little explanation about the overall long term policy goals.  But if professionals on both sides of the aisle are struggling to understand it, we’re not going to try. 

But we are going to take a stab at laying out what critical technology shifts are worth paying attention to for the rest of this year. You’ve already gotten a taste: How will AI affect the internet browser, social networking, copyrights, biology, healthcare, software development, augmented reality and the devices we use every day? 

Here are some additional questions we’ll be asking this year: 

What will competition from Chinese AI firms like DeepSeek and others look like? What are the national security implications? Will every country soon have its own LLM with answers rooted in their own interpretations of history and create the web balkanization many have feared for years?

What happens to Google search and web search generally if AI chatbots instead of a search box become the first place we go online? How will our internet browsers evolve to tackle these changes? Do we even need one anymore? 

What about the way we interact with our machines? The keyboard, like the paper book, has remained a persistent interface. But it sure has gotten easier and easier to interact with machines in other ways that have nothing to do with entering text on a screen - with touch, with voice, and with smartphone cameras. Look at the success of the Apple watch. Apple and Google have already built AI into their photo and cameras app. Yes, it’s still hard to imagine a world where our primary computer is built into our eyeglasses or a lapel pin as Meta is trying with its Ray-Ban partnership and Humane is trying with AIPin.  On the other hand, the internet created a wave of new ways for us to interact with our machines. The smartphone created another wave. It would be hard to argue that the AI revolution won’t do the same.

How sticky will our various social grafs be in this new world? Those network effects have proven super powerful and enduring over the past 20 years. But I don’t think it’s a coincidence that in the past month we’ve heard three entrepreneurs talk about how new open source protocols and identity standards are finally making it possible for us to build one social graph we control and use everywhere.

How will AI change drug discovery and healthcare - two industries with arguably the most to gain? We’re seeing companies like Abridge get traction in the nettlesome world of AI transcription. We’re also watching AI radically speed up the trial and error rate in the biotech industry, allowing drug targets to be isolated faster and enabling faster approvals.

And finally what does the end of the AI bubble look like? If it looks anything like the end of the internet bubble, it’s going to be ugly. Back then an entire generation of companies were built assuming expensive storage, processing power and bandwidth. Those that didn’t run out of money faced a new generation of competitors building at a fraction the cost. 

An event like that could turn companies like OpenAI into the biggest money pits ever created.  OpenAI just raised almost $40 billion at a roughly $300 billion valuation - ranking it about 50th in market cap next to Bank of America. What if we wake up one day and someone has done it better for half the price? It’s certainly happened before.

How much is ahead? If 2025 is like 1996, that was when Silicon Valley existed entirely in the shadow of Microsoft. The internet bubble had four more years to run. The U.S. vs Microsoft antitrust trial wouldn't start for another three years. Amazon was two years old and merely sold books online. Apple was on the verge of going out of business, and Steve Jobs hadn’t yet returned. Google and Meta hadn't even been founded yet. Then, when the internet bubble popped and Microsoft became distracted by antitrust attacks, that unleashed the flurry of innovation that birthed what remains the foundations of Silicon Valley today.

And those changes were arguably minor compared to what happened after Apple unveiled the iPhone in early 2007. The mobile and social revolution catapulted Silicon Valley’s companies from weirdo California nerd status to the most important and powerful companies on the planet. But again, it was hard to see that coming In 2009. The app economy was less than a year old. The iPhone had only just begun generating attention. But we were a long way from the iPhone  and smartphones running Google’s Android becoming the dominant computing platforms in the world. 

Android smartphones barely existed in early 2009. The same thing was true about Twitter. Meta was still just an interesting private startup rather than one of the biggest, most controversial companies on the planet. It hadn't yet purchased Instagram, and Oculus because they didn't exist. What’sApp and Uber had just started. Airbnb was just two years old. Amazon was still seen as just an online retailer instead of an engine room of the modern economy. Microsoft looked like a company that had lost its way. It was hard to imagine how these five companies would become among the most important firms in our economy, society and politics.

We’re not polyannas about the technology future. We understand that as factories, hospitals, warehouses, etc. use robots and AI more, a lot of jobs will disappear. AI obviates the need for as many high priced software developers too. All you have to do is take a Waymo cab in San Francisco, Los Angeles, Phoenix or Austin to understand that AI will soon kill off most of the taxi and chauffeur industry. Even Lyft and Uber aren’t safe. But while every technology revolution in history has killed off jobs, it’s also created new ones. 

A prevalent theme in the late 1960s was that the mainframe computer would turn American offices into ghost towns. Not only did that prediction not come true, computerization caused an explosion in new IT jobs. We’ve heard the arguments that say this time is different, or that AI will leave us only with soul-sucking, menial jobs. If that happens, it will be the first time in history that that prediction came true.   

We like to think about where we are right now like this:  1994 was the start of getting to all the information. 2007 was the start of getting everything else, anywhere, instantly. 2022 was the start of getting everything, anywhere not just instantly, but also personalized, augmented and automated.

That’s what we will be writing about for the remainder of the year and beyond!