Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
Open the editor’s digest for free
FT editor Rula Khalaf picks her favorite stories in this weekly newsletter.
There was nothing new to match the “wow” factor when we first used ChatGipt in 2024, but rapid improvements in the underlying technology still kept the field on edge. For 2025, this is how I see things panning out.
Will AI development hit a wall?
In the year By 2025, that pace will disappear. Some of the tech industry’s biggest optimists have acknowledged that in recent weeks, dumping more data and computing power to easily train large AI models — a reliable source of improvement in the past — has begun. Diminishing return. In the long run, this robs AI of a reliable source of improvement. However, at least in the next 12 months, other developments should outweigh the fatigue.
The most promising advances seem to come from models that take a series of steps before returning the answer, allowing them to refine initial responses to questions and provide more “reasonable” results. It’s debatable whether this really compares to human reasoning, but systems like OpenAI’s o3 still seem like the most exciting advancement since the advent of AI chatbots.
Google, which AI got the mojo again After two years of struggling to keep up with OpenAI at the end of the year, it also showed how new agent-like capabilities in AI can make life easier, such as tracking what you’re doing in your browser and then offering to complete tasks for you. All these demos and prototypes still need to be turned into useful products, but at least in the labs they show that there is more than enough to keep the AI ​​hype going.
Will AI’s ‘killer app’ emerge?
For most people, the development of generative AI means constantly seeing prompts that prompt you to complete your text or edit your photos in ways you didn’t expect — less than unwanted, rarely useful tools that will change your life.
Next year will bring the first demonstrations of applications that can intervene more directly: pulling in all your digital data and learning from your actions to act as virtual memory banks or to control entire aspects of your life. But concerned about the technology’s insecurity, tech companies are quick to make these available for mass use — and most users are equally wary of trusting them.
Instead of truly killer applications for AI, this means we’re stuck in the “AI in everything” world that tech users are already used to: sometimes intrusive, sometimes helpful, and still not delivering the truly new experiences that prove. The era of AI has truly arrived.
Will Nvidia GPUs still rule the tech world?
The chipmaker’s high profits have targeted high-tech companies, many of which are now designing their own AI chips. But Nvidia is moving too fast for its rivals, and with a quarter or two to go through a major production transition, Blackwell’s production cycles should comfortably carry it through the year.
That doesn’t mean others don’t fit in. According to chipmaker Broadcom, the three biggest tech companies will use their in-house chip designs with 1mn chips each by 2027. That’s 10 times larger than Elon Musk’s Colossus system, which is thought to be the largest cluster. Currently used AI chips.
Even though its market share has started to erode, Nvidia’s software still represents a big part of its business, and it should be in for another important new product cycle by the end of the year.
Will Stock Market AI Continue to Grow?
Big Tech remains one of the main forces behind the growth of AI capital spending, which leaders believe will shape the future shape of their industry amid the AI ​​race. Also, as some companies begin to make big claims – if not proven – of implementing the technology in their own businesses, many others feel the need to continue spending heavily, even if they don’t yet know how to use AI productively.
Whether that’s enough to keep investors throwing their money into AI is another matter. That will depend on other factors, such as the stock market’s confidence in the Trump administration’s new deregulation and tax cut intentions and the Federal Reserve’s readiness to continue easing monetary policy.
It all points to a highly volatile year with some big corrections on the way. But with enough liquidity, Wall Street may succumb to the AI ​​hype for a while.
