Welcome to a holiday week AI & Finance, in which we’ll review some of the biggest artificial intelligence stories of the past year. If we were trying to put together a top 10 list of specific stories, we’d have a very hard time picking which ones to report. As with any rapidly expanding and proliferating technology, a lot was happening. 
We feel our readers are better served by rounding up these individual stories into about 10 dominant trends that really set the course for the news over the past 12 months. Rather than focus specifically on financial services AI, we’re picking the overall artificial intelligence trends.
From the editors’ birds’-eye perspective, 2024 was a year in which AI really left science fiction for good and entered our everyday lives and lexicon, in the workplace, in our homes, on our devices and with our friends and families. Perhaps there’s an uncontacted tribe somewhere where peoples’ lives are completely untouched by AI and no one is aware of the technology, but across most of the developed world, it’s here, and everyone, even if they’re not embracing it yet, seems to be using it.
For me, your AI & Finance editor, it’s been one of the best 12 months of my life. At the endnen of 2024, I find myself happier, healthier and wealthier than I have been in nearly a decade, and I hope that all of our readers can say the same.
If not, let’s all try together to make 2025 a year of joy and positivity.
Let’s get to the first five of our top 10 AI stories of 2024.\
1. A Couple of Huge Funding Rounds
In the last half of the year, two huge AI-related venture capital announcements punctuated a very busy year in fundraising (and there will be much more on that to come, we promise). While many technology companies would make a splash with funding rounds in the tens or hundreds of millions of dollars, the biggest and most successful AI-centric companies are raising capital in the billions.
To begin, OpenAI announced a $6.6 billion funding round at a $156 billion valuation at the beginning of October. The developer of ChatGPT and several other widely used models will use the round to fund research, increase its computing capacity and introduce new AI tools. Joining the round were SoftBank, Thrive Capital, Fidelity, Microsoft and Nvidia. As part of the funding round, OpenAI announced intentions to transition to a for-profit entity.
Keep in mind that just last year, the company was valued at $29 billion, and in 2021, it was valued at $14 billion. Quite the expansion.
Not to be outdone, literally speaking in this case, Databricks raised a huge $10 billion Series J round at a $62 billion valuation. The round was led by Thrive and Andreeson Horowitz and joined by the likes of WCM Investment Management, DST Global, GIC and Insight Partners. The funding will be used for acquisitions, new products and expansion of its international go-to-market capabilities.
2. New Models Proliferate, but a Big One Is Delayed
OpenAI also made the news as it released a slate of new models, including its video generation model, Sora, which launched in December. Earlier in the year, OpenAI made headlines with the launch of o1, a model designed to achieve higher-level reasoning by “spending more time thinking.” o1 is intended to help users solve more difficult problems than OpenAI’s previous models. The company was far from finished—in fact, by launching GPT-4o, it moved in the opposite direction of o1 by accelerating the speed at which its flagship model can reason.
There’s still more from OpenAI, but that will have to come later. It’s worth mentioning here that perhaps its most anticipated launch, that of GPT-5, has been delayed at least until next year.
OpenAI, of course, was not alone. Google rebranded its Bard large language model to Gemini, released Gemini Advanced and launched the related Gemma open models. Amazon (Nova), Meta (Llama 3.1), IBM (Granite 3.0) and Anthropic (Claude 3.0 and Claude 3.5 Sonnet) also got in on the fun—and those only cover a portion of the larger launch announcements over the past 12 months.
3. AI Comes to the PC and Our Devices
The processing power to run these major AI models can only reside in data centers packed with GPUs or AI-oriented computer chips and using a tremendous amount of electricity. However, smaller AI applications do not require large banks of hardware to run—and, in fact, they tend to be more efficient if they’re run as close to the end-user as possible. Thus, AI has made its way to the so-called edge, and is now moving onto our personal computers and devices even as it also proliferates across the cloud.
Microsoft probably had the largest announcement in the space with the introduction of its Copilot+ PC, Copilot being the name of Microsoft’s large language model-powered chatbot. Billed as the “fastest, most intelligent Windows PC ever built,” with new system architecture combining the CPU, GPU and an AI-focused neural processing unit (GPU) that can run many AI applications locally (while others will continue to run in Microsoft’s Azure Cloud). Still to be worked out is the new computers’ “recall” function, which is billed as being capable of resurfacing any information viewed on the PC in the past—delays thus far have been due to concerns over information privacy and security.
In a similar announcement, Apple finally made public its entry in to the AI computing universe with the launch of Apple Intelligence in October. With an update to its operating systems, Apple Intelligence offers users on-board generative AI writing tools, image generation, notification summaries and a quick response tool.
4. AI Expands Geographically and looks for New Energy Sources
Most of the big work being done with artificial intelligence is not on our individual PCs and devices, however, and 2024 saw a major expansion of AI data centers, with much more to come over the next few years. In the four years leading into 2024, data center capacity doubled. It will not take another four years for data center capacity to double again.
One issue with the growth of AI and data centers is where to put them. Even our most energy-rich states in the U.S. have strained electrical grids. Data centers ideally need to be near where work is being done and energy is being produced—which means the geographic focal point of AI is likely to move well beyond California and Silicon Valley. Much of the AI news we’ve reported in 2024 bears that out, with many AI companies now locating in states like Wyoming, Oregon, Michigan, Wisconsin and Montana versus the heartland of the U.S. computer industry.
More data centers means more energy use—according to the International Energy Agency, data centers used 460 terawatt-hours of electricity in 2022, and will use more than 1,000 terawatt-hours in 2026. Thus, we’ve seen a keen interest not only in the next generation of renewable resources like wind and solar power, but also in restarting nuclear reactors and, after many years of decline, rejuvenating the First World’s nuclear energy programs.
This year, the AI-related energy headline that jumped out to us the most was the announcement that one of the decommissioned reactors at Pennsylvania’s Three Mile Island generation plant will be restarted, with Microsoft promising to buy the equivalent of its generating capacity for the first several years to power its own data center plans.
5. AI Threats Loom but Don’t Really Break Through
Let’s be honest with ourselves here: There are clear downsides and risks to the proliferation of artificial intelligence. Some of these risks follow the usual anti-technology mythology, with neo-Luddites fretting about lost jobs and displaced laborers. These are valid complaints. AI is going to disrupt the global economy across every sector, and people will be displaced, and, eventually, the meaning and value of work and labor will likely need to be reconsidered. But we’re a long way from that.
More relevant to us in 2024 is the use of AI in fraud and scam schemes. AI is being used to create ever-more sophisticated attacks on individuals and businesses via phishing, cyberattacks and social engineering. Generative AI, in particular, is accelerating these attacks because individual fraudsters will potentially no longer need to spend as much time studying their marks—it’s the kind of detailed, data-intensive task at which generative AI excels.
AI misinformation also proliferated this year. However, one major concern for 2024, an artificial intelligence-oriented assault on the U.S. presidential election, never really came to fruition. While we should all keep a careful eye on the use of AI to separate us from our hard-earned money, it’s not quite ready to create global political disruption. Yet.
