According to a report by JPMorgan released in September, tech giants Microsoft (MSFT), Meta (META), Google (GOOGL), and Amazon (AMZN) have collectively spent an astonishing $125 billion on A.I.-related capital expenditures (CapEx) and operational costs during the first eight months of 2024. This spending is expected to exceed $200 billion for the entire year, showcasing the intense focus these companies are placing on A.I. advancement.
Meanwhile, A.I. startups are witnessing an unprecedented influx of funding as investors rush to harness the lucrative potential of this technology. OpenAI is projected to conclude 2024 as the most well-capitalized A.I. company, recently valued at a staggering $157 billion. Its competitor, Anthropic, is also preparing for a fundraising round that could elevate its value to $40 billion.
With significant financial backing, leading A.I. firms face the challenge of demonstrating to both investors and the public that their substantial investments in this technology will yield substantial returns. As we look ahead to 2025, several trends are emerging in the A.I. landscape:
Agentic A.I. is poised to be the next major breakthrough
The term “agentic A.I.” refers to autonomous A.I. assistants capable of completing tasks without human intervention. The potential of these A.I. agents to enhance both workplace productivity and daily life has caught the attention of Silicon Valley, prompting companies like Salesforce to adopt agents as a central aspect of their product offerings.
Microsoft has also introduced various A.I. agents in recent months, unveiling a range of assistants tailored for its Microsoft 365 suite, including one capable of translating across nine languages.
OpenAI is joining the trend with an upcoming model designed to handle tasks like travel bookings and code writing. Sam Altman, OpenAI’s CEO, emphasized during a recent Reddit AMA that A.I. agents represent “the next giant breakthrough” in technology.
According to research firm MarketsandMarkets, the global market for A.I. agents is currently valued at over $5 billion and is projected to skyrocket to $47 billion by the end of the decade, largely driven by demand from enterprise clients.
Addressing A.I.’s data training crisis with test-time compute
A significant factor in A.I.’s success has been the vast amounts of data fed into its models. However, the supply of text, images, and videos on the internet is finite. To avoid stagnation, A.I. companies are exploring alternative training methods. One promising solution is “test-time compute,” wherein A.I. models enhance their performance by taking longer to reason through potential responses before answering. This concept was notably showcased by OpenAI’s o1 model.
During a November earnings call, Nvidia (NVDA) CEO Jensen Huang described OpenAI’s new model as “one of the most exciting developments” in scaling, noting that “the longer it thinks, the better and higher-quality answer it produces.” This sentiment is echoed by Microsoft CEO Satya Nadella, who recognized test-time compute as a new scaling law, while OpenAI co-founder Ilya Sutskever highlighted it as an evolution of A.I.’s pre-training phase.
Synthetic data emerges as a viable alternative
Another innovative approach to A.I.’s data challenges involves substituting traditional data with information generated by the technology itself. The synthetic data market is projected to reach $2.1 billion by 2028, marking a significant increase of over 450 percent since 2022, according to BCC Research.
Altman hinted at the promise of synthetic data last year, stating in an interview that “as long as you can get over the synthetic data event horizon, where the model is smart enough to produce good synthetic data, I think it should be all right.” Companies like OpenAI, Anthropic, Meta, Microsoft, and Google have reportedly begun incorporating synthetic data in various forms to train their models.
In October, A.I. startup Writer unveiled a new model trained entirely using A.I.-generated data, significantly reducing development costs to just $700,000, a stark contrast to the millions typically spent by other firms. For example, OpenAI’s GPT-4 model incurred over $100 million in training expenses.
Large world models are set to revolutionize 3D A.I. environments
Historically, A.I.’s visual outputs have predominantly remained two-dimensional. However, tech innovators are aiming to shift this paradigm in the coming years with the development of “large world models.” These models aspire to create interactive three-dimensional environments that could transform the realms of gaming, film, and simulation.
World Labs, a startup founded by Stanford A.I. pioneer Fei-Fei Li, raised $230 million earlier this year to pursue building large world models with “spatial intelligence,” a form of intelligence that understands and interacts with the real world. Li has previously illustrated this concept with examples, such as a cat knocking over a glass of milk, emphasizing the importance of anticipating consequences in such interactions.
In early December, Google DeepMind introduced its own large world model, named Genie 2, which simulates virtual environments designed for training and evaluating A.I. agents. This area is likely to become a key focus for the lab, particularly following the recent hiring of Tim Brooks, a former OpenAI researcher overseeing its video generator Sora. Google DeepMind CEO Demis Hassabis expressed excitement about collaborating with Brooks to realize the long-held dream of a world simulator.
A.I. search engines are set to transform online search
For years, Google has maintained its dominance in the search engine market. However, as artificial intelligence continues to evolve, a new wave of AI-powered search engines is emerging, poised to challenge Google’s supremacy.
Google itself is embracing this technological shift. In 2024, the company launched AI Overviews, which provide users with AI-generated summaries instead of traditional search results. CEO Sundar Pichai is optimistic about the feature’s potential, predicting it could attract over 1 billion monthly users and enhance overall search usage and satisfaction.
Nevertheless, Google faces fierce competition from other players in the search market, such as OpenAI and Microsoft, who are leveraging AI to enhance their services. Additionally, Meta is reportedly developing its own AI-powered search engine, while the startup Perplexity AI has emerged as a significant contender with a valuation of $9 billion and approximately 20 million queries handled daily, a considerable increase since early 2024.
This surge of AI-enhanced search engines indicates a transformative shift in the tech landscape, as companies strive to harness the power of artificial intelligence to redefine how we search for information online.
