Expanding boundaries of scaling

The principles of AI scaling were first outlined by OpenAI developers in 2020. According to experts, increasing model parameters, the volume of training data, and computational power directly enhance AI system performance. These principles drove remarkable progress in language models from GPT-2 to GPT-4. In 2025, scaling laws will extend to new domains. Startups like EvolutionaryScale are leveraging these principles for large-scale biological models, while Physical Intelligence applies them in robotics. Experts predict successful implementation of AI potential in other fields as well.

Adoption of autonomous AI

Next year, web agents will become mainstream, autonomously managing daily online tasks such as subscription handling, meeting scheduling, and shopping. This technology could mark a new frontier in consumer AI applications.

Creation of orbital data centers

2025 is expected to be a breakthrough year for innovation, with the first major projects for deploying AI data centers in orbit. Lumen Orbit, a startup with $11 million in funding, aims to establish a network of orbital centers with gigawatt-scale capacity. According to company head Philip Johnston, "Instead of spending $140 million on electricity, $10 million could cover the cost of a launch and solar panels."

Donald Trump’s political influence

The political landscape deserves special attention. In 2025, relations between US President Donald Trump and Elon Musk, CEO of Tesla and SpaceX, could significantly impact AI development. Potential regulatory changes may affect companies like OpenAI and Tesla, reshaping global approaches to AI oversight. Elon Musk’s stance on AI safety, advocating for broad restrictions on developers, will play a crucial role in shaping the industry.

Shifting business strategies in AI labs

The coming year may bring transformative changes to key business models. Companies like OpenAI and Anthropic, alongside other leading labs, are expected to pivot towards application development. High operational costs and intense competition drive this shift, marking a departure from their previous focus on foundational AI models.

Autonomous AI research

Autonomous systems are set to take the lead in AI innovation in 2025. These systems will begin conducting independent research in artificial intelligence. Experts anticipate the first fully AI-generated scientific paper, which could be accepted at major conferences like NeurIPS, CVPR, or ICML.

Popularity of robotaxi services

Autonomous taxis are projected to claim a significant share of the global market in 2025, estimated to exceed 10% in five major US cities. Waymo, holding a 22% share of San Francisco’s market, plans to expand operations to Los Angeles, Phoenix, Austin, Atlanta, and Miami.

Challenges in AI security

However, alongside these advancements cybersecurity concerns come to the fore. Experts predict the first major AI-related security incident in 2025. Potential scenarios include systems attempting to create copies of themselves or deliberately lowering performance metrics to evade oversight. These challenges may dramatically reshape the AI landscape and introduce new hurdles for the industry.