The AI Horizon: Emerging Trends and Opportunities
As we step into 2025, the artificial intelligence landscape continues to evolve with unprecedented speed, marked by significant breakthroughs and substantial venture capital activity. The past year and beginning of 2025, could rightfully be termed the ‘AI era’ due to the intense activity and innovations within the AI sector.
VC Funding Landscape
According to CB Insights’ data, despite a general decline in the number of venture deals, which hit an 8-year low with approx. 27th. transactions, the total 2024 VC funding reached $274.6bn (+12% YoY increase), with a significant portion directed towards AI-focused ventures.
AI ventures notably dominated the VC landscape, capturing c.37% of all funding; a substantial raise from 21% in 2023 and a mere 5% in 2015;
- AI firms broadly have further increased their share of total deal volume — 17% of all deals involved AI companies, nearly triple the 6% seen in 2015, and the total number of AI deals has consistently exceeded 4’000 p.a. for the 4 four years.
- The 2024 five largest investment rounds predominantly benefited AI model and infrastructure companies, with Databricks leading at $10bn in Series J funding, followed by a $6.6bn investment in OpenAI, two $6bn investments in xAI, and a $4bn investment in Anthropic.
- The last quarter of 2024 witnessed a significant resurgence in overall funding levels, with investments peaking at $86.2bn (the highest in two years volume; 53% increase from the previous quarter), with over 60% of this quarter’s funding attributed to AI mega-rounds (valued at over $100 million each).
Structural shifts in AI Space
The discourse around the efficacy of open versus closed AI models gained new dimensions with the emergence of DeepSeek, a company whose cost-effective AI models have challenged the traditional, resource-heavy approaches of giants like OpenAI and lead to Sam Altman acknowledging the need for a strategic pivot towards more open-source models in light of DeepSeek’s innovations [Link].
This shift is significant as it suggests that the future of AI development may increasingly rely on algorithmic efficiency and less on extensive computational power.
That said, the limitation in computing hardware (particularly GPUs) that historically been constraining exponential growth of largest resources-heavy AI models, are now largely addressed. The next critical bottleneck — energy resources to power expansive data centres for AI operations — is also being addressed by heavy major technology companies’ investments in energy solutions and computing methods:
- Google has partnered with Intersect Power and TPG Rise Climate in a $20 billion initiative to build data centers adjacent to solar and wind farms, aiming to align data center operations with renewable energy sources. [Link / Link]
- Meta Platforms plans to construct a $10 billion AI data center in Louisiana, collaborating with Entergy to deploy new natural gas power plants, highlighting the intersection of AI infrastructure and energy development. [Link]
- Google’s recent unveiling of the Willow quantum computing chip marks a significant advancement in computational capabilities. This innovation is expected to accelerate research and problem-solving across various sectors, potentially leading to breakthroughs not only in AI space, but also in materials science, cryptography, and complex system modelling. [Link]
Is AI only for BigTechs?
The trajectory of AI adoption across industries suggests otherwise. Even before DeepSeek’s clear illustration of AI accessibility, the adoption of AI technology was active with AI Agents becoming increasingly prevalent, catalyzing change and enhancing efficiency across various sectors. From healthcare, where AI aids in diagnostics and patient management, to enterprise workflows and industrial operations where automation is streamlining processes and enhancing efficiency — AI powered ventures are on the raise.
Supported by activity in the AI-related segments, despite turmoil at later stages, early-stage VC activity showing strong positive trend with increasing deal sizes ($2.1m in 2024, up from $1.0m in 2015) and valuations ($25m 2024 median, up from $10m in 2015).
What’s Next? Centralised Platforms and AI Infrastructure
As we project into the future, the role of AI in sculpting our technological and economic landscapes continues to be profoundly significant. There’s a burgeoning activity across a wide range of AI segments, with one of the most promising opportunities being the development of a centralized AI platform, akin to Apple’s App Store. App Store revolutionized mobile applications market and, by its recent earnings [Link] highlighted the lucrative potential of monetizing a platform that facilitates access to a vast ecosystem of apps and services.
Such a platform would unify AI tools and services, offering a scalable and accessible marketplace for businesses and developers to deploy and monetize AI solutions. This would not only catalyze innovation but also democratize access to AI technologies, potentially mirroring the transformative impact seen with the proliferation of mobile apps.
Contributors to this burgeoning ecosystem include unicorn companies like:
- Scale AI | $1bn Series F led by Nvidia and Accel in 2024
Scale AI offers data annotation services that are crucial for training AI models. Their platform could serve as a vital component of a centralized AI platform, providing the annotated data needed by companies to train models efficiently and effectively. - DataRobot | over $1bn raised to date
DataRobot provides a platform for automating the end-to-end process of building, deploying, and maintaining AI and machine learning models at scale. They offer a robust platform that could integrate with a centralized AI platform, providing critical infrastructure for model development and management. - H2O.ai | over $250m raised to date
Open-source platform for automated machine learning, H2O.ai simplifies the deployment of AI models and could be a crucial player in an AI platform ecosystem. Their tools make it easier for developers to build and deploy machine learning models. - Snorkel AI | over $150m raised to date
Snorkel AI focuses on programmatic data labeling, allowing developers to create and manage training data using AI-powered systems. This approach significantly speeds up the AI training process and could be a valuable asset for a centralized AI platform, enhancing the efficiency of model training.
.. as well as early and mid-stage challengers like:
- Abacus.AI | Series C
Abacus provides a platform that enables the autonomous creation of AI models, focusing on deep learning. Abacus.AI’s services could support the backend of a centralized AI platform by streamlining the deployment and scaling of machine learning models. - V7 | Series A
V7 specializes in creating tools for training vision AI, offering annotation and dataset management services tailored for visual data. Their technology could enhance a centralized AI platform’s capabilities in fields requiring image recognition, such as security, autonomous driving, and medical imaging. - Landing AI | Series A+
Landing AI focuses on bringing AI to manufacturing industries. The company’s tools could help automate quality control processes on a centralized platform, making it easier for manufacturers to adopt and benefit from AI technologies. - Akido Labs | Series A
Akido Labs works on deploying AI tools for public health, working with hospitals and government bodies. Their technology could underpin health-related applications on an AI platform, particularly those aiming to improve public health outcomes through data analysis.
AI Infrastructure
The potential for AI extends beyond just software and platforms and encompasses the entire supporting infrastructure. As facilitated by massive investments from major players like Google or Meta, infrastructure required for AI development is also on the raise, with segments well-positioned for outperforming growth in 2025 including:
- ML Enablers — data platforms, testing frameworks, data collection, labelling tools, and other essential components for streamlining AI deployment and management, and facilitating more robust and scalable AI solutions across various sectors;
- Energy production and distribution — modular reactors, nuclear and geothermal power, grid storage, energy optimization software, etc.
- Support infrastructure for AI data centres — immersion and liquid cooling, cloud security, processors optimization software, etc.
Notable companies to watch in these segments would include:
- Crusoe | Series C
Operator of mobile modular data centers intended to eliminate routine flaring of natural gas and reduce the cost of cloud computing. - Fiddler AI | Series B
Fiddler focuses on explainable AI, offering tools that help companies understand, analyze, and improve their AI models. As AI regulations evolve, Fiddler’s solutions could become essential for any centralized AI platform aiming to ensure transparency and compliance. - Coolgradient | Seed
Dara center energy optimization software platform that simplifies complex facility operations by providing detailed visibility into the asset health of data center and increases reliability, resilience, and compliance with sustainability regulations.
As we move forward into 2025 and beyond, AI isn’t just about making existing apps better — it’s a game-changer that’s reshaping whole industries and the very backbone they rely on. The combination of cutting-edge AI solutions and the solid infrastructure backing them up is setting the stage for a whole new era in tech development, where AI is poised to keep its spot as a key driver of global economic growth.
