Why The Next Phase of AI Adoption Hinges On AI-Enablers
We’re on the cusp of a new era of AI enablement. Google claims new AI training tech is now 13x faster and 10x more power efficient, while DeepMind‘s Jest is optimizing training data for increasingly impressive gains.
As AI models become more sophisticated by orders of magnitude each month and the costs for powerful LLMs continue to come down, we can expect to see even more adoption of AI across industries.
While the future of AI is still uncharted waters, we are starting to see clear parallels with the previous internet revolution and rise of software service apps. Here, it’s important to remember that it was infrastructure pioneers who paved the way for consumer-driven services and transformative industry solutions that later took shape.
In the early era of the internet, the dominant telcos of the time had to invest heavily in building the physical backbone required for widespread internet adoption, much like the hyperscalers of today. Without these investments, the internet would have struggled to get off the ground.
The promises of the internet and creation of new markets led to rapid technological innovation, albeit without established standards or protocols in place. This led to the rise of software infrastructure companies providing technologies such as database systems, networking infrastructure, security solutions and enterprise-grade storage.
Over time, these foundations paved the way for software infrastructure companies Cisco, Sun Microsystems and Oracle to become pivotal internet enablers as system design and protocols began to standardize.
We can see a highly similar pattern shaping up today when we examine the progress of AI adoption. Although we have the nuts and bolts of functional AI tools — often referred to as “point solutions” in venture circles — achieving widespread and meaningful adoption of AI will largely hinge on emergence of AI enablers: Foundational software infrastructure tools and components designed to support, scale and streamline AI-native applications and workflows.
At this pivotal moment for AI, there are three themes of AI-enablers to watch in 2025 that have the ability to support enterprise developers with the foundation and framework to build durable, reliable AI-native applications.
Durable cloud workflows: Enable reliable, reproducible results
As AI moves from the lab into the hands of enterprise-scale users and major public institutions, “good enough” won’t do. As with any service, AI needs to be reliable, offer reproducible results or idempotency, and minimize the risk of faults that undermine the viability of the technology during this pivotal moment of adoption.
From a workflow perspective, this challenge can be addressed by using software patterns that extend strong transactional guarantees ensuring strong consistency while developers focus on writing business logic instead of boilerplate code for failure scenarios.
Resource management tools key to efficient deployment
Although the overall ratio of cost to performance for AI models is coming down each year, companies still need to pay attention to how infrastructures are being built to maximize efficiencies.
The resource management tools we call AI enablers make it easier to use databases, streaming, storage and caching. Disparate tools cobbled together under poorly structured frameworks can not only drain both computing and financial resources, but also lead to a misallocation of engineering effort and focus. Current cloud deployments are optimized for non-AI applications, and as we transition to an AI-native world, these resource management tools would be helpful in building more efficient and capable workflows.
Leverage DevOps for agile AI solutions
Optimized DevOps tooling is essential for accelerating development cycles, unlocking developer productivity and enhancing software quality.
While we can see the incredible potential of rapid prototyping tools such as CoPilot, Cursor and Loveable, new and novel tools for automated testing and remote build execution are making the task of creating quality code much simpler. Taken together, these tools provide developers with 10x capability, turning ideas into production-ready systems faster than ever.
AI is poised to revolutionize industries, redefine how we interact with technology and, ultimately, power the next great leap in global connectivity. By ensuring sufficient investments in AI enablers and removing architectural bottlenecks, ambitious enterprises can leverage the full force of AI productivity at scale.
Sumedh Nadendla is a venture capitalist and investment lead at Pacific Alliance Ventures. He is also a mentor at USC Incubator and received his master’s degree from Columbia University.
Illustration: Dom Guzman
