Businesses are investing heavily to achieve AI-driven growth, seeking to quickly transition from experimental pilots to fully scaled projects that deliver value.
While some companies are seeing success. others are stalling as they navigate the evolving dynamics that underpin successful AI transformation.
According to Foundry’s AI Priorities Study 65% of IT decision makers report having a dedicated technology budget for AI, up from 49% the previous year. Another 97% say they are investing in or planning to invest in AI tools, signalling a shift from research to implementation.
But a third cite determining ROI as an implementation challenge, which could be one reason some businesses are not reporting success with their AI efforts so far.
In a new CIO webcast, Alexander Troshin, product marketing manager EMEA at AMD, reflects on the lessons IT leaders can take from the first wave of AI transformation as they plan for what’s next.
Enterprise AI adoption is moving to a practical phase, as organisations shift from experimentation to operationalisation. Hyperscalers and enterprises alike are investing in new infrastructure including AI supercomputers, data centres and edge devices, while IT leaders are focusing on making AI reliable, scalable and cost-effective.
Concurrently, the overall market is maturing. Enterprises seeing the most progress are treating AI as a coordinated programme that aligns data, infrastructure, security and operating models, rather than simply bolting accelerators onto their existing stacks.
Troshin explains: “AI is forcing IT leaders to revisit the fundamentals — where AI workloads should run, how quickly they need to move through the organisation and how efficiently mixed workloads can be managed. Leading organisations are building platforms and systems that are designed for repeatability and based on the right combination of data centre, cloud and device capabilities working together.”
As enterprise AI strategies mature, the debate is shifting away from cloud versus on premises and toward ensuring that workloads are deployed in the optimal environment.
Successful enterprises start with business needs and use cases before deciding where workloads should live based on factors such as performance, latency, data sovereignty and cost.
“AI infrastructure decisions today are about balance and flexibility,” says Troshin.
“Enterprises need platforms that scale across the full AI life cycle while efficiently orchestrating workloads across processors, accelerators and memory. Getting that foundation right is what allows organisations to scale AI reliably.”
The role of open ecosystems
As IT leaders build these foundations, open ecosystems are becoming vital. Models, frameworks and optimisation techniques are changing constantly, and enterprises need the flexibility to adopt new tools without rebuilding their entire foundation. Open ecosystems also enable freedom of vendor choice for enterprises while integrating compute, storage and networking in ways that suit their timelines and business needs.
Troshin comments: “AI stacks evolve incredibly quickly, so openness is critical. Enterprises need software portability, composable infrastructure and access to a broad partner ecosystem so they can adopt new models without disrupting their environment. An open ecosystem helps protect investments while keeping organisations agile as AI technologies evolve.”
As Troshin makes clear, success in AI is built not bought. As the next phase of AI gets into gear, the lessons of the past point the way forward. Businesses that invest in the right technical foundations and make use of open ecosystems and strategic partnerships will be well positioned to succeed.
Learn more about how to navigate AI transformation.