One Size Fits None: How Platform Engineering Must Evolve • William Rizzo & Colin Griffin • GOTO 2026

Platform Engineering is Not One-Size-Fits-All

Platform engineering principles like accelerating developer velocity and reducing cognitive load are universal, but their practical application varies significantly across different industry verticals. Generic frameworks must be tailored to the unique problems, requirements, and regulatory landscapes of sectors like fintech, telco, and automotive.

A common pitfall is “target fixation” on concepts like the Internal Developer Platform (IDP). An IDP is just one possible capability or outcome of platform engineering, not the universal starting point or end goal for every organization. The core challenge for platform engineers is to learn the unique language of the teams they serve and build a translation layer that connects disparate groups, creating specific perspectives rather than imposing a single, generic one.

Aligning Investment with Business Outcomes and Compliance

A critical challenge for platform practitioners is the failure to articulate how their work ties directly to business outcomes. Investments in platform engineering must be linked to a strategic path with tactical actions that deliver measurable business value. Without this clarity, securing and justifying investment becomes difficult.

This dynamic plays out differently across industries:

• Fintech: These organizations, while often having the financial means, are extremely cautious with platform investments due to stringent regulatory and auditing requirements. A new platform capability, if it causes an audit to fail, can lead to a system being pulled from production, directly impacting revenue. For them, any platform modernization must not break compliance; it must elevate it.
• Industrial: These sectors may be more open to pilots and experimentation, as fewer systems are under the same level of strict compliance, allowing for a longer-term investment horizon.

The EU vs. North America Divide

There is a noticeable divergence in drivers between the European Union and North America:

• EU: Organizations are already deeply engaged with compliance due to regulations like the AI Act. The complexity of the EU, where each member state has national authority, necessitates robust technological solutions to manage regulatory outcomes.
• North America: While compliance is on the horizon, the current environment is more of a “Wild West.” The immediate focus is often on simply “doing AI” and throwing money at the problem, sometimes with a shallow understanding of the organizational and technical readiness required.

The AI and Infrastructure Reckoning

The massive investment in GPUs for AI workloads is forcing a fundamental rethinking of infrastructure. Many organizations are now grappling with how to orchestrate and maximize the value of these expensive assets.

This has led to several key shifts:

• Hardware as a First-Class Citizen: Platform engineering for AI brings hardware, networking, and storage directly into the “golden path.” Orchestrating bare metal is becoming a critical skill. Infrastructure can no longer be an afterthought; it must be designed specifically for AI. You cannot simply buy GPUs and expect your existing infrastructure to be “AI-ready.”
• The Return of Network Engineering: The need for high-bandwidth, low-latency links to service AI models efficiently brings network engineers back to the table as first-class citizens. Their expertise is essential for wiring the infrastructure to maximize the monetary expenditure on every GPU.
• From Securing to Orchestrating: Initially, the challenge was securing GPUs. Now, the primary problem has shifted to orchestrating them effectively to deliver models as a service seamlessly, manage slicing and sharing (e.g., NVIDIA MIG), and abstract the underlying hardware complexity from the end user.

A Wish List for the CNCF Platform Engineering Guidance

Looking ahead, both the CNCF Platform Engineering White Paper and the Maturity Model are due for an update. The key desire is for the community to provide more specific, actionable guidance.

Top Wish List Items:

1. Verticalization: Both the white paper and maturity model should incorporate industry-specific guidance. Addressing the unique challenges of fintech, automotive, telco, and other sectors would make the documents far more practical.
2. End-User Feedback: There is a strong call for end-user organizations to come to the table and share what they found missing or no longer applicable in the first versions of the maturity model. This direct feedback is crucial for evolving the models to reflect real-world needs and challenges.

• Tags:
• Platform Engineering
• AI Infrastructure
• Cloud Native
• Fintech
• CNCF
• GPU Orchestration
• Maturity Model