Akamai's Q1 2026 AI Infrastructure and GPU Deployment Strategy

Akamai’s plans for AI infrastructure growth and GPU deployment strategies in Q1 2026

1) Core AI-infrastructure thesis: scale AI inference on a distributed “AI grid”

Akamai’s Q1 2026 message is that AI infrastructure growth is not meant to be “isolated AI factories,” but rather a unified, distributed grid for AI inference across its platform. ¹ Akamai describes this as integrating NVIDIA AI infrastructure into Akamai’s distributed platform and using intelligent workload orchestration so AI inference can run close to end users at the edge, reducing latency and cost. ¹

Akamai operationalizes this with its Inference Cloud strategy:

• Functions as a service runs serverlessly across all 4,300 locations. ²
• Managed container service is running in well over 100 cities and can run in all 700 cities (active in well over 100 today). ²
• IaaS capabilities exist in several dozen cities, with a couple dozen of those equipped with new 6000-series GPUs. ²
• The orchestration layer is designed to place workloads on the most computationally efficient resource as close as possible to the user. ²

This distributed grid approach is explicitly framed as being aligned with NVIDIA’s AI grid vision. ²

---

2) GPU deployment strategy in Q1 2026: “thousands” of RTX Pro 6000 GPUs and scaling capacity via contracted demand

In Q1 2026 communications, Akamai points to a major step at GTC in March, where it:

• unveiled the industry’s first global-scale implementation of NVIDIA’s AI grid, and
• announced the rollout of thousands of NVIDIA RTX Pro 6000 GPUs. ¹

The company then reinforces that its distributed platform is intended to support rapid scaling of AI inference workloads by placing compute where it provides the best benefit and by orchestrating whether workloads run on CPUs or GPUs. ³

Where Akamai is in deployment footprint (as of Q1 2026)

Akamai gives a footprint view for Q1:

• Inference Cloud spans 4,300 locations via serverless FaaS. ²
• Managed containers run broadly (in well over 100 cities and can run in all 700 cities). ²
• IaaS is in several dozen cities, and a couple dozen have 6000-series GPUs. ²

This indicates that GPU deployment is currently selectively deployed by city (IaaS GPU-enabled nodes in “a couple dozen” cities) rather than uniform GPU availability in every city. ²

---

3) Capacity growth plan: build out physical infrastructure + expand cloud sales/support, and reserve/serve GPU demand through a mix of on-demand and committed capacity

Management says Akamai expects to continue building out both physical infrastructure and cloud sales and support teams to satisfy strong and growing demand for cloud infrastructure services. ⁴

On customer capacity mechanics, Akamai states it supports both:

• on-demand, per-token or per-VM-hour access, and
• large tranche deals. ⁵

Critically, Akamai does not treat this as an either/or choice; it says that as it needs more GPUs it would purchase more. ⁵ The company also emphasizes that inference workloads typically require a mix of CPU and GPU, and application needs drive the CPU-vs-GPU allocation. ⁵

---

4) Q1 2026 “AI infrastructure growth” catalyst: a landmark AI-cloud commitment and follow-on pipeline growth

Akamai’s Q1 2026 plan for AI infrastructure growth includes a major revenue/commitment milestone:

• It announced a landmark seven-year $1.8 billion commitment for cloud infrastructure services by a leading frontier model company. ¹
• Management calls it the largest customer deal in Akamai history. ¹
• It frames this deal as part of “another major milestone for our cloud computing strategy,” alongside a previously announced $200 million CIS deal in February. ¹

Akamai also quantifies early Q1 momentum in CIS (Cloud Infrastructure Services):

• CIS revenue in Q1 started strong at $95 million, up 40% year over year (reported) and 39% in constant currency. ⁶

And Akamai highlights that the pipeline for AI-specific use cases is building rapidly. ⁶

---

5) GPU deployment and CapEx implications for Q1 2026: planned investment ahead of revenue + allowance for additional GPU orders later in 2026

A key part of Akamai’s infrastructure plan is spending to fund this growth trajectory:

• Management says it will be investing slightly ahead of revenue to capitalize on momentum and support accelerated growth expectations (with this reflected in updated capital expenditure and operating margin outlook). ⁶

For CapEx specifically (and therefore GPU deployment funding), management guides for full-year 2026 CapEx as a function of revenue, and also provides flexibility:

• It anticipates full-year capital of approximately 40% to 42% of total revenue, including the $700 million impact from the $1.8B contract mentioned earlier. ⁷
• Management says demand for CIS including GPU deployments is exceptional, and that its current pipeline for GPUs significantly exceeds existing and projected inventory. ⁷
• As a result, management states it may place additional GPU orders in the second half of the year to meet demand, and notes this is not factored into the current annual CapEx guide (meaning guidance would be updated on a subsequent call if another GPU order is placed before year end). ⁷
• It also states “all the CapEx we need to satisfy the $1.8 billion deal is in the guidance,” while separately acknowledging “potential additional GPU purchases tied to the broader pipeline.” ⁸

Demand sizing language underscores the GPU deployment scaling ambition:

• Management says opportunities range from “a couple hundred GPUs to a thousand or more per customer.” ⁸

Finally, management gives evidence of how their GPU-heavy capacity deals affect cost structure and margins:

• For larger dedicated-capacity deals, the biggest cost driver is depreciation over the period; cash gross margin costs are said to scale well, with expectations that cash gross margin and EBITDA margin expand over time. ⁸
• Operating margin may vary depending on mix; for very large deals, management notes margins may be below the 30% operating margin, while GPU-as-a-service/rental is described as tending to be much higher than company average. ⁸

---

Bottom line (directly answering the question)

In Q1 2026, Akamai’s plan for AI infrastructure growth is to expand a distributed AI inference “grid” across its Inference Cloud footprint (4,300 locations for serverless FaaS, managed containers across 700 cities capability, and GPU-equipped IaaS in “a couple dozen” cities). ² Its GPU deployment strategy includes rolling out thousands of NVIDIA RTX Pro 6000 GPUs initiated at GTC and scaling capacity for strong demand through a combination of on-demand access and large tranche/committed deals, with explicit willingness to purchase more GPUs as needed because inference requires a CPU/GPU mix depending on workload. ¹⁵

Akamai’s Q1 plan is also backed by a major seven-year $1.8 billion cloud infrastructure commitment tied to frontier model workloads, strong early CIS growth ($95M up 40% YoY), and management’s statement that its GPU pipeline exceeds available inventory, which could drive additional GPU orders in H2 2026 beyond what is currently in annual CapEx guidance. ¹⁶⁷