Skip to content

Reserved TensorNova capacity

Reserved Fleet capacity for buyers who need production posture, not opportunistic supply.

TensorNova Fleet is the reserved capacity rail for repeatable AI workloads. It sits above Marketplace partner supply with stronger day-to-day predictability, cleaner ownership language, and a more production-ready commercial path.

Reserved TensorNova-operated capacity Dedicated instances only Production-oriented commercial rail Escalates cleanly into Enterprise

TensorNova Fleet

Reserved instances for teams that need steadier availability than partner supply alone.

Fleet is for buyers who have moved beyond opportunistic sourcing but do not yet need a fully custom Enterprise deployment. The promise is straightforward: reserved capacity, stronger production predictability, and a cleaner escalation path when private controls become necessary.

See how Enterprise adds private deployment review →

Access model

Reserved dedicated instances

Best fit

Steady production workloads

Commercial mode

Public starting rates

Escalation

Enterprise when review deepens

01 · Fleet boundary

What Fleet is for

Fleet is for buyers who have moved past opportunistic sourcing and now need steadier reserved capacity, stronger operating discipline, and a cleaner production procurement story.

02 · Fleet boundary

Where Fleet stops

Fleet is still a public reserved-capacity rail. When the workload needs private connectivity, customer-directed controls, or contract-specific access scope, the request should move into Enterprise review.

TN-H100-GEN

NVIDIA H100

US East

VRAM

80 GB

Nodes

64

Network

800 Gbps

SLA tier

Business

  • • Dedicated instances
  • • Higher baseline capacity
  • • Reserved production path

Reserved capacity

Business

$3.39/hr from

Talk to Fleet Sales

TN-H200-SCL

NVIDIA H200

APAC Singapore

VRAM

141 GB

Nodes

32

Network

800 Gbps

SLA tier

Business

  • • High-memory Fleet
  • • Dedicated instances
  • • Reservation path

Reserved capacity

Business

$4.02/hr from

Talk to Fleet Sales

TN-B200-PRI

NVIDIA B200

APAC Singapore

VRAM

192 GB

Nodes

24

Network

800 Gbps

SLA tier

Business

  • • Blackwell generation
  • • Reserved instance capacity
  • • Enterprise priority available

Reserved capacity

Business

$6.32/hr from

Talk to Fleet Sales

01 · Fleet operating logic

Reserved capacity, not opportunistic supply

Fleet exists for workloads that have outgrown variable partner supply and now need steadier reserved instance allocation with clearer day-to-day expectations.

02 · Fleet operating logic

Production posture with clearer ownership

TensorNova Fleet is reserved capacity for repeatable workloads that need stronger production expectations, clearer ownership language, and a cleaner procurement story.

03 · Fleet operating logic

Enterprise remains the private path

Enterprise contracts add deeper review around private networking, key control, and direct infrastructure access when reserved public capacity is no longer enough.

Enterprise bridge

Fleet is the last public rail before the request becomes private infrastructure design.

Fleet is still public reserved capacity. Enterprise is where buyers scope dedicated machines or clusters, private connectivity review, stronger key-control posture, and direct infrastructure access where contracted.