Goldman Sachs: token demand growing 18x as prices collapse
Token growth alone is a misleading indicator of GPU demand — efficiency techniques can decouple the two, which reframes how the whole AI infrastructure cycle should be read. Investors tracking hyperscaler capex and operators planning inference capacity should watch pricing and compute-per-token, not just volume. The note also signals rising pressure on frontier-model economics as open-source alternatives absorb marginal demand.
Goldman Sachs research projects total AI token demand to grow 18x from December 2025 to September 2026, while frontier-model token demand grows only about 8-9x on the same index.
The widening gap means the marginal AI token increasingly comes from cheaper, smaller, open-source or routed models rather than compute-heavy frontier models.
As inference becomes commoditized, hyperscalers' capex math shifts: total GPU demand can grow far slower than token volume suggests.
The key variable is tokens consumed multiplied by compute required per token, reduced by quantization, distillation, speculative decoding, better caching and model routing.
AI is entering a volume economy: because token demand now has to outrun the price collapse underneath it.
Per Goldman Sachs research, total token demand is accelerating (by 18x from December-2025 to Sept-2026 ), while frontier-model token demand is growing much more slowly, at roughly 8-9x on the same index.
the bar for sustaining investment spending getting higher as inference becomes commoditized.
That means the marginal AI token is increasingly coming from cheaper, smaller, open-source, or routed models, not necessarily from the most compute-heavy frontier model. This changes the infrastructure math quite a bit.
That gap is so important for hyperscaler's capex-cycle, because frontier models remain an important source of hyperscaler compute demand, while increasingly competitive open-source models are helping push average token prices down.
So token growth alone is no longer enough to read the AI infrastructure cycle. The relevant variable is the interaction between token volume, token pricing, and the amount of compute required to serve each token.
What really matters is:
tokens consumed × compute required per token
If token volume grows 18x but more of those tokens move to models that need far fewer GPU cycles per token, total GPU demand can grow much slower than token demand suggests. Quantization, distillation, speculative decoding, better caching, smaller models, and model routing all push in that direction.
This report is categorized under AITOP's "Industry" desk, focusing on US frontier AI labs and Silicon Valley technology leaders. Within our cross-regional analytical framework, it directly maps to AI developer toolchains, coding agents, and engineering workflows and US-China foundation model benchmarks and commercialization comparison. Key entities and technical topics under continuous tracking include Goldman Sachs, token economics, open-source models, hyperscalers.
Source attribution: curated from rohanpaul_ai (@rohanpaul_ai). Cross-source corroboration: 3 independent sources have covered this development (including PolymarketMoney, Decoder), confirming industry-wide signal strength.
