The article critiques the telecom industry's reliance on data traffic growth as a primary economic driver, noting that revenue per gigabyte continues to decline despite rising usage. It argues that telcos must shift their business models away from volume-based metrics toward value-added services like AI token generation to remain economically viable.
The decoupling of volume and value suggests that infrastructure providers can no longer rely on a 'data tsunami' to justify capital expenditures. For the streaming ecosystem, this shift signals a move away from bit-pipe economics toward value-added services like AI token generation and edge AI services. As vendors like Qualcomm pivot toward integrated 6G servers and RAN acceleration, the industry must find new monetization units beyond the gigabyte to avoid further deflation. Watch for whether operators successfully transition to charging for intelligence and compute rather than just transport capacity in upcoming 6G specifications.
Qualcomm has been aggressively repositioning its infrastructure portfolio to address the revenue-per-gigabyte decline that telecom operators face. At MWC Barcelona 2025, Qualcomm unveiled its FSM100 and FSM200 RAN acceleration platforms designed for AI-native 6G networks, signaling a shift from pure connectivity silicon toward integrated compute and inference capabilities. The company's CEO Cristiano Amon has publicly framed this as a move to monetize intelligence rather than raw throughput, aligning with the broader industry argument that volume-based pricing is structurally broken. Ericsson, meanwhile, has pursued a parallel strategy through its Intelligent Automation Platform, which the company confirmed in mid-2025 had been deployed across multiple tier-one operators for autonomous network orchestration, targeting measurable reductions in operational cost per bit rather than top-line traffic growth.
The economic pressure driving this pivot is well documented by independent analysts. Dell'Oro Group reported that global telecom equipment spending declined for a third consecutive year in 2024, falling roughly 10% year over year, as operators deferred capex amid flat or declining ARPU. Tefficient's data reinforces the decoupling: the firm's 2025 analysis showed that mobile data traffic grew 22% globally while operator service revenue rose less than 2%, confirming that the elasticity between usage and income has effectively collapsed. Deutsche Telekom CEO Timotheus Höttges has been among the most vocal European executives calling for new monetization models, urging regulators at the 2025 Digital Summit to allow operators to charge hyperscalers for network usage, a proposal that remains contentious within EU policy circles but reflects the urgency operators feel as bit-pipe economics erode.
On the technical side, Qualcomm's X100 modem and its broader 6G roadmap represent an attempt to embed AI inference directly into the radio access network, creating new billable units beyond the gigabyte. Qualcomm's Hemanth Sampa and Durga Malladi outlined at the 2025 Snapdragon Summit how the X100 integrates on-device AI token generation with sub-6 GHz and mmWave connectivity, positioning the chipset as a platform for AI infrastructure inference optimization that operators could monetize per inference rather than per byte. Nvidia has entered this space from the GPU side, announcing its NVIDIA DSX AI factory platform integration with Ericsson and Nokia RAN software stacks in early 2025, creating a competitive dynamic where both chip vendors and traditional RAN suppliers race to define what the post-gigabyte revenue unit looks like. The outcome of this competition will likely shape whether operators can arrest the revenue-per-gigabyte decline before 6G deployments begin at scale.
Telecom data traffic is growing, yet revenue per gigabyte has hit historic lows. This decoupling of volume and value means infrastructure providers can no longer rely on data consumption to drive profits. Consequently, the industry is shifting toward monetizing AI intelligence and compute services rather than just raw transport capacity.
Global retail prices per gigabyte plummeted from $8.18 in 2019 to $2.59 in 2023. While mobile data traffic grew 22% globally, operator service revenue rose less than 2%, indicating that the elasticity between usage and income has collapsed.
According to Qualcomm VP Hemanth Sampath, a single AI assistant could consume 45 GB of data monthly, which could potentially limit 5G cells to only five users.
Deutsche Telekom CEO Timotheus Höttges urged regulators to allow operators to charge high-traffic platforms for network usage. However, he eventually abandoned these efforts after failing to secure payments from these companies.
Companies are pivoting toward integrated compute and AI inference. Qualcomm is developing RAN acceleration platforms for 6G, while Ericsson is deploying its Intelligent Automation Platform to reduce operational costs per bit.
Add StreamingMeme as a preferred source on Google to see more of our streaming news at the top of your Search results.
Add as preferred source