Technology / / 5 min read

AMD Details AT&T's 80% Agentic AI Token Cost Reduction as NASA Shows How LISA Gravitational Waves Can Detect Exoplanets

Following the 2026 Agentic AI Exchange in New York, AMD Chief Commercial Officer Philip Guido outlined a 'value per token' enterprise framework highlighting AT&T's 1-trillion-token Azure Instinct deployment and 80% gateway cost savings, while NASA Goddard scientists demonstrated how LISA gravitational-wave sidebands can reveal exoplanets.

By 全球伯乐 News

Frontier Compute & Space Science Desk

Share
On this page

Between November 11 and November 12, 2026, enterprise computing and space astrophysics researchers published parallel breakthroughs in signal routing and precision detection: following the 2026 Agentic AI Exchange in New York City, AMD Executive Vice President and Chief Commercial Officer Philip Guido published a strategic report titled 'AI Is the New Electricity. And the Meter Is Running.', proposing 'value per token' as the core metric for enterprise agentic AI infrastructure1. Simultaneously, astrophysicists at NASA's Goddard Space Flight Center (GSFC) released new simulations for the Laser Interferometer Space Antenna (LISA) mission, demonstrating how Doppler sidebands in gravitational waves from double white dwarf binaries can uncover sub-Jupiter and Jupiter-mass exoplanets thousands of light-years across the Milky Way2.

AI-generated illustration: Tabletop still life of three triangular brass optical mirror mounts on frosted glass, a matte clay exoplanet sphere suspended in a thin brass ring, and a dark-blue miniature GPU heatsink module on slate

AI-generated illustration: Editorial still-life concept depicting enterprise agentic AI token routing and space-based gravitational-wave astrophysics modeling, not an actual laboratory or spacecraft photograph.

The Meter Is Running: AT&T's 1-Trillion-Token AMD Instinct Deployment and 80% Gateway Savings

Tracing four rapid leaps in artificial intelligence—from conversational AI in 2022 to multimodal models in 2023, reasoning systems in 2024, and autonomous agentic workflows across 2025 and 2026—Guido noted that every enterprise productivity multiplier now acts as a direct compute and token multiplier1. As teams of AI agents reason over extended context windows and trigger actions across thousands of corporate workflows, chief information officers face escalating inference bills unless they match each task to appropriately sized open-weight or domain-specific models across cloud, on-premises data centers, edge servers, and AI PCs rather than routing every query to a massive frontier model1.

Highlighting concrete production economics, AMD disclosed that telecommunications provider AT&T has processed more than 1 trillion tokens on AMD Instinct GPUs running in Microsoft Azure to train open telco-specific models1. By pairing those specialized models with a cache-aware AI gateway that routes each request to the right model tier, AT&T reduced its overall AI operational costs by as much as 80%1. Based on these deployments, Guido outlined a three-step enterprise framework—identifying high-impact workflows and human-in-the-loop boundaries, proving total cost of ownership (TCO) via real-world proofs of concept, and scaling validated architectures across open platforms1.

Gravitational-Wave Sidebands: How ESA and NASA's LISA Observatory Can Spot Sub-Jupiter Worlds at 500 Light-Years

While enterprise architects optimize token streams across silicon accelerators, space scientists are learning to decode planetary signatures embedded in ripples of space-time. According to the simulation study led by NASA Goddard scientist Scott Noble, the space-based LISA observatory—led by the European Space Agency (ESA) with major NASA hardware and analysis contributions—will monitor thousands of tightly orbiting double white dwarf binary stars across our galaxy2. On its own, a close white dwarf binary emits gravitational waves at a nearly continuous, monochromatic carrier frequency; however, if an exoplanet orbits the binary pair, its gravitational tug induces a subtle reflex wobble in the stars' center of mass2.

Just as optical radial-velocity surveys detect planets via Doppler shifts in starlight, the binary's reflex motion imprints a Doppler frequency modulation onto the gravitational-wave carrier, splitting part of the signal energy into upper and lower frequency 'sidebands' analogous to AM/FM radio transmissions2. Evaluating a simulated test system 500 light-years from Earth, NASA researchers found that LISA could detect a sub-Jupiter-mass exoplanet under ideal conditions and identify Jupiter-mass worlds orbiting white dwarf binaries thousands of light-years away—opening an entirely new gravitational window on planetary populations that survive stellar evolution12.

Source notes

  1. 2026-11-11 · AI Is the New Electricity. And the Meter Is Running. · AMD Newsroom. ↩ ↩2 ↩3 ↩4 ↩5 ↩6 ↩7

  2. 2026-11-10 · How Gravitational Waves Can Reveal New Worlds · NASA Goddard Space Flight Center (GSFC). ↩ ↩2 ↩3 ↩4 ↩5