Observed Signal · Jun 18, 2026 · Podcast Interview · Source: AINews swyx · Impact: 3/5 · Sentiment: Positive
AMP’s Anjney Midha on Output‑Maxing AI Compute Grids
In a podcast interview published June 18, 2026, Anjney Midha (founder & CEO of AMP) argues that frontier AI scaling is increasingly a systems and utilization problem rather than just a GPU supply problem. Midha contrasts low reported MFU (Model FLOPs Utilization) in some labs with historical and best‑in‑class MFU figures (GPT‑3 ~21%, Gopher ~32%, PaLM ~46%; today’s best around 60–70%) and criticizes wasteful, non‑iterative data‑center practices. He describes AMP’s vision for an independent compute grid — analogous to an electric ISO — that pools supply and demand across clouds and silicon to make “FLOPs flow like megawatts.” AMP aims for roughly 1.2 GW base‑load capacity and ~6 GW spike capacity over several years. The conversation also covers responsible infrastructure, community buy‑in for data centers, research hoarding at large labs, end‑of‑life healthcare prediction as a priority application, and the concept of “output maxing” as a discipline for frontier systems.
Discussion introduces a grid‑style compute pooling model and concrete GW‑scale capacity targets; changes to how compute is provisioned and utilized could materially affect AI infrastructure costs, data‑center planning, and the chip/ecosystem market.
Track Google DeepMind Signals & Market Shifts in Real-Time
Polaris7 autonomous intelligence agents track regulatory filings, primary sources, executive changes, and deal flow 24/7. Create your free Explorer workspace to monitor these entities.
Key Takeaways & Evidence Grounding
- Anjney Midha is founder and CEO of AMP (AMP PBC) and guest on the podcast.
- Midha cites historical MFU examples: GPT‑3 ~21%, Gopher ~32%, Megatron‑Turing NLG ~30%, PaLM ~46%, and asserts best‑in‑class MFU today is ~60–70%.
- AMP positions itself as an independent system operator (ISO‑like) for compute, pooling supply and demand across clouds and silicon.
- AMP stated ambitions include ~1.2 GW of base‑load compute capacity and an estimated need for ~6 GW of spike capacity over the next four years.
- Midha has invested in or backed frontier AI teams including Anthropic, Mistral, Black Forest Labs and Periodic Labs.
Connected Companies & Entities
8 Entities mappedOntology Mapping & Concepts
Related Market Signals & Shifts
Recent verified developments and strategic activity across this market segment.
Executives: AI Demand 'Almost Unlimited' Amid Valuemaxxing
AI executives told CNBC that demand for compute tied to generative AI remains extremely strong and, in many cases, exceeds available capacity across chips and data centers. Leaders including Pat Gelsinger, Marc Boroditsky, Andrew Feldman and others said energy, data center space and component supply are primary bottlenecks. At the same time, enterprises are shifting from carefree 'tokenmaxxing' usage toward 'valuemaxxing' — focusing on return on investment and rationalizing AI spending by choosing models and workloads that match value. Market moves such as Meta and xAI selling or renting excess compute capacity have sparked debate about overcapacity, but several executives said those are isolated cases and do not reflect a broad slowdown.
Import AI: Control Inversion, Intelligence per Watt, 100k+ GPUs
This Import AI issue highlights three major developments: a new paper by Anthony Aguirre (Future of Life Institute) called “Control Inversion” arguing that increasingly capable, autonomous AI will tend to absorb power from humans rather than grant it, raising hard safety and governance questions; a Stanford + Together AI research effort that proposes an “Intelligence per Watt” metric for measuring on-device model efficiency and coverage, finding local open-weight models now answer 88.7% of single-turn queries and accuracy-per-watt improved ~5.3× over two years; and Meta/Facebook’s publication of NCCLX, a heavily customized NCCL variant designed to run synchronized training on clusters exceeding 100,000 GPUs (claiming up to 12% per-step latency reduction on some Llama 4 runs). The newsletter notes continuing cloud capability and efficiency advantages, caveats about single-turn metrics, and broader implications for compute scale, on-device AI, and AI safety.
AI leaders warn of chip, energy and architecture bottlenecks
At a Milken Global Conference panel hosted by TechCrunch, five executives across the AI supply chain — Christophe Fouquet (ASML), Francis deSouza (Google Cloud), Qasar Younis (Applied Intuition), Dimitry Shevelenko (Perplexity) and Eve Bodnia (Logical Intelligence) — discussed structural constraints facing the AI industry. They highlighted near-term chip supply limits, growing energy and cooling challenges (including exploration of orbital data centers), and debate over model architectures. Google Cloud cited rapid revenue and backlog growth, ASML warned the market will be supply‑limited for years, and Applied Intuition emphasized real‑world data scarcity for physical autonomy. Bodnia described energy‑based models (EBMs) as a different architecture that uses far fewer parameters and updates online. Perplexity outlined agent products with granular permissioning and approval flows. Panelists also raised geopolitics and sovereignty concerns for physical AI systems and noted the potential societal and workforce impacts of accelerating AI capabilities.
Track Real-Time Market Signals & Shifts
Set up custom watchlists to receive automated, evidence-grounded executive digests whenever material signals or shifts occur across your tracked landscape.