Eric Garofano

Entrepreneur / Executive

[Eric Garofano @eric_garofano]

@TheOneLanceB @ptj_official Such a great interview

[Eric Garofano @eric_garofano]

@TheOneLanceB Are you still planning to do this tomorrow?

[Eric Garofano @eric_garofano]

Data centers become model factories. Edge devices become the runtime. Innovation stays open. Cloud for frontier training. Edge for everything else. Hybrid wins.
This isn’t just engineering and economics. It’s about who gets to participate.

[Eric Garofano @eric_garofano]

Centralized AI inference is hitting a wall. Not just energy. Capital economics.
Data centers: 183 TWh in 2024 → 426 TWh by 2030. AI power demand: 4 GW today → 123 GW by 2035. Some regions seeing 5x more grid requests than total capacity.

[Eric Garofano @eric_garofano]

Billion-dollar data centers mean only a handful of corporations build the future. Distributed compute puts tools in the hands of innovators. A student with a laptop should be able to experiment and build. That’s how we got the internet. That’s how we should build AI.

[Eric Garofano @eric_garofano]

Training distributes too. Frontier models stay centralized. But fine-tuning, personalization, federated learning happen locally. Your device adapts without shipping data to the cloud.
This matters beyond efficiency. Centralized infrastructure means centralized control.

[Eric Garofano @eric_garofano]

NPUs won’t replace GPUs for frontier training. But inference is different. Forward passes, not backprop. NPUs run optimized matrix ops at watts, not kilowatts. Quantization and distillation already put billion-parameter models on device. The gap narrows every generation.

[Eric Garofano @eric_garofano]

Capex running 10-20x revenue. That gap closes or the model breaks.
Edge comp changes everything. Power distributes across existing grids. No new transmission. No concentrated loads. Hardware costs shift to consumers already buying devices on 5-7 year cycles. AI prov. shed capex.

[Eric Garofano @eric_garofano]

~$7T in global data center capex through 2030 (McKinsey). $1.5T in bonds for buildout (J.P. Morgan). $720B for grid upgrades (Goldman). Then it repeats.
Meanwhile, 2024 generative AI revenue: ~$16-26B. Industry spent $50B on Nvidia chips, generated $3B in AI revenue (Sequoia).

[Eric Garofano @eric_garofano]

The problem isn’t total energy. It’s concentration. Gigawatt-scale loads need transmission lines, substations, and generation that take years to build.
Then there’s GPU obsolescence. 3-5 year refresh cycles mean this isn’t one-time spend. It’s recurring.

[Eric Garofano @eric_garofano]

@tradertheory 🤣

[Eric Garofano @eric_garofano]

@TheShortBear I agree. Growth expectations are being questioned and it’s showing on the charts.

[Eric Garofano @eric_garofano]

@TheShortBear Quantum is years from a justifiable growth-stock profile. Depending on your horizon, the smart trade isn't in quantum stocks but in who controls the constraints future breakthroughs depend on.

[Eric Garofano @eric_garofano]

@TheShortBear I agree. Quantum tech isn’t ready.
It’s application-specific, physics-limited, and narrow in commercial scope. Each potential use: cryptography, material science, optimization depends on hardware that can’t yet scale or stabilize.

[Eric Garofano @eric_garofano]

@TheShortBear AI faces the same bottlenecks: chips, power, cooling, and talent but with vastly broader applications and immediate revenue paths. If those constraints are tightening AI today, they’ll strangle quantum tomorrow.

[Eric Garofano @eric_garofano]

@TheShortBear While capital pursues quantum technology, true scarcity lies upstream: advanced fabrication, materials, power, cooling infrastructure, precision cryogenics, and specialized labor bridging physics and computer science. These supply chains will reprice first.

[Eric Garofano @eric_garofano]

@MartinShkreli Medicine is too slow. Innovation is abundant but regulation suffocates even the most exciting advancements and gums the gears.

[Eric Garofano @eric_garofano]

@LovinTheTruth5 @MakisMD @elonmusk @grok Methylene blue is an often overlooked tool for brain health. It sharpens focus, boosts mental clarity, and can lift mood thanks to its impact on mitochondrial energy and neurotransmitters. It also helps protect the brain from oxidative stress.

[Eric Garofano @eric_garofano]

@TheOneLanceB 💯

[Eric Garofano @eric_garofano]

@str8outtaBNA @realDonaldTrump It’s may actually stop a devastating war and it’s not new. President Trump is acting within his legal authority. Under the War Powers Resolution of 1973, the President is permitted to initiate limited military action without a formal declaration of war.