greg copeland

Retired

[greg copeland @GregCopeland56]

@Simi_2210_ +/- 4

[greg copeland @GregCopeland56]

@curious_X_ @Rainmaker1973 There's a small airport in Ozark mo lined with houses with hangers in back and garage in front

[greg copeland @GregCopeland56]

And the plunder
Larry Fink (BlackRock CEO) has publicly stated that trillions of dollars needed for AI infrastructure—including data centers and new power generation—will largely come from ordinary Americans’ long-term savings, retirement accounts, and pension funds.4db551
The Core Statement
In recent remarks (around May 2026, including at events like the Milken Institute Global Conference), Fink highlighted massive capital requirements for AI:
Building a single 1 GW data center can cost $50–75 billion.
Overall U.S. needs for data centers, power grids, compute, chips, etc., run into trillions.
Governments face deficits and can’t fund it alone; tech companies also can’t shoulder everything.
Key quote/context: Much of this capital “will come from savings accounts and pension accounts.” BlackRock clarifies this refers to long-term investment vehicles like 401(k)s, IRAs, and pensions—not literal bank savings accounts with fixed interest.a1be74
Fink frames this as necessary for U.S. competitiveness in AI and views it as an investment opportunity (potentially creating new asset classes like “futures on compute”). BlackRock is actively involved via partnerships (e.g., with Microsoft, MGX, NVIDIA, xAI) to channel private capital into data centers and energy infrastructure, with potential for $100 billion+ in mobilized investment.294257
Context in the Broader Power/Data Center Debate
This ties directly into power concerns: Data centers’ surging electricity demand requires new plants, transmission lines, and upgrades. Traditional utility funding often socializes costs to all ratepayers (leading to higher bills). Fink’s point is that private markets—fed by retirement/investment dollars—must fill the gap through equity, debt, and infrastructure funds.f888f1
BlackRock’s actions: Acquiring utilities (e.g., stakes in AES, Minnesota Power via GIP) and data center operators (e.g., $40B Aligned Data Centers deal) to directly invest in and profit from the buildout.8c880e
Counterpoints/criticisms: Some see this as shifting costs/risks onto everyday investors and ratepayers, with potential conflicts if BlackRock-owned utilities prioritize data center loads. Others note it’s standard capital allocation—pension funds seek returns, and AI growth could deliver them. Tech firms (Microsoft, Anthropic) have pledged to cover incremental electricity costs in some cases to protect residential bills.6d2ce2
Bottom line: Fink’s comments reflect a realistic view of where the money for the AI power buildout is likely to come from—pooled retirement and investment capital managed by firms like BlackRock—rather than direct taxpayer bailouts or solely corporate balance sheets. This aligns with the massive scale of data center electricity needs discussed earlier. For the full context, check Fink’s recent letters or BlackRock’s AI infrastructure partnership announcements.

[greg copeland @GregCopeland56]

Power concerns for data centers in general center on their massive, always-on electricity demand—driven heavily by AI—which strains grids, raises costs for everyone, delays projects, and creates environmental trade-offs. This is a global issue, though most acute in high-growth areas like the U.S., Northern Virginia, Texas, Ireland, and parts of Asia.ce40bc
Scale of Electricity Demand
Data centers (especially hyperscale AI ones) are among the fastest-growing electricity users:
U.S.: In 2023, ~176 TWh (4.4% of total U.S. electricity). Projections for 2028 range from 325–580 TWh (6.7–12%). Data centers drove ~50% of U.S. electricity demand growth in 2025. By 2030, forecasts reach 400–600+ TWh in some scenarios.dabaa6
Global: ~415 TWh in 2024 (~1.5% of world electricity). Projected to ~945 TWh by 2030 in base cases (nearly doubling), or up to 1,050+ TWh by 2026 in higher estimates. AI workloads (especially inference) are a big driver.4079b3
A single hyperscale facility can use 100 MW+ (equivalent to 100,000 households). Large clusters rival entire cities or states in demand.f6bc27
Key Power-Related Concerns
Grid Strain and Reliability:
Interconnection queues are backlogged (wait times up to 7+ years in places). Nearly half of planned 2026 U.S. projects face delays/cancellations due to power availability.52798d
Sudden load fluctuations (e.g., AI training vs. inference) challenge grid stability. Examples include Virginia incidents where data center disconnections caused surpluses requiring emergency fixes.b853fb
Aging infrastructure (much of the U.S. grid from 1950s–1970s) exacerbates issues. Utilities are adding gas plants, batteries, and transmission, but this lags demand.973bae
Cost Impacts on Consumers:
New infrastructure (plants, lines, transformers) raises rates. In some markets (e.g., PJM), capacity costs surged dramatically, with much passed to households. Surveys show high public concern over rising bills.969af0
Data centers often get special deals, but costs can be socialized. Some states now require them to fund upgrades directly.264687
Environmental and Emissions Trade-offs:
Reliance on fossil fuels (gas, sometimes coal/diesel backups) increases CO₂, NOx, particulates, and air quality issues. Backup generators (thousands in some regions) add health risks even if "emergency-only."bc40ed
On-site gas generation is rising due to grid delays. Renewables procurement is common (many operators aim for 100% offsets), but additionality and timing matter—new demand often fills with whatever is available.c06712
Other Constraints:
Water tie-in: Power plants (for data center electricity) also use water; cooling for data centers compounds local strain (evaporative use, wastewater with chemicals/biocides).9b5622
Permitting, land, and supply chain bottlenecks (transformers, turbines).

[greg copeland @GregCopeland56]

AI data centers do generate contaminated wastewater through their cooling processes, though it's more accurate to call it "polluted" or "chemically treated" discharge rather than purely "toxic water" in the sensational sense. This is a real environmental concern tied to the massive scale of AI infrastructure, but the impacts vary by location, technology, and regulations.912a1b
How Data Centers Use (and Contaminate) Water
Data centers, especially those powering AI training and inference, generate enormous heat from servers. Traditional evaporative cooling (using cooling towers) is common:
Large volumes of water absorb heat.
Much of it evaporates (this is "consumptive" use — the water is effectively removed from the local cycle).
The remaining water becomes concentrated with minerals, salts, and added treatment chemicals. This "blowdown" or discharge water is then released.6f51f2
Common contaminants in the wastewater include:
Biocides (e.g., glutaraldehyde, isothiazolinones) — to prevent bacteria/algae.
Corrosion inhibitors (e.g., phosphates, molybdates, nitrites, azoles) — to protect pipes and equipment.
High total dissolved solids (TDS) — concentrated minerals, salts, and chlorides.
Heavy metals — traces of copper, zinc, lead from equipment degradation.
Other chemicals — like ethylene glycol in some systems, or refrigerants.dbe3fa
In some cases, this discharge can contribute to issues like eutrophication (algae blooms that deplete oxygen in waterways), harm to aquatic life, or added strain on municipal wastewater treatment plants.faace2
PFAS "Forever Chemicals" Connection
A growing issue is PFAS (per- and polyfluoroalkyl substances) used in:
Certain advanced cooling fluids (e.g., two-phase immersion cooling).
Semiconductors in servers.
Fire suppressants.
These don't break down easily and can lead to long-term environmental persistence. Data centers drive demand for PFAS-producing companies, indirectly increasing pollution near manufacturing sites (e.g., historical issues in North Carolina). Direct emissions from data centers are often limited because many systems are closed-loop, but leaks or disposal add risks.1fbcb6
Broader Context and Examples
Water consumption + quality: A single large facility might use millions of gallons daily. Much evaporates; the rest is often more polluted when discharged. In places like Oregon, data centers have been reported to concentrate existing agricultural pollutants (e.g., nitrates) in discharged water.eb7aff
Chip manufacturing: Separate but related — semiconductor production for AI hardware generates wastewater with heavy metals and other tough pollutants.406df6
Mitigation efforts: Many operators use recycled/graywater, on-site treatment, air cooling, or immersion cooling to reduce impacts. Some places (e.g., deals with xAI in Memphis) involve building dedicated treatment plants. Regulations require pre-treatment in high-density areas like Northern Virginia.9c1b09
Not all claims are equal — some social media exaggerates by saying data centers "poison the supply" outright, while others note that properly managed discharge is treated and regulated similarly to other industrial facilities (power plants, factories). However, rapid AI growth is outpacing infrastructure in some areas, leading to real strains on water resources and quality

[greg copeland @GregCopeland56]

@karthikponna19 Mumps

[greg copeland @GregCopeland56]

@ronsterd89 Missing the other half

[greg copeland @GregCopeland56]

@BrianHa24140826 @WallStreetApes Eminent domain is abused in most cases. The cowboy stadium land was acquired with eminent domain. For the benifit of politicians and the cowboy franchise

[greg copeland @GregCopeland56]

@overstreetha I thought it was democrats that were rich. Plenty of billionaires in silicon Valley and almost all are democrats

[greg copeland @GregCopeland56]

@LyndalScranton @weatherchannel

[greg copeland @GregCopeland56]

@rajyoga_ False

[greg copeland @GregCopeland56]

@CubPod79321 @Simi_2210_ Cube root of 27. 3 solutions.

[greg copeland @GregCopeland56]

@elonmusk @waitbutwhy Looks more than parabolic. Maybe exponential or exponential of a polynomial or is it more like double exponential. Uncharted

[greg copeland @GregCopeland56]

@Perry_smitzh After death

[greg copeland @GregCopeland56]

@TheThe1776 Believe it or not. Springfield mo. Was top 10 in murder and homoside in the us. Now it's not reported.

[greg copeland @GregCopeland56]

@melgibsonhq Justified imho. The protesters have demonstrated that they are violent. Have even killed in similar conditions. Regular citize some trust could be used

[greg copeland @GregCopeland56]

@itsjordanclark With starlink and tesla phones?

[greg copeland @GregCopeland56]

@ElofsonJess They plan to burn down isa on Saturday the 14. See https://t.co/8mkLuGnD76

[greg copeland @GregCopeland56]

@NicetryMYAH @Breaking911 @grok If they considered it a retirement fund and using historical s&p data it be worth 100k by 65

[greg copeland @GregCopeland56]

@LionelMedia All the data is already in the nsa Utah data center since 2014. Collects all phone conversations. Text messages. Cc purchases and nearly anything else you van think if