Futurist, accelerator, reader and occasional thinker. Dad x3, husband x1, founder @Theia3D, @cmndline & @keywire_io.
California, USA
Joined April 2008
- Tweets18.7K
- Following1.2K
- Followers14K
- Likes12.7K
X just revealed it uncovered a Chinese bot farm of roughly 200,000 accounts. Hundreds were used to influence the US debate over AI data centers and energy policy.
A foreign government running coordinated operations to slow American infrastructure development. Not military targets. Not elections. Data centers and energy policy.
This confirms what's been building for months.
In June, OpenAI found Chinese-origin ChatGPT accounts generating social media content claiming data center buildouts increase electricity costs. The campaign targeted Virginia and Iowa specifically, trying to fuel NIMBY opposition and create political costs for officials supporting data center investments.
Developers reported surges of anti-data center messages traced back to accounts with matching IPs and suspected bot activity.
Now X confirmed the scale. 200,000 accounts. Hundreds directly targeting infrastructure debate.
The strategic logic is obvious.
China made data centers a priority in its 15th Five-Year Plan. It requires 80% renewable power for new facilities in national computing hubs. It controls nearly half of global battery storage capacity. Its DC racks grew 30% annually from 2016 to 2023.
China is building. It has a strategic interest in making sure we don't.
163 US moratoriums. $286 billion stalled. Half of planned data centers at risk. Communities opposing infrastructure they've been told is dangerous, wasteful, unnecessary.
How much was organic? How much was manufactured?
The moratorium movement isn't fake. Real communities have real concerns. But a foreign government spending resources to amplify and weaponize those concerns against American infrastructure changes the conversation.
Every data center that doesn't get built here is competitive advantage for the countries building them. At least one was paying to make sure we didn't.
Escondido, California just ended its battery storage moratorium. Unanimously.
The city didn't ban BESS. It didn't extend the pause. It studied the technology, developed safety guidelines, and lifted the moratorium with real rules. 5-0 vote.
New requirements: plume modeling reports, flammability and toxicity evaluations for all failure scenarios, gas dispersion simulations, fire suppression, explosion control, 100-foot setbacks.
The moratorium started after a September 2024 fire at an SDG&E storage facility forced evacuations. The fear was real. Instead of turning a temporary pause into a permanent wall, Escondido wrote the rules and moved forward.
Meanwhile in the same timeframe:
Dresden, NY adopted a one-year moratorium August 10. North Elba drafting one. Panama, NY considering one. Oshtemo, Michigan extended theirs. New York now has 128 active BESS moratoriums.
Every one of these towns said the same thing Escondido said: we need time to study this.
The difference is Escondido actually finished studying.
A moratorium is supposed to be temporary. A pause to develop regulations. Not a permanent block dressed up as caution.
Escondido proved the process works. Study the technology. Write safety standards. Move forward with rules that protect communities and allow development.
128 towns in New York could learn from one city in California.
Liquid Death and Garage Beer just launched a campaign starring Jason Kelce asking Americans to mail their pee to data centers to help cool the servers because, they claim, data centers "waste millions of gallons of water."
The limited-edition "Data Center Coolant Collector" jars sold out at $18 each.
It's funny. It's absurd. It's great marketing. And it's built on a stat that completely falls apart under any scrutiny.
Data centers use approximately 0.3% of the total US public water supply. For context, golf courses use 476 billion gallons a year. Residential swimming pools use 200 billion. Lawn irrigation alone burns through 9 billion gallons per day. Agriculture accounts for 73 billion gallons daily. Data centers used roughly 17.5 billion gallons total in 2023.
And the "millions of gallons" framing ignores what's actually happening in the industry. Closed-loop cooling systems cut freshwater use by up to 70%. Microsoft deployed zero-water cooling in Phoenix and Wisconsin. Google uses reclaimed or non-potable water at more than 25% of its campuses. Amazon is expanding that approach to over 120 data centers by 2030.
The water problem is real for older facilities. It's being solved by newer ones. And the trajectory is toward less water use per unit of compute, not more.
But nuance doesn't sell beer.
The deeper issue isn't the joke. It's what the joke reveals about public perception. When a beverage company can build a viral campaign around data center fear, it tells you how deeply the misinformation has landed. People genuinely believe data centers are draining their water supply. The actual number is 0.3%.
This is what happens when an industry lets other people define its narrative. When you don't show communities what these facilities actually do, someone else will. And they'll do it with a jar of urine and a Jason Kelce song.
The data centers aren't the water problem. The communication gap is.
The US power grid is running out of margin.
PJM, covering 13 states and 65 million people, failed to procure enough power for the first time in its history. Its capacity auction came up 6,625 MW short of reliability targets for 2027/28. Reserve margin fell to 14.4%, nearly six points below the 20% threshold. Capacity prices hit a record $333/MW-day.
Data centers account for 94% of PJM's projected load growth.
ERCOT in Texas: reserve margin just 9.8%, below its 13.75% target. Load requests quadrupled from 63 GW to over 230 GW in one year.
NYISO in New York: 417 MW reliability margin this summer. Lowest in recent history. That's the buffer between normal operations and emergency conditions for the entire state.
NERC added 58 GW of new generation and still flagged New England and the Pacific Northwest as elevated risk. Their own reliability director: "The improved conditions shouldn't be interpreted as saying overall reliability risk is declining."
The grid is stretched thin in every major region.
What stabilizes a grid under this pressure? Battery storage. It manages peak demand, integrates renewables, provides the fast-response capacity that keeps brownouts from becoming blackouts.
New York has 116 municipal moratoriums on battery storage. Nationally, 140+ local groups have blocked or delayed $60 billion in energy projects.
The grid needs storage. Communities are blocking storage. The math doesn't work.
Every moratorium makes the margin thinner. Every blocked project makes the next heat wave more dangerous. PJM is already short. ERCOT is already tight. NYISO is at its lowest margin ever.
The grid doesn't care about local politics. It cares about physics.
$7 trillion in projected global data center investment between 2026 and 2030.
The top five hyperscalers are spending over $600 billion this year alone. Up 36% from 2025. The capital exists. The demand exists.
The question is where it gets built. And increasingly, the answer is: not here.
Almost half of US data centers planned for 2026 are expected to be delayed or canceled. Over 140 local groups have blocked or delayed $60 billion in US projects in just over a year. 116 municipalities have imposed moratoriums. Maine banned facilities over 20 megawatts outright.
While the US stalls:
Europe has a $569 billion pipeline. Brookfield and Data4 are putting $20.7 billion into France and Northern Europe. Hyundai is building a $6.3 billion AI hub in South Korea. Google is breaking ground on $15 billion in projects in India. The UAE is scaling AI-ready capacity through sovereign deals with Microsoft and G42.
The US constraint is no longer capital. It's transformers on 5-year delivery schedules, switchgear sold out through 2028, and a regulatory environment that treats data centers as problems to contain.
For companies whose AI workloads aren't latency-sensitive to US users, international sites now offer faster timelines and lower costs. Every moratorium, every blocked project, every new tax makes that math tilt further.
The US still leads with nearly 4,000 data centers. But leadership isn't permanent. If the most advanced AI systems end up trained and deployed from outside the US, not by choice but because we made it impossible to build here, the strategic fallout goes well beyond real estate.
Capital goes where it's welcome. It always does.
Bill Fishkin retweeted
we've had a lot of people hit us up with the "how do i get started" or the "i have a binder of old cards, how do i figure this out" so we decided to build a content series around it:
xitter.cf/keywire_io/status/2080…
When the grid can't serve you, you build your own.
That's exactly what's happening.
Chevron and Microsoft just signed a 20-year agreement to supply dedicated power for a data center campus near Pecos, Texas. On-site generation. One of the largest pairings of compute infrastructure and dedicated power in the country.
Meta locked in a 20-year power purchase agreement with Vistra for over 2,600 megawatts of zero-carbon energy from nuclear plants.
These aren't utility contracts. These are companies building parallel energy infrastructure because the existing grid can't deliver what they need on a timeline that works.
And that should concern everyone.
When the largest companies in the world start routing around the grid, it means the grid has failed as a platform. The public infrastructure that's supposed to serve everyone becomes something only smaller players rely on, while the companies with the deepest pockets build their own systems.
This is what happens at the end of a long chain of bad decisions.
Block battery storage and the grid gets less stable. Block data centers and the investment goes elsewhere. Let transformers sit on 5-year delivery schedules and switchgear sell out through 2028. Stack 104 moratoriums in New York, 14 proposed in other states, and $286 billion in canceled projects. And eventually the companies that need power will stop waiting for the grid and start building their own.
Microsoft isn't signing a 20-year deal with Chevron because it's the optimal outcome. It's doing it because the alternative, waiting for the grid to catch up, isn't viable.
The irony is sharp. Communities block infrastructure projects because they're worried about local impact. The result is that the infrastructure gets built anyway, just privately, at a scale and in a configuration that the public has even less input on.
The grid was supposed to be the great equalizer. If we keep making it impossible to build on, only the companies that can afford to build around it will survive.
#Infrastructure #DataCenters #Energy #PowerGrid #AI
A data center can be built in under three years. The power infrastructure to support it takes five to ten.
That's the gap that's breaking everything.
US data center electricity demand has nearly doubled in three years, from 23 gigawatts in 2023 to 42 gigawatts in 2026. The facilities are going up. The power systems behind them are not.
High-power transformers are on three to five year delivery schedules. Switchgear is sold out through 2028. Transmission upgrades that should take two years are stuck in permitting for five. The hardware bottleneck alone would be challenging. Layer in moratoriums, environmental reviews, and community opposition, and you get $286 billion in canceled or delayed projects.
The demand isn't slowing down. AI workloads are growing exponentially. Every major tech company is racing to build compute capacity. The electricity has to come from somewhere.
And this is where the battery storage problem and the data center problem become the same problem.
Battery storage is how you stabilize a grid under pressure. It's how you integrate renewables at scale. It's how you manage peak demand without building new fossil fuel plants. New York's own NYSERDA says 6 GW of storage by 2030 would save ratepayers $2 billion.
New York now has 104 moratoriums on battery storage.
The grid needs storage to handle the load. Data centers need the grid to operate. Communities are blocking both. And the timeline to build power infrastructure is already three to five times longer than the timeline to build the facilities that need it.
This is not a problem you can moratorium your way out of. The demand exists. The electricity has to come from somewhere. If you block storage, the grid gets less stable. If you block data centers, the investment goes elsewhere. If you block both, you get brownouts and lost economic competitiveness at the same time.
The facilities are ready. The power is not. And we keep making the power problem worse.
Last week I wrote that New York had 98 municipal moratoriums on battery energy storage.
The number is now 104.
Six new moratoriums in one week. New York accounts for 83% of all BESS moratoriums in the country.
But here's the part that should stop you.
In Iberville Parish, Louisiana, the council just imposed a two-year moratorium on battery storage. Their own parish president publicly said it could deter billions in investment and job creation.
The head of the local government is warning his own council that their decision will cost the community. They passed it anyway.
This is the pattern. The people closest to these decisions increasingly understand the cost. They see the tax revenue, the construction jobs, the infrastructure. And the moratoriums pass anyway because the fear is louder than the data.
104 moratoriums. Over 1 GW stalled. $2 billion in ratepayer savings at risk in New York. The grid approaching failure in multiple regions this summer.
I've watched this for 12 years in immersive tech. Communities block what they can't see. A parish president can say "this will cost us billions" and it doesn't matter because the residents can't visualize what these facilities actually look like or how they operate.
The data isn't winning. The economics aren't winning. Elected officials sounding the alarm aren't winning.
104 and counting.
This list should make you angry.
Empire State Building: 410 days.
The Pentagon: 16 months.
Hoover Dam: 5 years, finished 2 years early.
Alaska Highway, 1,700 miles through wilderness: 8 months.
Liberty Ships: 42 days average.
One built in 4 days.
That was a country that treated building as a form of national identity.
Now look at this week alone:
New York signed the first statewide data center moratorium.
New Mexico blocked a 2.5-gigawatt data center over 0.6 miles of pipeline. $286 billion in AI data center projects canceled or delayed since 2025.
Ninety-eight municipalities in New York have moratoriums on battery storage.
California lost a $3.2 billion shipyard to Texas. Companies are planning to build data centers in orbit because it’s easier than building here.
We went from the Hoover Dam finishing two years early to not being able to approve half a mile of pipe.
This isn’t about one project. It’s about what kind of society we’ve become.
The country that built the transcontinental railroad, the electrical grid, the interstate highway system, and the Apollo program did those things because it believed building was how you created prosperity. Infrastructure was the foundation of human flourishing. You built the dam so the valley had power. You built the bridge so a city could grow. You built the ships because there was a war to win.
That belief has been replaced by a default posture of suspicion. Every project is a threat. Every proposal gets a moratorium instead of a timeline.
The grid is approaching failure. NERC is warning about shortfalls this summer. We need battery storage for renewables. We need data centers for AI. We need pipelines to power them. We need shipyards for defense.
And the answer, in state after state: not here, not now, not without another decade of review.
The generation that built the Empire State Building in 410 days would not recognize us.
I’ve watched this pattern for 12 years in immersive tech. The problem isn’t that we can’t build. It’s that we’ve lost the ability to show people why building matters. Communities fear what they can’t see. And we’ve gotten very good at blocking things and very bad at showing what they make possible.
We used to be a pro-building society. A pro-flourishing society. A society that saw a river and thought “dam,” saw a strait and thought “bridge,” saw a continent and thought “railroad.”
That instinct isn’t dead. But it’s buried under moratoriums, reviews that outlast construction timelines, and a culture that treats every new project as something to stop.
We used to build.
We need to remember how.
how long it took America to build things:
transcontinental railroad: 6 years (1869)
Empire State Building: 410 days (1931)
golden gate bridge: 4 years (1937)
hoover dam: 5 years, finished 2 years early (1936)
the pentagon, largest office building on earth: 16 months (1943)
alaska highway, 1,700 miles through wilderness: 8 months (1942)
a liberty ship: 42 days average, one built in 4 days as a stunt (1943)
Since we're banning things that use water, here's the full list of things we need to shut down immediately:
"How do we save our farmland and stop all these dumb data centers?"
This went viral this morning. 6 million views and climbing. Tens of thousands of likes.
I understand the frustration. But nearly everything in that sentence is wrong.
Farmland first.
The US loses about 2,000 acres of farmland every day to non-agricultural uses. 75 million acres gone between 1997 and 2022. Data centers are a tiny fraction of that. All the land owned by data center companies, projected through 2028, would be 89 times smaller than the acreage we waste just on food that never gets eaten.
The farmland concern is real. Data centers are not the driver.
Now "dumb."
Data centers run the systems that route 911 calls, process Medicare claims, clear financial transactions, manage air traffic control, and power every AI tool you opened on your phone this morning. GPS. Telehealth. Fraud detection. Banking. All of it.
Calling them dumb is like calling power plants useless while your lights are on.
The economics.
Brookings: counties that get their first large data center see private employment rise 4 to 5 percent. Construction jobs jump 11 percent. Wages rise 3 to 4 percent.
In Mecklenburg County, Virginia, one Microsoft facility generated enough property tax revenue to fund $134 million in school renovations. Data centers now cover roughly 40 percent of property taxes in parts of the state.
An Ohio study found a single project can support 9,700 construction jobs and generate $2.4 billion in economic output.
Teachers hired. Roads paved. Property taxes lowered.
But here's the real problem.
Gallup reported in May that 71 percent of Americans oppose data center construction in their local area. The top concerns: natural resources, water, noise, energy use, quality of life.
Those concerns are legitimate.
They deserve real answers.
The industry has been terrible at providing them.
I've watched this pattern for 12 years in immersive tech. When communities can't see what's being built, fear fills the gap. A 90-slide environmental impact deck is not "showing people." It's burying them in data they don't trust from companies they don't know.
The solution isn't dismissing the concerns. It's letting communities actually see what these facilities look like. How they operate. Where the water goes. What the noise sounds like at the property line. What the tax revenue means for their specific school district.
The data centers aren't dumb.
The way we're communicating about them is.
"How do we save our farmland and stop all these dumb data centers?"
This went viral this morning. 6 million views and climbing. Tens of thousands of likes.
I understand the frustration. But nearly everything in that sentence is wrong.
Farmland first.
The US loses about 2,000 acres of farmland every day to non-agricultural uses. 75 million acres gone between 1997 and 2022. Data centers are a tiny fraction of that. All the land owned by data center companies, projected through 2028, would be 89 times smaller than the acreage we waste just on food that never gets eaten.
The farmland concern is real. Data centers are not the driver.
Now "dumb."
Data centers run the systems that route 911 calls, process Medicare claims, clear financial transactions, manage air traffic control, and power every AI tool you opened on your phone this morning. GPS. Telehealth. Fraud detection. Banking. All of it.
Calling them dumb is like calling power plants useless while your lights are on.
The economics.
Brookings: counties that get their first large data center see private employment rise 4 to 5 percent. Construction jobs jump 11 percent. Wages rise 3 to 4 percent.
In Mecklenburg County, Virginia, one Microsoft facility generated enough property tax revenue to fund $134 million in school renovations. Data centers now cover roughly 40 percent of property taxes in parts of the state.
An Ohio study found a single project can support 9,700 construction jobs and generate $2.4 billion in economic output.
Teachers hired. Roads paved. Property taxes lowered.
But here's the real problem.
Gallup reported in May that 71 percent of Americans oppose data center construction in their local area. The top concerns: natural resources, water, noise, energy use, quality of life.
Those concerns are legitimate.
They deserve real answers.
The industry has been terrible at providing them.
I've watched this pattern for 12 years in immersive tech. When communities can't see what's being built, fear fills the gap. A 90-slide environmental impact deck is not "showing people." It's burying them in data they don't trust from companies they don't know.
The solution isn't dismissing the concerns. It's letting communities actually see what these facilities look like. How they operate. Where the water goes. What the noise sounds like at the property line. What the tax revenue means for their specific school district.
The data centers aren't dumb.
The way we're communicating about them is.
$286 billion in AI data center projects have been canceled or delayed since the start of 2025.
Morgan Stanley.
And Polymarket is now pricing a 26% chance AI data centers get launched into orbit by the end of next year.
We have made it so difficult to build on the ground that the industry is seriously considering building in space.
SpaceX is targeting late 2027 for its first orbital data centers. Twenty meters tall, 70-meter wingspan. Google's Project Suncatcher is sending prototype satellites around the same time. A five-month-old startup just asked the FCC for permission to deploy 100,000 data center satellites.
This isn't science fiction. This is what happens when you block $286 billion worth of projects on the ground.
Meanwhile, the grid:
NERC is warning about electricity shortfalls across all three US interconnection regions this summer. New York's grid operator says they're expecting their smallest power surplus ever. PJM projects a 6 GW shortfall by 2027. Texas saw data center interconnection requests quadruple in a year.
The grid is buckling under current load. And we're blocking the two things that would fix it.
Battery storage stabilizes the grid. 98 municipalities in New York alone have passed moratoriums on it.
Data centers power the AI economy. Fourteen states have proposed moratoriums on those too.
So we're starving the grid of storage AND blocking the compute infrastructure the country needs to compete. The capital goes to Virginia. Texas. Georgia. Or now, apparently, low Earth orbit.
The companies building in space aren't doing it because it's better. They're doing it because we've made it nearly impossible to build here.
I've watched this pattern for 12 years in immersive tech. Communities fear what they can't see. Fear doesn't read spec sheets. And no one has figured out how to show these projects to people in a way that actually registers.
We have a grid approaching failure, an AI race we can't afford to lose, and we're actively blocking the infrastructure that would solve both.
This isn't a technology problem. It's a communication problem.
And it's now literally pushing human progress off the planet.
New York is now blocking both the energy infrastructure that AI needs and the energy infrastructure that clean energy needs.
The economic cost is staggering.
Today Governor Hochul signed the country's first statewide moratorium on hyperscale data centers. Any facility using 50+ megawatts is frozen for up to a year.
But New York was already doing this at the local level.
98 municipalities across 37 counties have passed moratoriums on battery energy storage. That's two-thirds of all BESS moratoriums in the entire country. Over 1 GW of projects stalled. Developers eating $2M+ in non-refundable costs every time a project dies.
New York's own target: 6 GW of storage by 2030. NYSERDA estimates that would save ratepayers $2 billion in transmission costs. Those savings are now at risk.
Now add data centers.
Data center construction has been the construction industry's golden goose in 2026. Everything else flatlined. These megaprojects surged.
The moratorium ensures that investment, those jobs, and that tax revenue flow to Virginia, Texas, Georgia. Fourteen other states have introduced similar bills. None signed them into law. New York went first.
The Pipe Trades Association is worried about construction jobs vanishing. The Data Center Coalition says the moratorium "sends a signal that the state is closed for business."
Here's what gets me:
New York wants to lead on clean energy.
Battery storage is essential to that.
New York wants to lead on AI.
Data centers are essential to that.
The state is now actively blocking both.
Not because the technology doesn't work. Not because the engineering is unsafe. Because communities don't understand what these facilities actually are, and the industries building them haven't figured out how to show them.
I've spent 12 years in immersive tech watching this same pattern across every industry. Fear doesn't read PDFs. A 90-slide safety deck doesn't win a town hall where parents are scared about fires or utility bills.
The developers who let communities see these projects at real scale, in context, are getting through. The ones showing up with data alone are getting moratoriums.
New York is leaving billions in investment and thousands of construction jobs on the table.
The storytelling problem just got a lot more expensive.
New York just became the first state to freeze data center construction. One year. No new permits for anything over 20 MW.
The coverage talks about grid strain and rate impacts. But anyone who's sat in the hearings knows what built the political will: communities being asked to accept infrastructure nobody ever showed them.
Every BESS town hall has the same moment.
Somebody stands up and asks about fires. Not politely. With the voice of a parent who read one headline and is now certain this thing is going to explode next to a school.
The developer pulls up thermal runaway statistics. NFPA 855 compliance. UL 9540A test results.
The room doesn't care.
The question was never about fire statistics. It was about fear. Fear doesn't process data. It processes images. It remembers the Arizona headline from 2019. It imagines the worst version of the thing it doesn't understand.
Then it votes accordingly.
And now it legislates accordingly. New York's interconnection queue went from about 1 GW of large-load proposals in 2022 to roughly 12 GW by late 2025, and almost none of those developers ever showed a community what that actually looks like from their street. The result wasn't 48 tough negotiations. It was one statewide moratorium.
The same dynamic is playing out for storage in New Jersey, Virginia, and Connecticut. The states that need grid-scale storage the most are blocking it project by project. New York just proved opposition can scale to an entire asset class overnight.
What actually changes the room:
Let residents stand in a virtual replica of their own neighborhood and see the facility at true scale. The setback distances. The landscaping buffer. The sound wall. From their kitchen window.
The fire question still comes up. But now instead of reading a statistic, you show the suppression system. The thermal management. The spacing between units. What "NFPA 855 compliant" actually looks like in three dimensions.
The conversation goes from "convince me this is safe" to "oh, that's what it actually is."
We've spent 12 years at @Theia3D watching the same pattern: when people can see it, they stop fearing it.
The developers who visualize before the first town hall get permitted. The ones with PDFs get moratoriums. New York just made that sentence literal.
We've been building interactive experiences for trade show floors since 2014. Twelve years of watching the same pattern play out.
Company spends six figures shipping physical prototypes across the world. Custom displays. Freight. Crates. Setup crew. Lighting rig.
Attendee walks up. Glances at it. Grabs a brochure. Moves on. Thirty seconds.
The booths that actually hold people are the ones where the attendee can do something. Tear apart an assembly. Swap configurations. Zoom into tolerances that a physical model can't show at any scale. Interact with the product instead of staring at it.
One shift I've watched accelerate over the last two years: the best exhibitors are treating their booth less like a display and more like a data collection instrument.
Every interaction in a digital experience is trackable. Which configurations get the most attention. Where people spend time. What they skip. You walk out of a show with actual buyer intelligence instead of a fishbowl of business cards.
The companies figuring this out are playing a different game than the ones still shipping crates and hoping people stop walking.
Trade shows aren't dying. But the static booth is.
Bill Fishkin retweeted
1/7 Anthropic dropped an infographic in February showing how AI eats every job vertical through 2040.
What surprised me wasn't the chart. It was the conversations after. Smart people who should know better, still talking like it's 2010.
So I walked the next 14 years 🧵
Bill Fishkin retweeted
The "why" behind why I'm launching @KeyWire_io
If you're a collector, you'll understand.
It's fun, it's an escape, it's joy, it's a part of your life and personality. You find like-minded people and they become your people. It's community. It's life.
I've loved comics and cards since I was probably 4 or 5 years old. The only things I love more than collectibles are my family and friends. That's it. It's always been that way. I vividly remember being maybe 5 and being obsessed with Spider-Man, Batman, and Star Wars. As an adopted only child that lived in a rural New England town in the pre-internet stone age, I spent a lot of time by myself. Imagination was my companion. Art and story were my reality escape. Batman, Spider-Man, and the cast of Star Wars were all I could think about. This is deeply ingrained in my DNA.
Clearly, that kid never really left. He just got older and the collections got bigger. The feeling when you pull a card, dig a key issue out of a longbox at a show, or close on a grail you've been hunting for years, that feeling is the same at 8 as it is at 38. It doesn't age. It's the closest thing to pure joy I know outside of the people I love.
My kids were raised on this. We've waited in line at Target on Thursday mornings for Pokémon. We stood in line seven hours at San Diego Comic-Con just for a shot at buying exclusive Legos. We dig through longboxes together. We'll never stop watching anime.
But somewhere along the way, the hobby got noisy. Prices move faster than anyone can track. Hype cycles spin up overnight and vanish by morning. Scalpers pump products and leave real collectors holding the bag. The useful information lives scattered across a hundred Discords, random X threads, private spreadsheets, and group chats you're not in. The actual signal gets buried every day. The fragmentation just seems endless.
All of this preprogramming led me to launching KeyWire.io.
What started as an idea for a newsletter grew quite a bit as a few friends and my boys got involved. Now we've built the collectibles site I've always looked for. KeyWire is the front page of collectible markets. One surface covering Pokémon, Magic, One Piece, Lorcana, Gundam, comics and soon sports cards, retro games, anything that grades and anything that moves. A Fear & Greed Index for every category. Original research under our Signal desk. Movers, velocity, grading pop deltas, mispriced flags. Giveaways, drops, leaderboards, community. The loud fun that makes collecting what it is.
We're opening the KeyWire doors slowly, starting now. I just spent the last 20 minutes clearing a bunch of folks who signed up. If you want full access to the site early, grab a subscription. Subscribers get the most exposure to product giveaways (with tons of extra entries), and our Buyers Club launches very soon.
WTF is the Buyers Club?
When we get Pokemon ETB's, One Piece cases and other product at or near cost, we turn around and offer it to Buyers Club members for just a bit above our cost plus shipping. You save real money over time and get stuff shipped right to your door. We hate scalpers just as much as you do (and we're not stopping there).
We're starting small. Twenty Buyers Club slots to begin with. FCFS at first, then offered to existing Pro subscribers, then Collectors and invite-only based on Quests and community activity. We scale the club alongside our supply. We'll learn as we go.
Buyers Club members also get access to cards from the KeyWire packs we rip, priced 20 to 30 percent below FMV. We've come up with a fun way to do this. More on that soon.
If any of this hits, go grab your subscription at keywire.io.
Signals, not noise.
I'm building this for the kid sitting on his bedroom floor in rural New England with a stack of comics and cards and nobody to talk to about them. There are millions of us out there now. We deserve somewhere serious to go.
Bill Fishkin retweeted
Blitz S1 wraps tomorrow at 12PM ET. You came, you saw, you shot each other up.
Congrats @cmndline for shipping 2 new maps, new features, and rebuilding the social layer from scratch. They listened, kept pushing, and we’re excited to see Blitz come back with even more in store.
