TechFlow Logo
Login/ Sign up
ETH Gas
Gwei
Fear
gas
Why Is the Largest Listed Mining Company MARA Pivoting to AI and Selling Bitcoin?

Why Is the Largest Listed Mining Company MARA Pivoting to AI and Selling Bitcoin?

2026.07.28
Share
Bitcoin mining will not disappear, but mining companies are shifting from computing power providers to energy platforms; the window for Bitcoin as a medium of exchange has passed, and high-frequency payments will be handled by stablecoins.

Written by: Wu Shuo Blockchain

In an interview with Natalie Brunell on July 23, 2026, MARA CEO Fred Thiel stated that as demand for AI computing power continues to grow, electricity has become a core resource contested by Bitcoin mining companies, cloud service providers, chip manufacturers, and large model developers. MARA is gradually transitioning from a single-product Bitcoin miner to an energy platform controlling land, power, and data center infrastructure. He believes Bitcoin mining will not disappear, but will increasingly rely on surplus power from data center commissioning, microgrids, and renewable energy sources. Bitcoin itself remains an important store of value asset, but due to its lack of yield and high price volatility, its window for developing as a medium of exchange may have passed; high-frequency transactions and machine-to-machine payments are more likely to be handled by stablecoins.

How Did I Become the CEO of MARA, the World's Largest Publicly Traded Bitcoin Mining Company?

Fred Thiel: I've worked in the tech industry for 40 years. I started as a software programmer, but didn't perform particularly well, so the company moved me to product and sales roles, where I performed much better. Since then, I've worked in almost every niche of the tech industry. Around age 40, I was lucky enough to lead an IoT company through its IPO, and later worked in private equity and venture capital for a while.

Later, I had basically retired, mainly serving on some boards of directors and advising large PE firms on tech acquisitions. It was around that time that people started talking to me about something called Bitcoin. At the time, I thought, "Bitcoin?" I remember working in fintech earlier in my career, but fintech back then was mostly bank software.

I believed Bitcoin could indeed solve many problems, but it wasn't regulated or sanctioned at the time, so I eventually put it out of mind. In 2014-2015, Bitcoin came back into my view. A former CTO I worked with invested everything he owned into Bitcoin. I mean literally everything, every single penny.

He was very anti-government and completely opposed to any form of centralization. He told me, "I'm putting all my money into Bitcoin. That's all I'm going to do next." At the time, he was still developing a lot of software leveraging the Bitcoin blockchain. I deeply respected this person, so I started thinking that maybe this was genuinely interesting.

After that, a series of coincidences occurred. Someone suddenly told me, "We're facing major issues trading Bitcoin; the real bottleneck is getting funds into exchanges." I'm naturally problem-solving oriented, so I started seriously thinking about this. In 2015, I partnered with a group to build an exchange, which you could think of as a wholesale exchange: users deposit funds into one account first, and we then distribute those funds across multiple different exchanges.

This way, since we didn't need to transfer funds frequently, we could trade directly and arbitrage across different exchanges. The issue was obtaining licenses. We spent nearly a year on this, finally securing one in Liechtenstein.

Why Liechtenstein? Because Liechtenstein had no stock exchange, and the local regulators viewed Bitcoin and ETH—there were actually only these two coins worth trading at the time—as foreign exchange currencies. Thus, we obtained an FX trading license there and began operations.

But the issue was that since we held funds for clients, we then needed a banking license. Ultimately, this entire business model couldn't sustain itself. Meanwhile, a friend contacted me saying, "Listen, someone invited me to take over a public company, and we plan to pivot it into a Bitcoin mining company. I know you understand crypto. Would you be willing to join the board and help us make this happen?"

So I joined. That was late 2017 to early 2018. The company was essentially what would later become MARA, then known as Marathon Patent Group, a so-called "patent troll" company. MARA, or Marathon at the time, had a very interesting history. It was initially founded as a vanadium mining company, genuinely engaged in metallurgy and mining operations.

Later, the company dabbled in oil and gas for a while, tried several other directions, and finally started acquiring patents. One patent they acquired was foundational for controlling mobile devices with limited voice command sets. You could think of it as technology similar to Siri or Alexa.

The company later sued Amazon and Apple for patent infringement. Apple ultimately paid a patent licensing fee, but the company then funneled all that money into litigating against Amazon. To this day, that lawsuit remains unresolved. In short, MARA back then was fundamentally a patent troll company. Like many shell-listed companies of that era, it was easy for a group to gain control, and a batch of people indeed did just that.

This group was precisely the team that initially pushed Riot to go public. They saw what they accomplished with Riot and decided to replicate the same model with MARA. They invited my friend to run the company, who then invited me to join the board.

Afterwards, we had to resolve some corporate ownership issues first, including some toxic convertible bonds. The balance sheet also had problems, and untangling those took time. Then, we began Bitcoin mining operations. The initial mining facility was located in Canada, operating as a hosting/tier facility.

There's a great video on our website titled "The Road to 23 EH/s". You can see that facility inside; the paint on the walls was already peeling off. Once the miners powered up, all the fans created massive negative pressure in the room, literally sucking the wall paint down and into the machines. Eventually, that entire batch of miners had to be scrapped.

Those were some memorable early lessons learned. After that, operations experienced some intermittent development, but we eventually got mining running. In April 2021, I left the board and officially became CEO, when I was the company's fifth employee.

From April 2021 to the end of 2023, we spent heavily on fundraising. That was a relatively easy period for mining company financing, especially for publicly listed ones. We raised billions of dollars and placed massive orders with Bitmain during tight hardware supply, allowing the company to expand rapidly.

At the time, we didn't invest in building or owning any mines. Instead, we fully adopted a hosting model, pursuing an asset-light route. This allowed our expansion speed to outpace any other company. Because under the traditional model, you'd invest 20% to 30% of capital into the site itself, leaving the rest for hardware, and building a mine typically takes 12 to 18 months.

Our approach was straightforward: find partners with available capacity and tell them, "We have capital and computing hardware; you have power and space capacity. Let's expand together." By the end of 2023, we had become one of the world's largest Bitcoin mining companies.

Why Must "Mining Companies Control Power"?

From late 2023 to early 2024, we began acquiring mines where we previously hosted miners, purchasing them at prices below replacement cost. By the end of 2024, we owned roughly 70% of our operational infrastructure. After that, we started thinking about directly controlling power, recognizing it would become a scarce resource.

I remember speaking at the Mining Disrupt conference in July 2021. At the time, I made a point that shocked many: you either have to become a power company or partner with one. Because by 2028 to 2032, after several Bitcoin halvenings, you must secure power resources. Electricity is the primary input cost in mining.

Almost everyone mocked that statement back then. But today, power is obviously the most critical resource. This is why we and many peers are transforming: if you use electricity for AI, each electron generates significantly higher revenue than when used for Bitcoin mining.

Nevertheless, we remain one of the world's largest Bitcoin miners and will continue mining. Even in a data center-centric world, Bitcoin mining is still an excellent way to optimize power utilization. In regions with free or low-cost energy, using Bitcoin mining to absorb power presents highly viable application scenarios. There are still many such localized opportunities globally. You just need to be willing to seek them out and capable of partnering with those who control these power assets.

There's an English term called NIMBY—"Not In My Back YARD". When we took over a mine in Granbury, Texas, the very first day we owned the facility, someone stood outside protesting with signs saying "Stop Bitcoin Mining Again". So this isn't anything new.

The new change is that people now believe AI and data centers will drive up electricity prices, consume scarce resources like water, and impact landscapes and noise levels. People are worried about these things. In a sense, fear simply stems from a lack of information. If you already know the final outcome, you can take corresponding measures. But if you have no idea what a data center actually is, and someone just tells you, "It will be loud and cause various problems," of course you'll be scared.

I believe the entire industry is currently dealing with this. Fundamentally, however, it remains an energy and power issue. Building a new power plant typically takes six to seven years. For conventional thermal, natural gas, or nuclear projects, forget it—those could take 20 to 30 years. Through Small Modular Reactors, or SMRs, we might potentially shorten construction cycles to five years in the future.

But if people already won't accept having a data center built near their homes, imagine how they'll oppose a nuclear reactor nearby. Therefore, I believe SMRs are more likely to be built on government-owned land. They might also be operated by Indigenous communities or tribes, who might be more willing to deploy SMRs on their own lands.

Natalie Brunell: Could you explain to viewers unfamiliar with SMRs exactly what they are?

Fred Thiel: Sorry, SMR stands for Small Modular Reactors. The US Navy has used nuclear power to propel its fleet for decades. Traditional commercial nuclear plants are usually custom-designed projects. Each plant is bespoke, requiring specially manufactured components.

But in naval nuclear systems, standardized parts are used. Currently, several commercial companies founded by different teams have secured backing from investors like Bill Gates. Their business model essentially replicates what Ford did to the auto industry, applied to nuclear reactors.

This means adopting assembly line production to cut costs, unifying designs, and applying this model to nuclear reactors. This allows manufacturing compact, water-free cooling nuclear reactors. These units can also consume spent fuel from traditional reactors, eliminating the need to buy more uranium while helping solve the disposal problem of waste fuel rods.

Because they don't use water cooling and incorporate a fail-safe mechanism that automatically shuts down the reactor upon anomaly, they are extremely safe. The US Navy has operated such reactors for decades with very few incidents, because they've designed and optimized the technology to a highly mature and reliable state. I believe SMRs will ultimately become an excellent solution to energy challenges. The main hurdle is that society first needs to accept nuclear energy.

Following the Fukushima disaster in Japan and Three Mile Island in the US, public acceptance of nuclear energy has been a persistent issue. Often, public perception alone dictates everything. But looking at actual conditions in the US, you'll find America has sufficient power to meet AI, residential, and industrial demands simultaneously.

The real issue isn't an overall lack of power, but rather the timing of when that power can be utilized. Duke University published a study last year arguing that total AI power demand is roughly 40 GW, while the US actually has about 72 GW available, provided data centers can curtail load less than 2% of the year.

Consider that 2% of 24 hours is a very short window. Data centers could easily maintain operations using backup generators during those periods. Historically, though, data center operators preferred facilities within 100 miles of so-called "NFL cities"—meaning near major metropolitan areas.

They also aim for "five nines" availability, meaning 99.999% uptime. This requires redundant power and infrastructure. But as usable power slowly depletes, people are beginning to genuinely explore and innovate. They're asking: "How can we operate more like a Bitcoin miner? How can we use energy in a way that proactively sheds load when needed, making data centers a flexible load? How can we utilize surplus power already existing in the grid?"

Remember, building a new natural gas plant takes six to seven years. Even if Google says it wants to build a data center and deploy behind-the-meter generation, it still needs to purchase gas turbines, complete construction, and obtain permits. All of this takes time.

Therefore, immediately available power in the market is extremely limited, yet demand chasing it far exceeds supply. Many don't realize this power constraint affects not just data centers, but every piece of equipment inside them and the entire underlying supply chain.

Why Did the AI Compute Race Ultimately Become a Power Grab?

Fred Thiel: Assume existing power can only support building 10 GW of data centers. Then the market can only absorb 10 GW of compute hardware, because beyond that, devices physically cannot plug into the grid. This caps NVIDIA's potential chip sales. Meanwhile, NVIDIA competes with companies like Google and AMD in chips and compute.

So chipmakers are now saying, "We must lock in power early, or we'll lose market share." A fascinating competition is underway. The contenders grabbing power aren't just hyperscale cloud providers needing compute and data center capacity, but also chip vendors and frontier model companies.

For example, OpenAI and Anthropic compete fiercely. If Anthropic fails to secure enough compute, its token usage prices must rise sharply. Suddenly, open-source models become suddenly attractive. So multiple wars exist right now: at the token cost level, frontier closed-source models versus open-source, open-weight models; plus chip wars and battles among hyperscalers.

Those holding land and power can sit back and say, "I'm in a highly advantageous position." MARA has accumulated substantial related resources over recent years. Currently, we operate about 1.1 GW of power. If we expand existing sites, capacity could grow past 2 GW.

Add to that the Long Ridge deal we announced earlier this year, including a 505 MW gas-fired power plant. Long Ridge also features over 1,600 acres. We can build a massive data center there without stopping power delivery to the grid, since we can continue increasing plant capacity.

We also control a substation, allowing us to tap additional grid power. Additionally, we just announced acquiring a project from HIF. The project boasts up to 2 GW of power capacity and a large campus adjacent to a nuclear plant in Texas. The campus connects to roughly 13 to 15 high-voltage transmission lines, engineered to withstand severe weather and hurricanes.

Originally designed around industrial uses like e-fuels, it's actually an ideal data center site. We're very excited about this. So currently, our controlled power resources exceed 4 GW.

Does Shifting to AI Mean MARA Will Shut Down Bitcoin Miners?

Natalie Brunell: We've recently heard a lot about mining companies pivoting to AI. Does this mean you'll actually shut down Bitcoin miners since profitability shifts to supplying AI companies? And what impact would that have on the Bitcoin network?

Fred Thiel: First, this isn't an overnight transition. Building a data center takes 18 to 24 months. Meanwhile, when structuring relevant deals, we ensure Bitcoin mining can continue until power truly needs to switch to data center operations. We strategically partition sites to allow sustained mining runs.

In some cases, we retain a certain scale of Bitcoin mining capability at a single site, because data center power demand fluctuates based on specific use cases. Are they training models or handling inference? Is it critical IT load? Power curves differ across scenarios.

A key advantage of Bitcoin mining is our containerized deployment model, meaning we can genuinely move equipment from one site to another. Take the HIF project we recently acquired in Texas: it has 2 GW capacity, and local energy prices still support profitable Bitcoin mining.

Theoretically, we could retrofit all current mines into AI data centers, relocate every Bitcoin miner to that site, and keep the entire mining operation profitable. And that doesn't even factor in any future AI business deployed there.

Right now, we operate roughly 1.1 GW of Bitcoin mining load. Even migrating all of it to the HIF project leaves about 900 MW of spare capacity there, while all our existing other assets remain available for conversion into AI infrastructure.

However, as I said, this process takes years. Even with signed tenant leases, completing data center construction still requires 18 to 24 months.

Is AI Draining Capital and Attention from the Bitcoin Market?

Fred Thiel: My hair is gray enough; the early internet build-out phase was when my career was highly active. I worked in data communications—you could call it Ethernet-related fields. When internet infrastructure started expanding, the market seemed to demand virtually infinite routers, Ethernet switches, networking gear, fiber optic cables, and data centers. Yet later, part of that demand never fully materialized.

Today's situation differs because we've seen actual market demand for AI compute actively growing. I believe the true watershed moment that made people realize this was Anthropic launching Claude Cowork, alongside products like OpenClaw demonstrating genuine Agent capabilities.

Around February this year, everyday consumers suddenly realized what AI actually is. Investors started observing this too, recognizing that if demand climbs this high while supply stays this low, someone is definitely going to make a fortune.

Consequently, investors began mapping the entire value chain built around AI. NVIDIA CEO Jensen Huang proposed a fascinating pyramid model. At the bottom sits power; above that, chips and compute—the domain of NVIDIA; followed by infrastructure, including buildings, cooling systems, etc.; and at the very top are various software layers.

Only one factor restricts growth across all upper tiers: power. That's why Bitcoin mining companies suddenly became highly attractive, as they can potentially supply both power assets and data center footprints. Financially, building a Bitcoin mine—including infrastructure and compute hardware—costs roughly $1 million per megawatt total.

But constructing an AI data center, infrastructure-only excluding compute hardware, costs $10 million to $15 million per megawatt depending on size. Plus, compute hardware gets deployed atop that infrastructure later. So it's an exceptionally capital-intensive business.

In the past, Bitcoin miners struggled to get bank credit. Banks viewed Bitcoin as insufficiently regulated, lacking clear intrinsic value, and highly volatile. They also disliked the "post-loan asset takeover" scenario, refusing to hold miners if borrowers defaulted.

This happened exactly during the 2022 market trough. Many lenders ended up seizing Bitcoin mining equipment anyway. But in AI, clients typically carry investment-grade credit. If Google contracts with you, funding actively flows toward you. This usually involves 15-year lease agreements generating recurring revenue, making it considerably lower risk from financial institutions' perspectives.

Private credit firms show strong interest too. So although you need heavy capital, you can simultaneously secure it. This pulls the entire capital market into play. Now look at retail investors. Historically, retail drove valuation surges for many Bitcoin concept stocks.

But retail investors now see: "Bitcoin prices are falling, AI-related assets are rising, so I'll rotate funds there." Consequently, selling pressure emerged across Bitcoin and related equities. Investors first rotated into gold, then shifted toward AI. Now, they're pulling from gold again, further overweighting AI.

I believe SpaceX's IPO sent a valuable warning globally: not all assets appreciate forever. Companies going public only after reaching extreme maturity often harbor large numbers of early investors desperate to exit. These stakeholders hold shares for 8-10 years, naturally seeking profit realization before rotating to the next opportunity rather than holding indefinitely.

Thus, I anticipate significant changes in future IPO markets. Listed companies will increasingly consist of mature enterprises, with noticeably fewer young ventures. Meanwhile, AI sector competition grows fiercer daily, whether at the model layer or compute tier.

Yet one market remains strongly protected for quite some time: power, land, and infrastructure capacity. If we summarize primary constraints into three aspects, they are capital, capacity, and compute. Data centers represent capacity.

If you possess immediately usable power, can build data centers on your own land, and achieve operational readiness by late 2027 or 2028, you occupy a highly favorable position.

Is AI a Bubble? How Long Can the Power Bottleneck Sustain the Industry?

Fred Thiel: I believe certain market segments may indeed harbor bubbles, like memory chips. Most viewers here are probably too young to have experienced the late 1980s to early 1990s PC boom. As PCs genuinely mainstreamed, memory became a critical bottleneck.

Suddenly, every PC required RAM. Every time Intel upgraded processors from 286 to 386, 486, then Pentium, the required memory capacity increased further. Thus, the memory sector consistently endured stark boom-bust cycles. Check Micron's stock chart since the 1980s—it constantly rose, fell, rose, fell.

Why? When market demand outpaces capacity, memory makers prolong enjoying high prices and margins. Eventually, a company builds a new fab, capacity increases, supply catches up to demand, then overshoots. Markets bottom out and crash.

Later, a new tech wave emerges, restarting demand growth. But makers stay disciplined, avoiding rapid expansion because they fear repeating supply gluts and crashes. Yet as demand keeps climbing, they're inevitably forced to build more fabs.

What we see today mirrors this: Korean memory makers reap massive profits, while Chinese enterprises construct fabs at breakneck speed. This essentially signals memory markets will likely enter another downturn within the next 18-24 months.

HDDs and flash storage experienced identical cycles historically. These patterns repeat endlessly. But currently, a limiting factor exists: power. Power constraints cap how large the overall market can expand. Were infrastructure capacity infinitely supplied, the AI market would grow faster, and price wars would trigger earlier and harder.

Precisely because capacity is constrained, the market faces artificial restraint, keeping costs and profits elevated for a period. I believe the power bottleneck will realistically ease only in four to five years.

What Are the True Difficulties in Building an AI Data Center?

Fred Thiel: It depends entirely on your starting point. If you're originally a Bitcoin miner, you already hold land and power, with connections established. That saves two to three years instantly. Next comes permitting. But first, you must find tenants. Without knowing how the project will be built, you can't apply for permits, and designs heavily depend on tenant specs.

Unless you adopt a bare-metal cloud model like IREN: build the DC, deploy GPUs, then rent compute to others. But with more traditional hyperscaler leasing models, tenant confirmation must come first.

The leasing process alone takes ~6 months, involving site inspections, detailed engineering, etc. Only after completion do you submit permit applications. Depending on region, approval ranges from 90 days to two years. Then construction begins, averaging ~18 months. Duration varies by scale and location.

Labor shortages compound challenges. Simultaneously, you must order equipment timely to ensure transformers, PDUs, racks, etc., arrive on schedule. This explains why we chose to enter AI via partners like Starwood rather than solo. Going independently requires hiring experts convincing hyperscalers to sign; engineers designing/building DCs; and teams managing contracting, EPC works, etc.

Post-completion, you must master facility stabilization/commissioning, ensuring hyperscalers accept delivery. Crucially, all this occurs under brutal timelines; missing deadlines triggers severe penalty clauses.

We deliberately delayed market entry believing we lacked clear advantages across all mentioned domains. We sought a partner leveraging relationships with DC owners, operators, and hyperscalers, alongside inherent EPC/construction/design capabilities, to rapidly deploy projects while injecting robust credit endorsement. Starwood perfectly matched our criteria.

Natalie Brunell: I was about to ask if you regret not pivoting earlier. Seeing some public miners possibly start shifting to AI years ago, does not having gone earlier feel like a missed opportunity to you?

Fred Thiel: Hindsight is always easy. Sure, I could say: "If we acted a year earlier, how would stock reaction look?" Perfect timing doesn't exist. But I firmly believe luck is self-created, because luck essentially equals opportunity meeting preparation. Back then, we weren't ready.

I've repeatedly stated publicly: if I approached Google then saying, "I want you as tenants for our upcoming DC," they'd likely show me the door. You're just a Bitcoin miner; you wouldn't know how to build Tier IV DCs, right?

Hence, collaborating with true experts executing this workflow was mandatory. Starwood excels—they've constructed ~7 GW of DCs for Google, Amazon, Microsoft. They know exactly what they're doing, and collaboration proceeds smoothly.

Examining financial structures reveals this model heavily favors us, enabling ultra-light participation. This means avoiding excessive equity issuance to fund projects, thereby preventing severe shareholder dilution.

Concurrently, abundant debt financing is accessible. Even at 80% LTV, the model remains highly cap-efficient. Therefore, from capital efficiency standpoints, it's highly effective, unlocking larger capacity development. So we persist focusing on core competencies: acquiring land/power assets, converting them into maximum-value utilities.

Why Did MARA Sell Approximately 20,000 BTC?

Natalie Brunell: What prompted that initial decision? We frequently observe miners liquidating BTC immediately post-mining. Current operating environments are tough; influenced by BTC pricing, many public/private miners actually mine at losses.

Fred Thiel: We've consistently clarified we aren't a digital asset treasury firm. Holding BTC remains purely a treasury management strategy—a cash alternative. Actually, our first BTC purchase occurred January 2021. We bought directly spot. Balance sheet carried $150M cash, but zero miner supply remained purchasable.

Thus, allocating capital to BTC served as cash preservation, anticipating price appreciation. Looking back, Jan 2021 BTC traded ~$15k, making this undeniably correct. Flip to Q3 last year; I noted publicly during earnings calls that pricing turned overheated, which proved accurate.

Then, we held ~55k BTC. Valued >$100k/unit, balance sheet exposure reached $5-6B. Truly successful strategy.

Yet as BTC declined, we simultaneously carried ~$3B convertible debt. ~$1B matures in 2027 due to investor put options. Our longstanding practice: buying back CBs below par when traded discounts makes financial sense.

Hence, liquidating ~20k BTC serviced ~$1B convertible debt. Given prevailing discounts, transacting ~$80k/BTC proved economically rational. Naturally, it signaled sentiment shift. But shortly after, Michael Saylor executed identical moves.

Does Selling BTC Mean MARA Loses Bullish Conviction?

Fred Thiel: Markets genuinely shifted, evolving rapidly. Honestly recalling election-period interviews, BTC continuously hit ATHs, everyone preparing for $200k-$300k targets. But circumstances abruptly changed. Proliferation of BTC treasuries emerged; now 1.5 to 2 years later, strategies diversified, miners pivoted hard to AI. Even homebound staunch holders occasionally face informational overload.

Fred Thiel: Correct. Gold persists as an asset class across extended epochs. Certain periods witness overwhelming enthusiasm; others languish unloved. Recently, gold enjoyed peak spotlight moments too.

Natalie Brunell: Worst performing asset classes this year were gold and Bitcoin.

Why Hasn't BTC Become a Widely Adopted Payment Currency?

Fred Thiel: Perspective demands extended horizons. As an asset class, BTC faces fundamental challenge: zero yield generation. Value stores solely depend on holder-vs-seller ratios. Essentially driven purely by supply-demand dynamics.

Satoshi's original whitepaper framed BTC as currency. Discourse now rightly focuses on network security budgets, especially amid widespread miner diversification designs originally stipulated transaction fees replacing block subsidies progressively compensating miners post-halvings. Reality diverged significantly.

Natalie Brunell: Agreed. Transaction fees now hover near historical lows, correct?

Fred Thiel: Affirmative. Block space fees severely depressed, approximating mere basis points (~5 bps historically cited). Fundamentally, mining rewards themselves compressed drastically compared to prior blocks. Positive spin: newly minted BTC entering markets reduced. Conversely, market absorption demand weakened concurrently.

Reviewing post-election periods through 2025: spot ETF approvals birthed, digital asset treasuries proliferated. All catalyzed fresh demand streams against static issuance schedules.

Consequently, price appreciation persisted exceeding organic demand fundamentals justify. Justifies Q3 remarks regarding overheating necessitating reversion towards statistical medians. Rational midpoint estimated ~$90k.

Post-extreme rallies typically mandate full correction cycles returning median baselines before resuming uptrends. Process requires patience. Still retains appeal during geopolitical fractures/conflicts offering rapid cross-border liquidity transfer utility.

Regarding AI/tech commercial integration within crypto spheres, consensus leans heavily toward stablecoin settlement dominance counterparties desire explicit exchange rate certainty.

Barter analogies (eggs/milk) highlight subjective valuations contrasting USD/priced goods revealing mutual pricing consensus despite inflationary erosion rates. High frequency thousands/sec magnifies minor fluctuation impacts instantly altering realized values, intensifying margin compression risks accelerating transaction velocity compounds severity proportionately.

Conclusion: path dependency suggests missed optimal maturation windows functioning mediums-of-exchange capacities.

Alternative paradigm positions enduring value-preserving instruments escaping centralized custodial frameworks transferring portability characteristics retaining relevance provided fiat兑换渠道 remain unimpeded maintaining tailored demographic portfolio optimization flexibilities.

MARA lacks ongoing accumulation mandates presently.

Fred Thiel: Confirmed. Mining economics currently generate structural losses reinvesting reserves necessitates external liquidity injections equity dilution avoidance protocols dictate strategic pauses.

Natalie Brunell: Compared quinquennial benchmarks, reduced bullish outlooks warranted contemporary assessments?

Fred Thiel: Negative confidence degradation merely repositioning fiduciary obligations prioritize maximizing stakeholder returns harnessing land/power portfolios converting AI infrastructure yields exponentially surpass legacy mining arithmetic mandates concentration capitalizing emerging avenues.

Stakeholder matrices encompass personnel/institutional/allied ecosystems demanding maximal value extraction long-term projections reference prior conference observations predicting residual electricity waste utilization mainstream adoption phases decentralized micro-generation nodes industrial balancing mechanisms regulating consumption endpoints inherently providing grid stability utility ancillary revenues supplement incentive structures marginal cost zero-operation efficiencies guarantee sustained hashrate continuity cryptographic security guarantees persist conditional ongoing community codebase maintenance avoiding technological ossification trajectories.

Commissioning sequences traditionally employ dummy loads dissipating energized capacity substituting ASIC clusters eliminates expenditure differential establishing economical zero-marginal-cost operational paradigms naturally migrating workloads optimizing geographic/temporal alignments securing continuous network protection post-transition queries reduce exclusively determining collective governance willingness sustaining programming evolution resisting architectural rigidity enforcement.

Can BTC Evolve Into Multi-Trillion Dollar Asset Classes?

Natalie Brunell: Narratives suggest bounded proliferation restricting categorizations contrasting hyper-targeted valuations ($20T/$50T/$100T scenarios).

Fred Thiel: Projecting establishment reaching approximate $2T classification constitutes monumental achievement already ranking amongst global top-ten asset strata.

Maturation timelines dictate organic growth trajectories. Spot ETF vehicles compressed twenty-year gold-equivalent expansion metrics into twelve-month implementations establishing resilient foundations requires institutional bedrock formation progressing incrementally.

Generational divides manifest traditional financial encumbrances constraining behavioral adaptations juxtaposed digitally native cohorts initializing early exposure accumulating compounded familiarity families/maturation processes redirect allocations towards index instrumentation diversified portfolio constructions spanning multiple stratifications.

Progeny generations inherit normalized participation frameworks treating allocations uniformly absent revolutionary fervor characterizing foundational architects generational transmissions diminish ideological intensity proportions.

Natalie Brunell: Motivations driving repositioning parameters clarify chronological developments disrupting expected pathways obstructing universal adoption trajectories displacing incumbent monetary architectures institutional barriers preventing systemic integration mechanisms.

Fred Thiel: Monetary sovereignty resides consolidated banking/government apparatuses hypothetical sovereign reserve recognition regulatory easing facilitating balance sheet incorporations interbank clearing permissions utilizing cryptographic verification protocols revolutionizes settlement architectures fundamentally mismatched micropayment frequencies contrast macro-level bilateral settlements international balance adjustments demonstrate suitability advantages entrenched incumbents blocking formal assimilation processes prevent operational embedding institutional compliance requirements dictate exclusionary policies.

Natalie Brunell: Skepticism regarding eventual paradigm shifts ignoring evolutionary progression probabilities speculative reserves implementations evidence chain analytics suggesting preemptive accumulation initiatives intelligence gathering activities.

Fred Thiel: Institutional experimentation programs actively piloting excess capacity monetization methodologies transforming theoretical innovations scalable deployments bridging execution gaps remaining paramount objectives sustained demand creation buyer-supplier asymmetry appreciation mechanics positive feedback loops reinforcing perceived scarcity hedging inflationary depreciation maintaining purchasing power parity correlations analogous precious metals cyclical seasonality fluctuations characterize current market phases resembling autumnal/winter transitions anticipating springtime recoveries mirroring historical precedent recurrence patterns.

Five-Year Horizon: MARA Development Roadmap?

Fred Thiel: Continuous partial engagement anticipated subsequent halvings biennial/quasi-decadal intervals reducing block subsidies below unity thresholds elevating operational expenditures disproportionately favoring alternative utilization vectors integrating computational modules HVAC systems embedded appliances harvesting wasted energy gradients solar PV arrays ubiquitous microgrid decentralization consumer empowerment direct generation utilization west coast experiences illuminate distributed renewable integration complexities utility tariff structures malfunctioning theoretical grid optimization strategies employing photovoltaic installations paired battery storage architectures dispatching stored kinetic/resistive elements mitigating oscillations filling peaks valleys saturation points triggering redundant generation cascades incentivizing cryptographic proof-of-work algorithms absorbing otherwise stranded potentials eliminate opportunity costs zero-marginal-exploitation efficiencies democratizing accessibility removing economic barriers fostering ubiquitous environmental harmonization.

Early-stage minority stakes injected residential rooftop deployments coupled miniaturized hashing rigs autonomously diverting surplus outputs bypassing domestic consumption/charging requirements capitalized intensive upfront capital expenditures ultimately failed viability validation innovative concepts retained intellectual merit demonstrating foundational principles feasibility.

Predictions emphasize inevitable ASIC circuitry miniaturization homogenizing diverse appliance categories simplifying electronic architectures isolating power management complexities diminishing relevance non-industrial scaling contexts creating abundant harvesting possibilities utilizing waste streams rendering cost considerations irrelevant exploiting gratis atmospheric thermodynamic differentials.

Natalie Brunell: Acknowledged paradigm inversion remarkable condensed timeframe witnessing rhetorical displacement apex narratives evaporating completely redirected attention allocations concerning computational density ecological thermal burdens planetary warming apprehensions muted comparative apathy normalizing ubiquity masking concerns.

Fred Thiel: Historical parallels recur cyberpunk dystopian warnings dismissing utility potential demonizing content classifications obscuring transformative benefits until irreversible dependency establishes normalization processes mirror evolutionary adaptation trajectories cryptographic networks undergo parallel metamorphoses commoditizing utility embedding seamlessly diverse portfolios hedging inflationary pressures preserving real estate purchasing power maintaining equilibrium positioning beneficial outcomes intrinsically aligned societal progress trajectories.

Natalie Brunell: Valuation ceilings hypothesized tipping points signaling exhaustion thresholds?

Fred Thiel: Realize pricing dynamics bifurcate value equation denominators assume aggressive fiat debasement eroding nominal wealth twenty percent annually equivalent absolute appreciation neutralizes purchasing parity modifications question fundamentally addresses broader monetary ecosystem valuation shifts commodity revaluation phenomena manifest collectibles/hard assets appreciating dramatically scarcity replication difficulties compounding cyclic reinforcement patterns precious metals cryptos function substitution hedges inflationary regimes navigating generational cohort preferences rejecting physical custody complexities preferring digital programmability accepting volatility premium excitement factor correlation amplitude inversely proportional comfort zones.

Quantum Computing: Genuine Threat to Bitcoin?

Natalie Brunell: Perspectives regarding BIP-110 proposition fracturing community governance discourse minimal miner endorsement percentages (<1%) subjective evaluations?

Fred Thiel: Foundation positions articulated separately deem debates counterproductive sector development quantum vulnerabilities wallet architectures substantiated imminent timelines approximately five-decade projection horizons distinguishing ledger integrity concerns versus client-side exposure limitations selectively affecting subset functionalities proposing mitigation strategies implementing temporal delays triggering specific transaction classes circumventing centralized coordination mandates conflicting foundational decentralization ethos personal opinion representing individual stance exclusively advises migration early legacy wallets generating fresh addresses preserving public key obfuscation mitigating cryptographic compromise risks forgotten passphrase scenarios rendering dormant holdings immune unauthorized access attempts acknowledging misfortune attribution hypothetical scenarios involving genesis address activity speculation activation claims contrasting concerns prioritizing secondary effects market reactions precipitated sudden movement occurrences.

Natalie Brunell: Movements implying cryptanalytic breaches necessarily?

Fred Thiel: Assumption invalidates alternative motivations autonomous redistribution intentions authorial agency determinations unspecified.

Natalie Brunell: Uncertainty permeates plausible custody configurations contemporary existence hypotheses undocumented.

Fred Thiel: Affirmative archival research collaborations documentary productions unveiling obscure chronological events forecasting intriguing releases illuminating formative chapters deciphering email correspondence revelations posteriori analyses indicate inaccessible wallet controls fragmented private key distributions malfunctions precluding restoration impossibility assertions.

Natalie Brunell: Probability acknowledged ultimate disposition resolutions mandated?

Fred Thiel: Analogy drawn legal proceedings Craig Wright controversies establishing possession proofs demonstrating administrative authority secrecy motivations deterring disclosure inquiries transcending cryptographic discussions referencing past executive tenures chairmanship prominent European encryption hardware manufacturers supplying HSM modules NATO/corporate/banking sectors deploying stringent security protocols collaborative engagements NIST evaluating post-quantum algorithmic alternatives superseding RSA architectures conceptualizing asymmetric cryptographic frameworks computational asymmetries encrypting transmitting storing protecting sensitive banking/email/web authentication credentials valuing compromised datasets exceeding genesis coin estimates intercepting SSL encrypted packets logging ISP traffic decrypting subsequently extracting plaintext credentials circumventing brute force methodologies compromising accounts universally accessing financial/securities portals stealthily siphoning marginal amounts delaying detection durations aggregating cumulative theft volumes vastly exceeding conspicuous single-transaction withdrawals broadcasting global awareness universally distinguishing targeting priorities national entity supremacy assuming offensive capabilities hypothetical possession immediate capital appropriation strategies Bitcoin last priority conversions facilitating laundering operations identifying superior exploitation vectors illustrating discernible indicators monitoring vendor certificate upgrades Microsoft/Google transitioning quantum-resistant standards signaling imminent existential threats validating preparedness escalations.

Natalie Brunell: Fascinating mortality dispositions claiming elimination boosting deflationary mechanics misunderstanding upgrade necessities relinquishing custodial control transferring vulnerabilities adversaries retroactively invalidated.

Fred Thiel: Convergence plausible hardware enclave isolation employing fresh addressing methodologies preserving comprehensive security profiles.

Natalie Brunell: Posthumous emergence preventing transitional migrations hypothetically destroyed assets potentially recovered malicious actors confirming vulnerability exposures?

Fred Thiel: Inquiry targets multi-billionaire demographics acquiring substantial holdings utilizing sophisticated estate planning instruments converting digital balances exchange-traded products IBIT exempting taxation designations collateralization possibilities familial trusts inheriting complexities navigating executor administration procedures traditional finance integration supporting underlying cryptocurrency preservations legitimizing ownership structures demonstrating prudent foresight anticipatory governance solutions.

Natalie Brunell: Candidly evaluating preceding assessments projecting bearish undertones apparent?

Fred Thiel: Negation reaffirms conviction cornerstone personal portfolio unwavering belief cyclical volatility acceptance pride participating maturation trajectories organic evolution passion-driven retail participation transitioning institutional/rational investor coexistence longevity affirming continued existence adjusting inflation-adjusted valuations establishing sustainable equilibriums dismissing lunar speculation narratives contingent macroeconomic catastrophes capital flight inductions positioning anti-inflationary hedge utility portfolio constituents inherently positive developmental advancements characterizing evolutionary milestones.