1. Podcasts
  2. Fintech Stacks
  3. 🏗️ BofA Builds a $1/4T War Chest. Riot Mines a $9 Billion Deal With Anthropic. BitGo Beats the Top Line.
Fintech Stacks

13 Aug 2026 · 19m · 11 chapters

Episode artwork

🏗️ BofA Builds a $1/4T War Chest. Riot Mines a $9 Billion Deal With Anthropic. BitGo Beats the Top Line.

In short

Using mid-August Q2 earnings as a “front office” lens, the episode argues that tech/crypto winners are building structural leverage in physical infrastructure and durable revenue, not just chasing headline growth.

Guests

No guest names or bios are provided in the transcript; it’s a two-host discussion.

Key claims

  1. Bank of America’s $250B “critical infrastructure finance initiative” (data centers, energy, semiconductors through July 2027) signals AI’s bottleneck is power/buildout, not software.
  2. Riot Platforms’ $174.2M revenue beat and $237M net loss are reframed as a pivot: a 20-year $9.1B lease with Anthropic for 191MW, plus AMD deal for 241MW total AI capacity.
  3. BitGo’s $4.3B revenue but ~$42.5M gross profit is explained by trading-volume accounting: top line reflects notional volume; fees are basis-point tolls, while security/compliance overhead crushes margins.

Notable examples

NBA “quiet window” analogy; impairment charges from obsolete mining rigs; comparison to Coinbase’s early profitability struggle.

Written by AI. May contain mistakes.

Chapters

0:41

Analyzing Bank of America's Initiative

Discover Bank of America's unprecedented $250 billion initiative in infrastructure.

“And, well, welcome to today's Deep Dive, because it turns out the exact same phenomenon is happening right now in the financial and tech sectors.”
1:56

Understanding the AI Infrastructure Needs

Examine how the AI industry's infrastructure demands are reshaping financial investments.

“And that means looking at Bank of America, ticker symbol BAC.”
5:06

Riot's Transition to AI Infrastructure

Explore how Riot is pivoting from Bitcoin mining to AI infrastructure with a major lease deal.

“But wait, if a legacy bank is committing a quarter of a trillion dollars to this, does this mean they view the AI build out not as a speculative bubble, but like as the undeniable center of the next decade of growth?”
10:54

BitGo's Revenue Insights

Analyze BitGo's staggering revenue and the implications of their operating model.

“But, you know, Riot's success here comes from stepping outside the pure digital asset ecosystem to service AI.”
14:03

The Costs of Institutional Crypto Operations

Explore the significant costs and challenges faced by crypto institutions.

“You are constantly paying for elite cybersecurity teams, redundant cold storage facilities, intense regulatory compliance across dozens of different jurisdictions, and continuous legal counsel.”
14:37

Lessons from Coinbase's IPO Struggles

Learn how Coinbase's early challenges mirror those of BitGo.

“It is the identical evolution that Coinbase went through when they first debuted on the public market.”
15:06

BitGo's Growth Pains and Investor Relations

Discuss the challenges BitGo faces in communicating with investors.

“It's a tough conversation to have with investors.”
15:27

Understanding Structural Leverage in Crypto

Identify who holds the structural leverage in the crypto sector.

“But their bottom line is quietly whispering that the business model of securing it hasn't figured out how to actually keep the profits.”
15:54

Physical Infrastructure vs. Digital Success

Analyze the importance of physical infrastructure for digital firms.

“Right, because BitGo proves that institutional crypto adoption is undeniable.”
16:26

Bank of America's $250 Billion War Chest

Discover Bank of America's strategy to secure physical infrastructure.

“Meanwhile, the entities focused on physical reality are locking up the board.”
Show all 11 chapters
17:11

The Limitations of Power Grids on Tech Development

Consider the implications of power grid limitations on tech ecosystems.

“It is a remarkable restructuring of where global capital is flowing.”
A man walking through an airport, asking his phone a question

Ask about this episode

Get the gist, explore an idea or ask about something said.

Claude

On Fintech Stacks, what are the key points of this episode?

Searching your listening

Transcript

Automatic transcript. May contain errors.

0:00You know, if you follow the NBA, there is this very specific, very quiet window. in mid-August, the free agency frenzy is just completely over. Right. Yeah. The blockbuster trades have all settled down by then. Exactly. And the draft is, you know, firmly in the rearview mirror. But if you look really closely, this is the exact moment when the smartest front offices are quietly restructuring their rosters. Because the casual fans have totally tuned out. Yeah, they really have. But the actual architects of the game. Yeah. They are making these foundational, under-the-radar moves that basically dictate who wins a championship, you know, three years from now.

0:39It's all about the long game, right? Right. And, well, welcome to today's Deep Dive, because it turns out the exact same phenomenon is happening right now in the financial and tech sectors. It really is. We have this fascinating dispatch today from a fintech newsletter, and it's covering the mid-August Q2 earnings. So we are going to look at these numbers with you through that exact same front office lens. Which I think is a great way to frame it. Thanks. Yeah, the mission here is to look way beyond the surface level earnings headlines, you know, to figure out which of these companies are actually building long term structural depth.

1:13Versus the ones that are frankly just running it back and hoping the market stays friendly. Exactly. Because it is, well, it's a critical exercise, right? Oh, absolutely. Because surface numbers in tech and finance, I mean, they are notoriously deceptive. Right. A company can post a staggering revenue figure and still possess a fundamentally broken internal architecture. Or, you know, inversely, they can post a heavy quarterly loss but be laying the groundwork for a multi-decade monopoly. So it all comes down to the underlying mechanics of where the capital is actually flowing. Precisely. You have to follow the money, not just the headlines.

1:49OK, let's unpack this, because if we're looking for foundational moves, we really have to start at the absolute bedrock of the economy. The capital layer. Yeah, the capital layer. And that means looking at Bank of America, ticker symbol BAC. Their Q2 announcement is just anything but traditional. No, it's pretty unprecedented. They just announced a$250 billion critical infrastructure finance initiative. Which is just a massive number. It's a quarter of a trillion dollars running through July 2027. And it's aimed specifically at U.S. data centers, energy and semiconductors. Yeah. And when a financial institution operates at that scale, allocating a quarter trillion dollars, that is a monumental pivot.

2:32Right. Because Bank of America, I mean, they're headquartered in Charlotte. They were founded back in, what, 1904? Yep. 1904. This is a 120-year-old institution. They have weathered world wars, global depressions, you name it. So they don't just throw money around. Exactly. They do not allocate that kind of liquidity on a whim or, you know, just to chase a fleeting tech trend. So I have to ask, why them? I mean, we usually think of legacy banks underwriting traditional corporate mergers or like suburban mortgages. Right. The safe bets. Yeah. Why is a century old bank stepping in to underwrite the physical foundation of artificial intelligence?

3:08Yeah. Instead of, say, a consortium of Silicon Valley venture capitalists? Well, because the venture capitalists are starting to hit a wall. Really? A wall? Yeah. Over the past two years, Bank of America has been quietly observing a severe bottleneck in the market. They are watching hyperscalers, you know, the colossal cloud providers and AI labs just devour capital at an unprecedented rate. Right. But the realization is that these tech giants are exhausting, that capital faster than physical utilities can actually build the power plants required to run the servers. Oh, wow. Because AI requires an entirely different scale of electricity.

3:45Precisely. The AI revolution has evolved past being purely a software endeavor. It is now, for all intents and purposes, a heavy industry. Right. It's not just code anymore. No, it requires copper. It requires millions of tons of concrete, intricate liquid cooling systems and, like, entirely new power grids. And VCs aren't set up for that. Not at all. The venture capital world is built to fund software, which scales infinitely and, relatively speaking, cheaply. They are not built to fund utility-scale infrastructure. So Bank of America sees the gap. Exactly. By stepping in with this initiative, Bank of America is positioning itself as the literal financing backbone for the next tech revolution.

4:26They aren't trying to build the AI models. They are funding the physical reality that allows those models to exist. You hit the nail on the head. That perfectly mirrors our NBA analogy, actually. This is the equivalent of a team drafting strictly for need instead of just, you know, picking the most hyped player available. Right. They aren't going for the flashy rookie. No, Bank of America isn't trying to draft the flashy point guard, like the hot new AI software startup. They are loading up on the seven foot centers who could just block shots and grab rebounds. The foundational pieces. Yeah, they are hoarding the unglamorous infrastructure picks that every other company is going to be scrambling for.

5:05It's a really smart play. But wait, if a legacy bank is committing a quarter of a trillion dollars to this, does this mean they view the AI build out not as a speculative bubble, but like as the undeniable center of the next decade of growth? What's fascinating here is that a commitment of this size is a definitive declaration. How so? Well, when a bank of the sheer magnitude commits that level of capital to a unified theme, it is no longer a hedge. A hedge is, you know, a few billion dollars scattered across emerging markets just in case. Right. Just testing the waters. Exactly. A quarter trillion dollars is a planted flag.

5:42They are signaling to you, to the broader market and to their competitors, that the next decade of macroeconomic growth sits squarely in domestic semiconductors, energy generation, and data centers. That makes total sense. Yeah. But Bank of America is only providing the checkbook, right? I mean, writing a check doesn't magically turn the lights on. No, it certainly doesn't. someone actually has to secure the physical megawatt capacity on the ground to build these facilities. And that reality leads to one of the most bizarre earnings reports I have seen in a really long time. I think I know where you're going with this.

6:16Yeah. Let's look at Riot platforms, ticker symbol ROT. The pure play Bitcoin miner. Supposedly. Supposedly a pure play miner. But let's look at their Q2 numbers because the math is wild. It really is a fascinating case study. So on the top line, they posted a revenue beat of$174.2 million. That is a 14 % year-over-year jump, comfortably beating analyst estimates of, I think,$152 million. Right, which looks great on paper. Exactly. If you just read the headline, you assume they had a phenomenal quarter. But then you look at the bottom line, and they posted this brutal$237 million net loss. Ouch.

6:55Yeah. And a chunk of that was driven by$28 million in impairment charges. Right. So let's break down that impairment charge first because it is crucial context here. Please do because it sounds bad. Well, in the mining industry, an impairment charge usually means an asset on your balance sheet has permanently lost value. For a company like Riot, that often refers to older generations of Bitcoin mining rigs. Oh, I see. Because the tech moves so fast. Exactly. As the global network becomes more competitive and the algorithms become harder to solve, older machines just become mathematically obsolete.

7:27Like trying to run modern software on a computer from 2010? Basically, yeah. You spend more on electricity to run them than you actually earn back in Bitcoin. So you have to write their value down to zero on your books, which obviously hits your net income hard. OK, so they beat on revenue. But between operating costs and writing off all this dead equipment, they bled out$237 million. Correct. For almost any publicly traded company bleeding out more money than you actually brought in, would send investors, like, sprinting for the exits. You would think so. But here's where it gets really interesting.

8:00Despite that heavy net loss, Riot's stock jumped. At one point, it was up 22%. Which defies traditional logic, right? Totally. Why on earth is the market rewarding a company that just lost nearly a quarter of a billion dollars? Because the market recognized a fundamental transformation in their corporate architecture. What did they do? Riot announced they signed a 20-year,$9.1 billion lease with Anthropic. Right. $9.1 billion. Yep. Riot is going to provide 191 megawatts of capacity at their Rockdale, Texas campus to physically house Anthropic's AI operations. Oh, wow. And when you combine that with an existing deal they have with AMD, Riot now has 241 megawatts of total capacity dedicated strictly to AI infrastructure.

8:48I am stuck on something here, though. Riot has spent years cultivating their entire corporate identity around being a pure play Bitcoin miner, right? They absolutely have. They sold investors on the premise that they mine Bitcoin at an industrial scale. So by carving out 241 megawatts for AI companies, aren't they basically abandoning the core thesis that their shareholders bought into? It looks like abandonment on the surface, I'll give you that. But mathematically, it is an essential structural hedge. How so? You have to consider the mechanics of the global hash rate. In Bitcoin mining, your revenue is intrinsically tied to two highly unstable variables, the daily price of Bitcoin and the global computational competition, which is known as the hash rate.

9:31OK, so a lot of moving parts. Right. If the price of Bitcoin drops by 4 percent or if a competitor brings a million new machines online and makes the puzzle harder to solve, your profit margins can just evaporate overnight. That sounds incredibly stressful to manage. It is. So Riot is choosing the exact same survival instinct as some of their peers, companies like MRI and CleanSpark. So they are securing a reliable stream of fiat currency to offset the wild swings of the crypto market. Exactly that. A 20-year lease with Anthropic provides an incredibly predictable, multi-billion dollar bedrock of traditional revenue.

10:10Which gives them breathing room. Right. It guarantees that a sudden crypto bear market won't just sink the entire company. They didn't just beat their top line earnings, they fundamentally altered their DNA. That's a great way to put it. They have transitioned from being exclusively a Bitcoin miner into a diversified data center infrastructure company that still utilizes its excess power to mine Bitcoin. That makes total sense. They realized their primary asset isn't actually the Bitcoin they produce. Their primary asset is their access to industrial scale electricity. Bingo. And right now, AI labs are obviously willing to pay a much higher premium for that electricity than the Bitcoin network is.

10:46Far ahead. So the market saw a volatile crypto company suddenly lock in a$9 billion guarantee, and they rewarded the adaptability. It's all about securing the future. But, you know, Riot's success here comes from stepping outside the pure digital asset ecosystem to service AI. What happens to a company that refuses to pivot, like a company that stays entirely entrenched in the crypto plumbing? Well, that brings us to a much harsher reality check regarding financial mechanics. And that is our third case study today, BitGo, ticker symbol BTGO. Right, the custodians. Yeah. They IPO back in January at$18 a share.

11:25They are a digital asset custody and infrastructure company, and their Q2 numbers are, frankly, mind-boggling, but for an entirely different reason. The numbers are pretty jarring. On the top line, they posted$4.3 billion in Q2 revenue, driven almost entirely by digital asset trading volume and institutional flow. And$4.3 billion in a single quarter implies a tremendous amount of institutional capital utilizing their rails. It does. But I am looking at these numbers and my brain is just stalling on the math. They have$4.3 billion in revenue, but their gross profit only rose to$42.5 million. It's a huge drop off.

12:03It's essentially a 1 % gross margin. Yeah. And from there, I mean, it only gets worse. Their operating loss came in at$17.4 million and their overall net loss was$19.0 million. Right. Their diluted earnings per share was negative 16 cents, which more than tripled the negative 5 cents that analysts were expecting. So my question is, how do you generate over$4 billion in institutional revenue and wind up with a$42 million gross profit? It's the classic rookie basketball analogy you mentioned earlier. Yeah, scoring 40 points but fouling out of the game. Exactly. But if we connect this to the bigger picture, this is a classic misinterpretation of how revenue is booked in institutional trading in custody.

12:44Okay, clarify that for me. So in traditional software, if you sell a license for$100, that$100 is your revenue. But in financial exchanges and custody platforms, the revenue figure often represents the total notional volume of the assets moving through the platform. Oh, I see. So if an institution moves a billion dollars of Bitcoin through BitGo's infrastructure, BitGo might record a significant portion of that volume as top line revenue depending on their accounting structures. They don't get to keep that billion dollars. Right. The actual fee they capture for facilitating that trade, you know, the spread or the custody fee is only a fraction of a single basis point.

13:22Ah, OK. So the$4.3 billion is just the water flowing through the pipes. But the$42.5 million in gross profit is the tiny toll they actually collect for letting the water pass. Precisely. And that tiny toll has to cover an exorbitant amount of overhead. The reality of the crypto adjacent thesis is that building the plumbing for digital assets is just staggeringly expensive. Break down that thesis for me. Why is the overhead so high if it's all just digital? Because securing billions of dollars in digital bearer assets requires military grade operational expenditure. It's not just a password on the server.

13:58No, not at all. Your gross profit isn't just sitting in a bank account. It is immediately eaten up by the operating expenses. You are constantly paying for elite cybersecurity teams, redundant cold storage facilities, intense regulatory compliance across dozens of different jurisdictions, and continuous legal counsel. That adds up fast. It does. The top line volume scales instantly when institutional money flows in, but profitability severely lags behind because the security and compliance costs scale almost in tandem. So they're basically trapped in a scenario where they have to prioritize gross over margins just to keep the platform secure and legal.

14:37It is the identical evolution that Coinbase went through when they first debuted on the public market. Oh, right. They had a rough start, too. They posted explosive trading volume but struggled for years to find consistent quarterly profitability because the infrastructure overhead was so punishing. But eventually, they figured it out. They did. But the difference here is that BigGo is experiencing that entire painful learning curve compressed into their first two quarters as a newly public entity. I mean, they IPO'd in January and by August, they are sitting in earnings calls trying to explain to shareholders why facilitating billions of dollars in trades still left them$19 million in the hole.

15:18It's a tough conversation to have with investors. Totally. Their top line is screaming from the rooftops that institutional demand for crypto is very real. But their bottom line is quietly whispering that the business model of securing it hasn't figured out how to actually keep the profits. It perfectly illustrates the dichotomy we've been tracking today. There is a vast difference between simply participating in a booming sector and owning the structural foundation beneath it. So what does this all mean? We started this deep dive talking about how the smartest front offices use the quiet periods to build foundational depths while everyone else is distracted by flashy statistics.

15:54For you, the listener, whether you are walking into a corporate strategy meeting this week, evaluating your own portfolio, or just trying to navigate this incredibly complex tech landscape, the lesson here is all about identifying who holds the structural leverage. Exactly. Who actually owns the board. Right, because BitGo proves that institutional crypto adoption is undeniable. I mean, you don't accidentally process$4 billion in trading volume. No, you do not. But high volume does not equal a sustainable, profitable engine. They are racking up impressive stats, but they haven't figured out how to secure the win.

16:30Meanwhile, the entities focused on physical reality are locking up the board. Exactly. Bank of America, an institution that literally predates the invention of the microchip, recognize that the true bottleneck of artificial intelligence isn't software. It is physical infrastructure. They saw the writing on the wall. They did. They are deploying a$250 billion war chest to own the physical reality of the next decade. And Riot Platforms, a company that literally built its name on the extreme volatility of digital currency mining, secured its survival by pivoting hard into the physical world. Landing that$9.1 billion trade partner in Anthropic to house servers.

17:10Yeah, they are the adaptable players who realize that without the hardware and the megawatts, the digital world just doesn't exist. It is a remarkable restructuring of where global capital is flowing. But if we pull back even further, looking at both the ambition of the banks and the pivot of the miners, this raises an important question. What's that? Well, it's something for you to deeply consider as you watch these markets evolve over the coming months. If legacy institutions are emptying their vaults to fund the physical AI build out and former crypto purists are rewriting their corporate identities to physically house these supercomputers.

17:42Yeah. What happens to this entire technological ecosystem if the physical U.S. power grid simply hits its absolute finite limit before these 20 year leases even mature? Oh, wow. That is the ultimate invisible ceiling, isn't it? It really is. You can have a quarter of a trillion dollars in capital, and you can secure the most lucrative long-term lease in corporate history. But if the local grid literally cannot handle the megawatt draw required to cool the servers, the whole system just grinds to a halt. The physical limit is the final boss. It's like you can build a championship roster, but if the arena's lights won't turn on, you still can't play the game.

18:22Exactly. Thank you so much for joining us on this deep dive. As always, keep questioning the underlying mechanics behind the numbers and pay close attention to who is quietly building depth while everyone else is distracted. We will catch you next time.

From the publisher

Bank of America just committed a quarter-trillion dollars to the AI infrastructure race, Riot Platforms turned a bitcoin mining campus into a $9 billion Anthropic data center deal, and BitGo's revenue exploded while its losses widened. We break down what each move signals for where fintech capital is actually flowing in the back half of 2026.


More from Fintech Stacks

All episodes