Per Capita Profit of $85 Million, Outshining Goldman Sachs and Nvidia: The World's Most Profitable Business Isn't AI

marsbitPublished on 2026-02-25Last updated on 2026-02-25

Abstract

An article titled "Per Capita Profit of $85 Million, Surpassing Goldman Sachs and Nvidia: The World's Most Profitable Business Isn't AI" discusses the remarkable financial performance of Tether, the company behind the USDT stablecoin. In 2024, Tether reported a net profit of $13 billion with only about 150 employees, resulting in a staggering per capita profit of approximately $85.62 million. This figure is nearly 300 times that of Goldman Sachs and 85 times that of Nvidia. Tether's highly profitable business model is described as a "stablecoin float game." Users exchange $1 for 1 USDT, and Tether uses this capital to purchase U.S. Treasury bonds, which yield over 5% annually. Since USDT pays no interest to its holders, Tether captures the entire spread. By the end of 2025, Tether's U.S. Treasury holdings reached $141 billion, making it the 17th largest holder globally. It also holds significant assets in gold and Bitcoin, which have contributed substantial additional gains. The article positions Tether as a global "shadow bank" that operates 24/7, offering a crucial service, especially in emerging markets with high inflation and capital controls. It highlights the inefficiencies of the traditional SWIFT payment system, which can take days and charge high fees, compared to USDT transfers, which are nearly instantaneous and cost very little on networks like Tron. The concept of "Pay-Fi" (Payment Finance) is introduced, where next-generation financial protocols enable value...

By | Clow

Author:白话区块链

In 2024, a company named Tether presented a performance report that left Wall Street astounded.

$13 billion in net profit, with approximately 150 employees.

Per capita profit is about $85.62 million, nearly 300 times that of Goldman Sachs and 85 times that of Nvidia.

This is not an AI unicorn, nor a top-tier hedge fund. It is merely a stablecoin issuer—the company behind USDT.

When these numbers circulated in financial circles, many people's first reaction was: How is this possible?

But if you understand Tether's business model, you'll realize that this is not only possible but almost inevitable.

01. The World's Most Profitable Business

Tether's profit logic is known in the industry as the "stablecoin float game."

The rules are simple: You give Tether $1 and receive 1 USDT in return. Tether takes your money and buys U.S. Treasury bonds.

The annualized yield of U.S. bonds has long remained above 5%, while USDT never pays any interest.

The difference in between goes entirely to Tether.

As of the end of 2025, Tether's total exposure to U.S. bonds reached $141 billion, making it the 17th largest holder of U.S. bonds globally, surpassing the scale of sovereign nations like Germany and South Korea.

U.S. bonds alone contribute over $4 billion in cash flow to Tether annually.

And this is just the first layer.

The second layer is gold and Bitcoin. Tether holds approximately $17 billion worth of gold and over 96,000 Bitcoins. The sharp rise in gold prices in 2025 brought an additional $5 billion in floating profits.

The third layer is the liquidity premium. What do those who forgo the 5% bond interest get in return? It is the ability to use digital dollars anytime in Turkey, Argentina, Nigeria, and other countries. For markets plagued by high inflation and foreign exchange controls, this liquidity is more valuable than a 5% annual return.

Tether is essentially a global "shadow bank" without branches, tellers, or downtime, specifically capturing the huge interest rate differentials left behind by the inefficiencies of the traditional financial system.

02. Breaking Through the Walls of Traditional Payments

The SWIFT system was built in the 1970s. Half a century later, its core logic remains largely unchanged: relay through correspondent banks, passing through multiple nodes one after another, taking at least 3 to 5 business days, with comprehensive fees as high as 7%.

A payment from the U.S. to Nigeria must go through layers of remitting banks, intermediary banks, and receiving banks, each taking a cut in handling fees.

Moreover, these banks have business hours. A remittance initiated on Friday night won't start "processing" until the following Monday.

In contrast, a USDT transfer can reach the recipient's wallet in less than 30 seconds on the Tron network for a fee of under $1, operating 24/7, 365 days a year without interruption.

The cost difference is particularly staggering. Traditional comprehensive fees for B2B cross-border payments range from 1.5% to 7%, while personal remittances can sometimes exceed 11%; the comprehensive cost on stablecoin networks is typically only 0.5% to 2%.

The deeper impact lies in "reach."

Hundreds of millions of adults worldwide still lack bank accounts. But with a mobile phone and internet access, they can create a crypto wallet and connect to global trade. In Africa and Latin America, USDT has become a common tool for small and medium-sized enterprises to pay international suppliers.

In 2025, a new generation of Web3 POS systems began using NFC technology to enable "tap-to-pay," bringing crypto payments to retail store counters.

This wall is being breached from all directions.

03. Pay-Fi: The Logic of Money Is Being Rewritten

Payment + Finance, this combination has a new name: Pay-Fi (Payment Finance).

Traditional payments solve the problem of "moving money from A to B." Pay-Fi aims to solve the problem of "money moving from A to B while generating interest along the way."

Protocols like Huma Finance are working on tokenizing corporate accounts receivable, providing instant financing through on-chain liquidity pools to address prepaid capital pressures in cross-border trade. As of early 2026, the cumulative transaction volume of the Huma protocol has exceeded $10 billion, and its T+0 real-time settlement capability is attracting increasing attention from traditional financial institutions.

The battle for underlying infrastructure is also underway. Ethereum L2 significantly reduces on-chain transaction costs through Rollup technology; Celestia and EigenDA further lower fees at the data storage layer, making large-scale micro-payments possible. The Tron network, with its vast USDT reserves and extremely low transfer fees, remains the world's busiest stablecoin settlement network.

The stablecoin market itself is also diversifying. USDT dominates offshore payments and emerging markets with about 59% market share; USDC wins over U.S. licensed institutions with compliance and transparency, occupying the vast majority of share in institutional-grade, compliance-first transfer/settlement scenarios. PayPal's PYUSD targets the retail end with its merchant network, and Ripple's RLUSD aims for interbank large-scale settlements.

This market is no longer dominated by a single player but is rapidly moving toward specialized division of labor.

04. The Ambition and Boundaries of Tether

Having made so much money, what does Tether plan to do with it?

Buy mining farms. In Uruguay, Paraguay, and El Salvador, Tether has invested over $2 billion to establish 15 energy and Bitcoin mining sites, aiming to become the world's largest Bitcoin miner.

Buy AI. Through channels like Northern Data Group, Tether has invested over $1 billion in AI computing infrastructure.

Buy robots. At the end of 2025, Tether invested €70 million in Italian AI robotics startup Generative Bionics; simultaneously, it is considering an investment of up to $1.15 billion in German robotics company Neura, with the goal of producing 5 million humanoid robots by 2030.

The logic behind this is not hard to understand: in an economy operated autonomously by AI agents and robots, the value exchange between them requires an instant, programmable digital currency. And USDT is already the most obvious candidate for this role.

Regulators are also adding weight to this story. In July 2025, the U.S. "GENIUS Act" was officially signed into law, opening legal channels for regulated institutions to issue stablecoins and explicitly excluding stablecoins from securities and commodities. The EU's MiCA framework was fully implemented the same year, bringing stablecoins from the "gray area" into the mainstream regulatory视野 (vision/scope).

Core Wall Street circles have also begun to enter the game. U.S. bond primary dealer Cantor Fitzgerald holds about 5% of Tether's equity, and its CEO Howard Lutnick has repeatedly publicly endorsed the authenticity of Tether's reserves. This deep binding means: Tether is no longer just a crypto project; it has quietly embedded itself into the interest network of traditional finance.

05. Summary

From a stablecoin issuer to one of the top 20 global holders of U.S. bonds, and then to an investor in robot factories—every step of Tether's expansion points in the same direction:

The definition of money is quietly shifting from the printing presses of sovereign nations to digital networks that offer higher efficiency and lower friction.

This process is not a revolution but an infiltration.

SWIFT is still operating, banks are still open, and the Fed is still adjusting interest rates. But another system is growing at an astonishing rate in the gaps between them.

For everyone involved, it might be worth asking yourself one question:

In the next decade, in which system will the money in your hands operate?

Trending Cryptos

Related Questions

QWhat is Tether's business model and why is it so profitable?

ATether's business model, known as the 'stablecoin float game,' involves users exchanging 1 USD for 1 USDT. Tether then uses that USD to purchase U.S. Treasury bonds, which yield over 5% annually. Since Tether does not pay any interest to USDT holders, it keeps the entire spread. This, combined with investments in assets like gold and Bitcoin, results in extremely high profitability.

QHow does Tether's profitability per employee compare to major financial and tech firms like Goldman Sachs and NVIDIA?

AWith approximately 150 employees and a net profit of $13 billion, Tether's profit per employee is about $85.62 million. This is nearly 300 times the per-employee profit of Goldman Sachs and 85 times that of NVIDIA.

QWhat advantages do stablecoins like USDT offer over traditional cross-border payment systems like SWIFT?

AStablecoins like USDT offer significant advantage in speed, cost, and accessibility. A USDT transfer on the Tron network costs less than $1 and settles in about 30 seconds, operating 24/7. In contrast, SWIFT payments can take 3-5 business days and have much higher fees, ranging from 1.5% to 11%. USDT also provides financial access to the unbanked who have a mobile phone and internet connection.

QWhat is the concept of 'Pay-Fi' (Payment Finance) as described in the article?

A'Pay-Fi' refers to the combination of payment and finance that solves not just the movement of money from point A to B, but also allows that money to earn interest while in transit. Protocols like Huma Finance tokenize corporate accounts receivable to provide instant financing through on-chain liquidity pools, offering T+0 real-time settlement capabilities for cross-border trade.

QWhat are some of Tether's major investment areas with its significant profits?

ATether is investing its profits into several strategic areas: purchasing and building Bitcoin mining facilities to become a major miner, investing over $1 billion in AI computing infrastructure, and funding AI robotics companies like Generative Bionics and Neura with the goal of producing humanoid robots for a future economy where autonomous agents use digital currencies like USDT for value exchange.

Related Reads

Wall Street Morning News: V-shaped Rebound at Month-end, but Nasdaq Suffers Worst July in 12 Years; Funds Accelerate Concentration Towards Cloud Giants

Despite a V-shaped rebound at the end of July, the Nasdaq posted its worst July since 2004, while the S&P 500 had its worst July since 2014. Markets were jolted by geopolitical shifts, as President Trump canceled a planned strike on Iran, leading WTI crude to plunge over 8%. This, alongside OPEC+ announcing a supply increase, reversed crude's sharp July gains. Treasury yields surged, with the 10-year yield rising over 30 basis points in July—its largest July increase since 2005. In a rare move, the US and Japan jointly intervened to weaken the USD/JPY, aiming to prevent potential Japanese sales of US Treasuries. While the tech sector faced deleveraging pressure throughout July, cloud giants staged a massive rally on strong earnings. Microsoft, Amazon, and Google collectively added nearly $1.5 trillion in market value last week. Amazon soared over 15% on accelerating AWS growth, Microsoft extended historic gains, Google fully recovered post-earnings losses, and Meta ended an 11-day losing streak. In contrast, Apple tumbled over 7% on supply chain and guidance concerns, ceding its "world's most valuable company" title to Nvidia. The memory and storage sector corrected sharply. Gold edged up 0.91% in July, with analysts viewing the ~30% pullback from January highs as a potential basing period, supported by long-term central bank demand. Key events to watch this week include earnings from Palantir, AMD, SpaceX (its first post-IPO report), and memory giants like Western Digital. The US July non-farm payrolls report on Friday will be critical for gauging the Fed's policy path. SpaceX also faces a significant lock-up expiration, testing market liquidity.

marsbit38m ago

Wall Street Morning News: V-shaped Rebound at Month-end, but Nasdaq Suffers Worst July in 12 Years; Funds Accelerate Concentration Towards Cloud Giants

marsbit38m ago

Rubin Ultra Makes Major Cuts, Even Nvidia Can't Handle Memory Price Hikes?

NVIDIA's Rubin Ultra, the top-tier variant of the newly announced Rubin AI accelerators, has reportedly seen significant specification downgrades, according to an industry report from SemiAnalysis. Initially designed with four compute dies (4-die), the Rubin Ultra is now said to be reduced to a 2-die design. Key changes highlighted in the report include: * **No increase in peak theoretical compute performance**, remaining at 35 PFLOPs like the standard Rubin. * **Severe reduction in memory capacity** to 192GB using 8-Hi HBM stacks, which is less than the standard Rubin's 288GB using 12-Hi stacks. * **Negligible memory bandwidth improvement** of only 1 TB/s. * **Slightly higher chip-level power consumption**. * The **primary upgrade is a massive increase in scale-up interconnect capacity**, supporting connections for up to 576 GPUs via NVLink, compared to 72 for the standard Rubin. The report suggests the redesign is primarily a cost-optimization move driven by the sharp rise in HBM (High-Bandwidth Memory) prices. By reducing the expensive HBM content and shifting investment towards enhanced system-scale networking, NVIDIA aims to maintain the platform's value for large-scale AI training clusters while managing soaring material costs. The news reportedly triggered a sell-off in South Korean memory stocks, with SK Hynix and Samsung shares falling around 8%, as markets grew concerned that NVIDIA—a major HBM buyer—might be reducing its reliance on high-capacity memory, potentially capping future pricing power for memory makers.

Odaily星球日报58m ago

Rubin Ultra Makes Major Cuts, Even Nvidia Can't Handle Memory Price Hikes?

Odaily星球日报58m ago

Can Generative Models Finally Be Trained End-to-End? The Core Is a For Loop

This article introduces a novel training paradigm for generative models called Explorative Modeling (XM), which enables true end-to-end training. Traditionally, powerful generative models like autoregressive and diffusion models are not trained end-to-end. They are trained to predict a single small step but require iterative multi-step sampling for inference. This "exposure bias" leads to error accumulation and limits performance. The core challenge XM addresses is "mode blurring." In generative tasks, a single input (e.g., "generate a dog") corresponds to many valid outputs (multiple modes). Standard training objectives like reconstruction loss force the model to average these modes, producing unrealistic, blurry outputs. To avoid this, existing models break generation into many small, almost deterministic steps, sacrificing end-to-end training. XM tackles this by restructuring the training loop itself. Its key insight is to amplify "generative expressivity." For each training input, instead of generating one sample, the model generates K candidate outputs. Only the candidate closest to the real data is used for computing the loss and updating the model via backpropagation. This simple "best-of-K" mechanism is implemented as a short for-loop. By exploring multiple possibilities, the model learns to distribute its guesses across different modes rather than collapsing to their uninformative average. The paper demonstrates that "exploration" acts as a new, powerful scaling axis. Gains from XM increase with model size, data scale, and compute. Experiments show improvements in FID scores for image generation and significant efficiency gains, sometimes outperforming larger models without exploration. When pushed to the limit, XM enables fully single-step, end-to-end generative models. In robotics tasks, an "Explorative Policy" matched the performance of a 100-step Diffusion Policy with a single forward pass, drastically improving inference speed. While the best-of-K concept is not entirely new, the authors' contribution lies in formally understanding it as a direct method to boost generative expressivity without fragmenting the generation process. This work suggests that as models scale, enhancing exploration during training may become crucial for overcoming fundamental performance bottlenecks.

marsbit1h ago

Can Generative Models Finally Be Trained End-to-End? The Core Is a For Loop

marsbit1h ago

Trading

Spot

Hot Articles

Discussions

Welcome to the HTX Community. Here, you can stay informed about the latest platform developments and gain access to professional market insights. Users' opinions on the price of AI (AI) are presented below.

活动图片