Regulation, AI, And The Next Market Structure

Coinbase crypto exchange app and other cryptocurrency exchange apps on smartphone

getty

Bitcoin slipped below $90,000 on December 1, 2025, erasing $1 billion in leveraged positions within hours. The same week, Bitnomial launched the first fully CFTC-regulated leveraged crypto exchange in the U.S., a milestone in institutional legitimacy arriving amid retail carnage. The contrast captures crypto’s defining tension: the industry is simultaneously building regulated infrastructure and automating itself into machine-speed volatility amplifiers. The question is which transformation defines the next decade.

For more than a decade, exchanges competed on two blunt weapons: liquidity and listings. Who could move the most volume? Who could list the fastest? Who could offer the most leverage?

Now that model is breaking. Trading fees have compressed toward zero, and a pure transaction-fee business no longer scales.

Exchanges are bifurcating along two radically different paths. Coinbase is building institutional infrastructure through regulatory compliance. Bitget is scaling retail participation through AI-powered trading automation. One bets on trust. The other bets on intelligence.

The FTX Catalyst

FTX’s November 2022 collapse set off a chain reaction. One centralized platform after another—Celsius, Voyager, and BlockFi—revealed the same structural weakness: customer funds held without regulatory oversight. $8 billion disappeared from FTX alone, and with it, any credible claim that centralized exchanges could self-regulate. FTX proved liquidity without custody trust is worthless.

Custodial platforms faced a binary choice: pursue U.S. institutional markets through domestic regulatory compliance, even if that meant slower product cycles and expensive infrastructure, or operate in offshore jurisdictions with different regulatory frameworks, optimizing for global retail velocity.

Liquidity Has an Address: Binance

Binance remains crypto’s gravitational center. Despite regulatory pressure, leadership transitions, and its 2023 DOJ settlement, Binance still commands roughly 35–40% of global spot volume and an even larger share of derivatives trading. It remains the primary venue for price discovery. When Bitcoin moves, it often moves on Binance first. Institutional desks watch their order books closely.

No exchange can compete with that liquidity directly, so they’re building around it instead. Coinbase is targeting what Binance cannot legally reach: regulated institutional capital, offering substantial liquidity and a growing role in price formation, particularly in U.S. spot markets. Bitget is layering AI engagement tools over Binance-driven price action. Both strategies accept Binance’s liquidity dominance as structural reality.

Coinbase: Building the Regulated Financial Operating System

Coinbase made the compliance bet early. While offshore exchanges launched 20x leverage perpetuals and listed hundreds of speculative tokens weekly, Coinbase spent on regulatory licenses, custody infrastructure, and CFTC-supervised derivatives platforms. Post-FTX, that plumbing became a moat.

Coinbase now spans custody, staking, institutional prime brokerage, international derivatives, and on-chain settlement through its Base network, its Ethereum layer-2 network that processes transactions faster and cheaper than Ethereum’s main blockchain. Each layer serves the same structural goal: moving beyond a fee-only exchange model that no longer scales. Increasingly, Coinbase monetizes how capital sits, moves, and settles across the crypto economy.

That strategy rests on four pillars: custody for ETFs, institutions, and corporates; prime services combining execution, financing, and reporting; subscription and services revenue from staking, interest, and custody fees; and embedded crypto rails for banks, ETFs, and asset managers.

But this positioning trades market risk for regulatory risk.

Coinbase’s roughly $300 billion in custody assets, expanding ETF access, and institutional distribution all scale only if the current U.S. regulatory framework holds. Securing that framework required the crypto industry’s largest political wager to date. Coinbase contributed $75 million to Fairshake and affiliated PACs in the 2024 election cycle, part of the industry’s $245 million in crypto-related political spending, making crypto one of the top corporate political donors of the year.

The bet paid off. Fairshake and allied crypto PACs won 53 of 58 targeted races. The GENIUS Act, the first federal stablecoin regulation bill, advanced. The SEC’s February lawsuit was dismissed. Vanguard reversed its crypto ETF ban. Bank of America announced 1–4% crypto allocation recommendations for wealth clients. Bitnomial launched under full CFTC oversight.

What happens if a future administration turns hostile to crypto again?

The risk is real, but no longer binary: time is on the industry side. Crypto’s embeddedness is now compounding along two tracks: institutional infrastructure and voter alignment.

First, infrastructure. Stablecoins already process $5.7 trillion annually in cross-border payments, settling in minutes, operating 24/7, and increasingly serving as dollar-denominated savings in inflation-prone economies. Total stablecoin supply has grown from $5 billion to $305 billion in five years, with $32 trillion in transaction volume in 2024. This is no longer speculative finance. It is operational plumbing.

Tokenization is following the same path. BlackRock’s tokenized money market fund, BUIDL, has surpassed $2.5 billion and is now accepted as collateral across crypto and traditional derivatives venues. Major banks including JPMorgan, Fidelity, State Street, Goldman Sachs, HSBC, Citi, and Deutsche Bank have launched tokenization platforms. Base processes over 12 million daily transactions.

This infrastructure creates inertia. It becomes costly to unwind as capital, collateral, and risk management depend on it. On December 11, DTCC received an SEC no-action letter to tokenize certain DTC-custodied assets, an implicit acknowledgment that blockchain is entering regulated market infrastructure.

Second, voters. Recent surveys estimate that 55–65 million Americans (21–28% of adults) currently own or have ever owned crypto, with ownership rates highest among younger generations: around half of Gen Z (up to 51%) and Millennials (up to 49%). Younger cohorts dominate the owner base, reflecting both higher adoption and recency of entry into the market. Ownership is broadly bipartisan, with similar rates across political parties or only modest differences. For many younger holders, crypto represents a meaningful portion of investable assets.

Every quarter of spot ETF inflows, every expansion in stablecoin settlement volumes, and every new tokenized security deepens crypto’s integration. With years before the next electoral reset, time favors platforms like Coinbase that built compliant infrastructure early.

Bitget: Testing the AI-Native “Universal Exchange”

If Coinbase is engineering crypto’s regulated foundation, Bitget is experimenting with a more speculative frontier: AI as the primary trading interface.

According to CoinGecko, Bitget holds roughly 7–8% of global spot volume and ranks among the top five exchanges in derivatives. That makes it smaller than Binance and Coinbase by traditional metrics, but size misses the point. Bitget is betting that AI-powered automation can sustain retail engagement. Whether that model scales beyond early adopters is uncertain, but the direction is clear: automation is crypto’s competitive battleground, not just liquidity or listings.

Bitget claims 120 million registered users and dominance in copy trading: 1.1 million followers, 200,000+ strategy providers, 110 million executed trades. These figures come from company disclosures without an independent audit. Recent product launches like tokenized stock futures generated over $10 billion in trading volume within a month according to the exchange.

That scale translates into meaningful retail presence, particularly in emerging markets where traditional banking infrastructure is weak. But unlike Coinbase’s institutional positioning or Binance’s universal dominance, Bitget’s competitive edge depends on sustaining user engagement through automation rather than raw liquidity or regulatory access.

Bitget is positioning itself as what CEO Gracy Chen calls the “Universal Exchange,” essentially an AI-mediated interface for existing spot, futures, and tokenized asset trading. The flagship is GetAgent, described in company materials as a conversational assistant executing trades via natural-language prompts, recently expanded with “AI avatars”—automated trading strategies running live accounts with public P&L that users can copy with one click. Bitget monetizes through trading fees on AI-generated volume, betting that AI sustains retail activity even in flat markets.

Bitget’s AI trading avatars, launched November 28, 2025, show positive returns across several strategies during their first eleven days of operation, too brief to evaluate performance during significant volatility. These figures are self-reported from the exchange’s Model Arena, and lack third-party verification. The broader question is replicability and stress performance. AI trading experiments under independent observation show inconsistent results: in nof1.ai’s Alpha Arena, where six foundation models traded crypto with real capital over 17 days during late October’s drawdown and early November’s weakness, four of six lost money, with GPT-5 down 63% and Gemini down 67%.

These performance metrics carry legal disclaimers: when copied strategies fail, followers bear full losses; Bitget’s ToS disclaims liability for AI-generated trades. More systemically, correlated failures (AI avatars piling into the same momentum play) could trigger cascading liquidations.

Bitget operates as an offshore exchange registered in the Seychelles, with a Cayman Islands entity. The exchange holds various licenses across Europe, the UK, El Salvador, and Australia while pursuing MiCA compliance. It blocks the U.S., Canada, and OFAC-listed jurisdictions, mandating KYC globally. No major enforcement actions have materialized, but the structure remains vulnerable to sudden regulatory shifts. Customer protections rely on a $300M+ protection fund, not FDIC-style guarantees.

The Uncomfortable Question: When AI Trading Fails, Who Bears the Risk?

In an AI-mediated trading environment, liability is undefined. If thousands of users deploy similar algorithmic strategies through AI agents and copy systems, risk becomes mechanically correlated. Discipline rapidly becomes herding.

If an AI agent propagates a flawed signal and triggers cascading losses, who is responsible? The trader? The exchange? The model developer? Traditional finance answers this through fiduciary duty, capital requirements, and liability frameworks tested through decades of litigation. Crypto exchanges answer it through Terms of Service disclaimers, not because they’re evasive, but because legal precedent doesn’t exist yet. Early internet companies navigated user-generated content liability without established case law; fintech platforms tested banking regulations designed for physical branches. The pattern repeats: fast-moving technology outpaces regulatory frameworks. The risk isn’t intentional evasion: it’s that when thousands of AI agents execute correlated strategies during the next liquidation cascade, no court has established who bears responsibility for algorithmic failures that individual users authorized but didn’t design.

Automation also complicates exit behavior. Humans hesitate. Machines do not. In a synchronized unwind, speed becomes an amplifier rather than a safeguard. What looks like empowerment during bull cycles can resemble automated herding during stress.

From this perspective, Bitget’s model is not just a product experiment. It’s a live systems test of whether AI absorbs volatility, or whether it ultimately compresses human risk into machine-speed feedback loops.

What Comes Next

The landscape has stratified. Coinbase owns U.S. institutional infrastructure. Binance dominates global liquidity. Bitget is testing AI-driven retail automation.

Each model carries a distinct risk. Coinbase inherits regulatory dependency, its growth conditioned on political continuity. Bitget inherits correlated automation risk, machine-speed herding during stress.

Coinbase is betting that institutional capital ultimately demands trust, compliance, and integration into existing financial systems, even at the cost of slower innovation. Bitget is betting that intelligence and automation can substitute for liquidity—that traders will delegate decision-making to AI agents rather than deeper order books.

Both bets may succeed. Both may prove incomplete. The more uncomfortable possibility is that the risks each embeds—regulatory dependency on one side, algorithmic correlation on the other—surface before either model fully matures, shaping the next crisis that defines crypto’s market structure.

Disclaimer: Companies mentioned are illustrative examples only. The author has no affiliation with any companies referenced. This does not constitute investment advice. Readers should conduct independent research.