Disruption Capital - 15 March, 2026

Every technology cycle produces a phase where the narrative outruns the infrastructure. The pattern is consistent: maximalist claims dominate the discourse, capital floods in at historically unprecedented levels, and the technology eventually settles into applications that almost nobody at the peak anticipated.

This week, two developments illustrate both sides of that dynamic. The AI sector is exhibiting the financial signatures of peak-cycle investment — the widest gap between capital deployed and revenue generated in modern technology history. Meanwhile, the incumbent exchange infrastructure is quietly building the tokenised securities rails that may define the next decade of capital markets. One story is about narrative. The other is about plumbing. The plumbing tends to win.

There is a well-documented phase in every transformative technology cycle where the dominant narrative inflates past what near-term delivery can support. AI has entered that phase. And for anyone who has watched previous cycles closely, the structure of the current claims is remarkably familiar.

The Bitcoin precedent

Between 2015 and 2017, the dominant narrative in cryptocurrency was that Bitcoin would replace the global financial system entirely. Central banks would become obsolete. Commercial banking would disappear. The 1.7 billion unbanked adults worldwide would leapfrog traditional financial infrastructure. These were not fringe beliefs. They were mainstream positions held by the industry's most influential voices, backed by packed conference keynotes projecting five-to-ten-year timelines for complete disruption.

During the 2017 bull run, Bitcoin went from roughly 1,000 dollars in January to 19,700 dollars in December. The maximalist narrative felt inevitable.

Then transaction fees surged to 30-50 dollars per transaction. Block confirmation times stretched to hours. The network could process roughly seven transactions per second while Visa handled thousands. Bitcoin did not replace banking. It settled into its actual role: a speculative store of value.

What emerged instead was the use case that almost nobody in the maximalist camp anticipated — stablecoins. Dollar-denominated tokens became the primary practical application of blockchain rails. The stablecoin market now exceeds 313 billion dollars and processes transaction volumes approaching Visa's scale. The technology's dominant application turned out to be the exact opposite of the maximalist vision: the dollar's most efficient distribution mechanism.

The AI maximalist claims follow the same arc

The current consensus in venture capital and technology commentary reads something like this: SaaS will disappear because AI allows everyone to build their own software. Most knowledge work will be automated within three to five years. Traditional employment structures will collapse. Money itself may become obsolete.

These claims have an almost identical structure to Bitcoin maximalism. And the spending data reinforces the parallel.

In 2026, Google, Microsoft, Amazon, and Meta are projected to spend between 635 billion and 690 billion dollars on AI infrastructure. Google alone has committed 175-185 billion dollars. Amazon roughly 200 billion dollars. Including Oracle, the five largest U.S. cloud and AI providers have collectively committed approximately 660-690 billion dollars in capital expenditure this year, approximately 2.2% of U.S. GDP directed at a single technology category from five companies.

For context, during the dot-com boom, telecom companies collectively spent over 500 billion dollars on fiber optic infrastructure, accumulated over roughly five years. In 2001, an estimated 95% of that fiber sat dark. Telecom stocks lost over 2 trillion dollars in market value. The fiber eventually proved essential. The companies that built it were mostly destroyed. The infrastructure survived. The investors frequently did not.

The AI build is already larger than the fiber buildout in both absolute and relative terms. And it is concentrated in a single year rather than across five.

The market is already repricing

The financial signatures of peak-cycle deployment are visible. Between January and February 2026, approximately 2 trillion dollars in SaaS market capitalization evaporated. The SaaS index fell 6.5% while the S&P 500 rose 17.6% — the widest divergence in over a decade. Forward price-to-earnings ratios for software companies collapsed from roughly 35x to 20x. Apollo Global Management cut their software exposure in private credit funds from 20% to 10%. Amazon's AI spending is projected to push free cash flow negative by 17-28 billion dollars in 2026.

David Cahn at Sequoia Capital calculated that AI companies needed to generate approximately 600 billion dollars in annual revenue to justify the infrastructure spend, against actual AI revenue of roughly 100 billion dollars. That gap between capital deployed and revenue generated is the defining feature of peak-cycle investment.

What actually materialises

Global SaaS spending is still projected to grow from 318 billion dollars in 2025 to 576 billion dollars by 2029. The incumbents — Salesforce, ServiceNow, Adobe — are embedding AI into existing platforms, absorbing the new technology rather than being displaced by it. This is the pattern that repeated with cloud computing, mobile, and social media in previous cycles.

The claim that everyone will build their own software mirrors the claim that everyone would become their own media brand. YouTube and TikTok made video production accessible to anyone. The creator-to-consumer ratio on the internet has held roughly constant at 9-to-1 for two decades. Democratising tools expanded the creator base. It did not eliminate demand for curated, professional, reliable products.

For institutional allocators, the implication is structural: the technology is real, the infrastructure being built will eventually prove essential, and the actual applications that endure will be more mundane, more embedded in existing systems, and more dependent on human judgment than the current maximalist narrative suggests. The spending may prove justified over a long enough timeframe. The specific claims about what AI will do, how fast, and to whom will not survive contact with reality. They never do.

While AI maximalism dominates the technology narrative, a quieter structural shift is accelerating in capital markets infrastructure. Nasdaq has partnered with Kraken to distribute tokenised versions of listed stocks globally — with full shareholder rights including voting, dividends, and proxy ballots. Settlement runs through the Depository Trust Company. Launch is scheduled for early 2027.

The significance is in the distribution architecture. A crypto-native platform is now a primary distribution channel for Nasdaq-listed equities. This is production infrastructure, not a pilot programme.

Three signals worth examining

First, ICE — the parent company of the New York Stock Exchange — invested in OKX at a 25 billion dollar valuation during the same week. Two of the world's largest exchange operators are building crypto-native rails simultaneously. When incumbent infrastructure providers commit capital to a new distribution paradigm, the signal carries a different weight than when startups make similar claims.

Second, U.S. banking regulators — the Federal Reserve, FDIC, and OCC — recently affirmed that tokenised securities receive identical capital treatment to traditional assets. Kraken has secured Federal Reserve payment infrastructure access. The regulatory gap that previously separated tokenised and traditional securities is narrowing rapidly.

Third, the architecture being constructed enables 24/7 global equities trading on blockchain rails — built by the incumbent exchanges themselves, not by the startups that originally envisioned disrupting them.

The institutional implications

This development echoes the stablecoin pattern in a different asset class. The maximalist narrative for tokenised securities — that decentralised exchanges would displace traditional market infrastructure — is giving way to a more pragmatic reality: incumbents adopting blockchain rails to extend the reach and efficiency of existing systems.

For family offices and wealth advisors, the question has shifted from whether tokenised securities will reach institutional adoption to how rapidly the transition occurs. The regulatory framework, the settlement infrastructure, and the distribution partnerships are now being assembled by the entities that control the existing plumbing.

The architecture for global, continuous equities trading on blockchain-native rails is being built. The builders are not startups. They are Nasdaq and NYSE.

The connective thread across this week's signals is the gap between what technology narratives promise and what technology infrastructure actually delivers. AI maximalism is following the same arc that Bitcoin maximalism traced — extraordinary spending, extraordinary confidence, and specific claims that will not survive contact with reality. Meanwhile, the incumbents are quietly building the rails that will carry the next generation of financial infrastructure. Nasdaq is distributing equities through a crypto exchange. Stablecoins are processing 1.8 trillion dollars in monthly volume. Florida is regulating digital dollar issuers.

The maximalist vision — whether for AI or for crypto — consistently overestimates the speed of disruption and underestimates the adaptability of existing institutions. The technology is real. The applications that endure tend to be more pragmatic, more incremental, and more embedded in existing systems than the peak narrative anticipates.

If you're exploring how stablecoins fit into your capital allocation strategy, get in touch with the team here: https://brava.finance/contact

Sources: