The trillion-dollar headline is the standard opening move in a financial newsletter pitch, and the summer of 2026 has five of them running at the same audience at once. Louis Navellier’s Golden Dawn presentation leads with $100 trillion in economic disruption. Jim Rickards’ American Birthright leads with $150 trillion in federal mineral wealth. Jeff Brown’s Anthropic thesis quotes the $4.4 trillion McKinsey generative AI figure. Addison Wiggin’s Grey Swan Dollar 2.0 pitch leads with $6.6 trillion in bank deposits shifting to stablecoins. Mark Skousen’s America Reloading leads with a $1.5 trillion defense buildout.

Five pitches, five numbers, five different methodologies, one shared structural move. Each takes a real analyst figure, strips the scope and time horizon, and presents the number as addressable revenue for a stock the subscription will name. The numbers are not fabricated. The connection between the number and the stock is where the work happens, and that connection is the part the presentation never explains.

What the AI Market Actually Is

There is no single AI market size. Depending on what you count, the 2026 figure ranges from $37 billion to $15.7 trillion, a 425-to-1 spread where every number is correct and none are interchangeable.

Gartner’s January 2026 forecast puts worldwide AI spending at $2.52 trillion for 2026, a 44 percent year-over-year increase. That figure measures total procurement across the full stack: hardware, software, services, platforms, and models. It is the broadest spending number and the one most often quoted as “the AI market.”

Stanford HAI’s 2026 AI Index measured global corporate AI investment at $581.7 billion in 2025, a 130 percent single-year surge. That figure tracks capital flowing into AI companies through venture capital, private equity, and mergers and acquisitions. It is not spending on AI; it is investment in AI companies. The distinction matters because investment capital and procurement spending resolve on different schedules and produce different returns.

Precedence Research pegs the global AI market at $638 billion in 2026, a 35 percent increase from 2025. That figure covers hardware, software, and services but uses a narrower scope than Gartner’s full-stack procurement measure. The difference between Precedence’s $638 billion and Gartner’s $2.52 trillion is roughly $1.9 trillion, and the gap exists because the two firms count different layers of the stack.

Bloomberg Intelligence’s June 2026 forecast puts the generative AI market at $2.3 trillion by 2032, representing 22 percent of total technology spending. That is a 2032 projection, not a 2026 figure, and it covers generative AI specifically rather than all AI.

McKinsey Global Institute’s 2023 study estimates generative AI could add $2.6 trillion to $4.4 trillion in annual value across 63 use cases, with the largest pools in customer operations, marketing, software engineering, and research and development. That figure measures annual economic value creation, not market size, not revenue, and not corporate profit. McKinsey’s broader AI estimate, combining generative and nongenerative AI, reaches $17.1 trillion to $25.6 trillion annually when including productivity gains across knowledge-worker activities.

PwC’s 2017 “Sizing the Prize” study estimates AI could contribute $15.7 trillion to global GDP by 2030, a 14 percent increase against a no-AI scenario, with China at a 26 percent GDP boost and North America at 14.5 percent. That figure was published before generative AI existed as a commercial product, covers all AI rather than generative AI specifically, and has not been superseded by an equivalent updated analysis, which is why it still circulates in pitches nine years after publication.

Menlo Ventures tracks enterprise spending on generative AI tools at $37 billion for 2025. Exponential View’s bottom-up model puts the generative AI ecosystem’s actual revenue at $110 billion over the past 12 months, with a $175 billion annualized run rate. Goldman Sachs, cited by Axis Intelligence Research, notes that hyperscalers will spend an estimated $500 billion annually on AI infrastructure through 2027, against a revenue base that does not yet justify that spending on a standalone return basis.

The US Bureau of Economic Analysis reported current-dollar GDP at an annualized $31.87 trillion in the first quarter of 2026. Stanford’s Digital Economy Lab estimates the US consumer surplus from generative AI tools at $172 billion annually by early 2026, most of it accruing to users of free or near-free tools rather than to the companies spending hundreds of billions on infrastructure.

Each of these numbers answers a different question. Gartner answers what organizations will procure, Stanford HAI answers what capital flows into AI companies, Precedence answers what the hardware-software-services market is worth, Bloomberg answers what the generative AI market will be in 2032, McKinsey answers what generative AI could contribute in annual economic value, PwC answers what all AI could add to global GDP by 2030, Menlo Ventures answers what enterprises spend on generative AI tools, and Exponential View answers what the generative AI ecosystem actually earns in revenue. None of these is wrong, and none of them is interchangeable with any other.

How the Pitches Use the Numbers

The structural move is identical across all five campaigns: each takes a real analyst figure, strips the scope and time horizon, and presents the number as addressable revenue for a stock the subscription will name.

Navellier’s $100 trillion Golden Dawn figure stacks McKinsey’s $4.4 trillion generative AI estimate with Bloomberg’s $40 trillion energy and fusion forecast, adds medicine and biotech and quantum computing, and applies a 36,000 percent acceleration multiplier to produce a composite number that is roughly 25 times McKinsey’s high-end annual value estimate and roughly 4 times PwC’s 2030 all-AI GDP contribution forecast. The number traces to Navellier’s own framing of cumulative economic disruption across multiple sectors over an unspecified time horizon, not to any single analyst’s market-size estimate. The pitch then names a chipmaker building the Oak Ridge supercomputer as the stock that captures the thesis.

Rickards’ $150 trillion American Birthright figure traces to a 2013 Institute for Energy Research report funded by ExxonMobil and the American Petroleum Institute, which estimated the gross value of federal mineral estate. Snopes verified in April 2025 that the statutory layers are real: the General Mining Act of 1872, the Chevron Doctrine overturn in June 2024, and the federal mineral estate calculation all check out. The $150 trillion is the gross value of minerals in the ground across federal land, not a market, not recoverable at scale, and not a stock return. The pitch describes a Houston-based seismic sensor manufacturer as one of the picks.

Brown’s Anthropic thesis quotes McKinsey’s $4.4 trillion generative AI figure as the addressable market for a single privately held AI company. McKinsey’s number measures annual economic value creation across 63 use cases globally, not the revenue addressable to any single company. Anthropic’s actual revenue run rate, per reporting in The Information and Reuters in 2025, was in the single-digit billions. The gap between the $4.4 trillion macro figure and the company’s revenue is roughly three orders of magnitude.

Wiggin’s $6.6 trillion Grey Swan Dollar 2.0 figure traces to a Treasury Department framing of bank deposits potentially shifting to stablecoins. The stablecoin market’s total capitalization in mid-2026 is roughly $250 billion, per CoinGeoper and DefiLlama tracking. The $6.6 trillion is a projection of a projection, applied to a pitch naming three publicly traded crypto-infrastructure companies by their roles in the stablecoin stack.

Skousen’s $1.5 trillion America Reloading figure is a multi-decade defense rearmament projection, tied to the fiscal year 2027 defense budget and the January 2027 NDAA critical-minerals ban. The number is real as a budgetary commitment. The pitch describes five unnamed defense and minerals picks as the way to play it.

The 1999 Parallel

The last time headline market-size figures drove a promo cycle at this scale was the dotcom boom, and the mechanism was identical.

Geoffrey Moore’s 1991 book “Crossing the Chasm” formalized the Total Addressable Market slide as a standard element of venture capital pitches. By 1998, the TAM slide had migrated from VC pitch decks to sell-side analyst reports to retail newsletter presentations. Morgan Stanley’s Mary Meeker published an “Internet Trends” report in 1998 projecting a $1.3 trillion internet economy by 2002, and Forrester Research forecast online retail sales of $108 billion by 2003. Those numbers were real estimates with real methodologies.

The pitches that ran against those numbers did what the 2026 pitches do. They took the TAM figure, stripped the time horizon and the scope, applied it to a single company, and sold the subscription. Cisco’s market cap exceeded $500 billion in March 2000, priced as if the entire projected internet infrastructure spend would flow through its routers. Analysts published “internet economy” forecasts reaching $7 trillion by 2004 and applied them to companies whose actual revenue was a fraction of one percent of that figure.

The Nasdaq fell 78 percent from its March 2000 peak over the following 30 months. Cisco dropped 88 percent from peak. The internet economy Meeker and Forrester projected was real, and it arrived roughly on schedule. The companies that survived the washout went on to define the next two decades. The investors who bought the TAM-justified valuations at the peak waited 15 years to break even on some positions, and many never did. The market size was correct. The stock application was the error.

Why Market Size Is Not a Stock Return

The gap between a market-size figure and a stock return has four layers, and each one filters the number down by a factor that the pitch never mentions.

A market size is the total dollars flowing through a category. A company’s addressable revenue is the share of that market the company can realistically capture, which depends on competition, geography, distribution, and pricing power. A company’s earnings are its revenue minus costs, and in capital-intensive categories like AI infrastructure, the costs can consume most of the revenue during the buildout phase. A stock’s return is the earnings multiplied by the valuation multiple the market assigns, and multiples contract when interest rates rise, when competition intensifies, or when the growth narrative breaks.

The McKinsey $4.4 trillion generative AI value figure is annual economic value creation across 63 use cases globally. To get from that number to a single stock’s return, you pass through roughly six layers of filtering: the share of the $4.4 trillion that is addressable revenue rather than cost savings accruing to users, the share of that revenue captured by a specific company rather than its competitors, the company’s profit margin on that revenue, the multiple the market assigns to those earnings, the time horizon over which the revenue materializes, and the discount rate applied to future earnings. Each layer reduces the number. The pitch quotes the first figure and skips the other five.

The Stanford Digital Economy Lab’s $172 billion consumer surplus figure is the counter-evidence the pitches never mention. Real economic value is being created by generative AI. Most of it is accruing to users of free or near-free tools rather than to the companies spending hundreds of billions on infrastructure to provide them. That structural tension, which Goldman Sachs has described as the central economic question of the AI buildout, is the reason a large market-size figure does not translate into proportional corporate earnings.

What the Numbers Tell You

The trillion-dollar headline is a signal that a real macro shift is happening, and the shift is usually worth paying attention to. The AI infrastructure buildout is real, the federal minerals reclamation push is real, the stablecoin regulatory framework is real, and the defense rearmament cycle is real. Each of these is a genuine structural change in how capital flows through the economy.

The headline does not tell you which company captures the shift, what the company earns, what the stock is worth, or when the value materializes. Those are separate questions, and they answer to different math than the market-size figure answers to. The pitch conflates them because conflation is what sells the subscription. The reader’s job is to hold them apart.

The investors who navigated 1999 successfully held the technology thesis and the valuation question as separate claims. The technology was real and arrived on schedule. The valuations were wrong and corrected violently. Both conditions held simultaneously, and the investors who acted on only one side of that either missed the recovery or rode the crash down. The 2026 pitches are running the identical structure against the identical audience, substituting AI for internet and trillion for billion, and the pattern holds even as the names change.

See the promo watch board for more newsletter-campaign coverage.

See the guides index for more publisher and thesis explainers.