The Four Fates: How Companies Die, Shrink, or Pivot in the Age of AI

Generated illustration

Published: April 15, 2026
Reading: 7 min read
Written by: Nicolas SiatrasCEO & Founder

Last week I argued that AI won't consolidate industries — it will fragment them. When the cost of building drops to near zero, a million small players emerge to serve niches that were always there but never economically viable. The big players don't get disrupted by a competitor. Their market evaporates.

If you haven't read that piece, start there. This one is about what comes next.

Because if the market is fragmenting, every incumbent lands in one of four places. History has run this experiment before — with beer, with publishing, with photography, with media. The pattern is remarkably consistent. And it's not kind to most of them.

Fate 1: Die (The Kodak Pattern)

When your entire value proposition is tied to the scarcity that AI just destroyed, you die.

Kodak invented digital photography but couldn't escape its film-based business model. Nearly 90% workforce decline. Bankruptcy. Encyclopaedia Britannica — 244 years in print — was outpaced by volunteers on Wikipedia and stopped the presses. Monastic scriptoriums lost their reason to exist after the printing press, though scribes kept copying by hand for decades. Blockbuster had the chance to buy Netflix and passed.

The pattern: the incumbent's value was never the product. It was the difficulty of producing the product. Once that difficulty disappears, so does the business.

In software, this is the company whose pitch is "we built a generic thing and you have no alternative." The moment a 50-person company can build its own — and they can now — that pitch is dead.

Fate 2: Shrink but Survive (The Big Beer Pattern)

The US went from 1,300 breweries in the 1990s to over 9,000 today. Brewing technology got cheaper. Distribution got easier. Suddenly a 3-person operation serving a neighborhood was a viable business.

AB InBev didn't die. But it stopped growing. The entire growth segment went to thousands of small players serving niches.

AB InBev tried to adapt. They bought ten craft breweries between 2011 and 2017. In 2023 they sold eight craft brands to Tilray, three of them from that buying spree. The reason is structural: in one US market, craft made up 41% of beer SKUs but just 5% of gross profits, according to Beer Business Daily. Niche economics don't work inside a machine built for mass production. The business models are, as beer writer Jeff Alworth put it, "largely incompatible."

The same thing happened to network TV. NBC, CBS, ABC still exist, still profitable — but where the big three once drew more than 90% of American TV audiences, all broadcast TV combined now holds about a fifth of US viewing time. The Big 5 book publishers still dominate bestseller lists while 92% of America's self-published print books went through Amazon's own platform in 2018.

In software, this is SAP. This is Oracle. They'll keep their Fortune 500 clients. But the SMB and mid-market? Gone. Those companies are building their own tools, or getting them from small niche players who understand their specific world.

The big players survive. They just stop mattering for the part of the market that's actually growing.

Fate 3: Pivot to Platform (The Adobe Pattern)

This is the smart play — and very few pull it off.

The smartest incumbents stop making the product and become the infrastructure for the fragmented new market. Adobe went from desktop publishing software to Creative Cloud — the platform millions of small creators use. AWS went from retailer infrastructure to the platform enabling millions of small software companies. Shopify doesn't sell products. It enables millions of small businesses to sell theirs. YouTube doesn't make content. It's the platform that millions of 2-3 person teams build on.

The insight these companies grasped: when the market fragments, the real money shifts from producing end products to enabling the ecosystem that produces them. Picks and shovels, not gold.

In the AI era, this means providing the AI infrastructure, the dev tools, the deployment layers. Don't make the software. Be the platform for the millions who do. This is why big tech is collectively pouring well over $500B into AI infrastructure this year. They're betting on being the platform layer, not the product layer.

But here's the catch — there's only so much room at the platform level. Not every incumbent can pivot here. Most won't.

Fate 4: Try to Acquire Your Way Out — and Fail

This is the most tempting path and the most doomed.

Constellation Brands paid $1 billion for Ballast Point in 2015. Sold it for pennies on the dollar in 2019. AB InBev's ten craft acquisitions? Three already sold off. Big media conglomerates kept acquiring smaller outlets and couldn't prevent fragmentation.

Acquisition fails because the structure of a large company — its distribution, its cost base, its decision-making speed — is optimized for scale. Niche operations need the opposite: agility, specificity, low overhead. You can't bolt a niche business onto a mass-production machine and expect it to thrive. The economics are fundamentally incompatible.

In software, large incumbents will try to buy small AI-built tools and integrate them. It will mostly fail, for exactly the same structural reasons it failed in beer, in media, and in publishing.

This Isn't Just Software

The four fates apply to every knowledge-intensive industry where AI is collapsing production costs.

Legal services. Large firms bill by the hour. Lawyers expect AI to save each of them about five hours a week — roughly 240 hours a year, worth some $20 billion to the US legal industry. But the more efficient AI makes them, the less they can bill. They'll shrink-but-survive, keeping Fortune 500 clients. Meanwhile, five-person AI-augmented firms serve millions of small businesses that currently do zero legal work. Total legal work consumed goes up. The big firms just don't capture any of the growth.

Architecture. Between 2015 and 2023, small firms' share of national billings fell by half. AI reverses this. Solo practices outsource rendering and compliance to AI, competing on agility. Every restaurant renovation, small office, and custom home that couldn't justify professional design now gets it from a 2-person studio.

Marketing. Every 10-person e-commerce brand now produces campaigns that previously required a $50,000/month agency retainer. Large agencies keep Fortune 500 clients. But the growth — every Etsy seller, every local restaurant, every niche SaaS running sophisticated campaigns — fragments across a million small producers.

Accounting. AI adoption jumped from 9% to 41% in a single year. Small businesses that fly blind on generic QuickBooks get AI-powered analysis tailored to their industry. The Big 4 pivot to platform. Mid-tier firms get squeezed.

The pattern is the same everywhere. The big players keep the top. The long tail explodes. The middle gets crushed.

The Part Nobody Expects: Quality Goes Up

Here's the counterintuitive truth about fragmentation.

Most people hear "a million small players" and think "lower quality." A million small operations can't match the polish of a well-funded enterprise product, right?

Wrong. Because quality isn't about polish. It's about fit.

In the old model, three companies make three generic products. Each is maybe 60% fit for any given user. In the new model, each company builds exactly what it needs — 95% fit. The average quality of software experienced by users goes up, even if individual products are less polished.

A local IPA from a 3-person brewery isn't "worse" than Budweiser. It's better — for the person drinking it. A niche YouTube tutorial that solves your exact problem is more valuable than a generic TV segment, even with lower production value.

Fragmentation doesn't mean lower quality. It means higher relevance. And relevance is what users actually care about.

This flips the entire conversation. The future isn't a few giants making polished software that sort of works for everyone. It's a million small players making specific software that works exactly for someone.

So — Which Fate Is Yours?

The inversion is coming. The economics guarantee it. The data confirms it's already started.

If you're a large incumbent: which of the four fates are you heading toward? And are you choosing it deliberately, or drifting into it?

If you're mid-market: the squeeze is real. You're too big to be nimble, too small to be a platform. What's your move?

If you're small or just starting: the "worth building" threshold just dropped to near zero. What niche — what audience that was always there but never economically servable — are you going to own?

The demand was always there. The cost finally dropped. The only question left is what you do about it.

Which fate is yours?

Sources

Eastman Kodak, Form 10-K for 2011 (SEC): about 17,100 employees at the end of 2011. sec.gov
Rochester Business Journal, Eastman Kodak chronology (27 January 2012): a workforce peak of 145,300 in 1988. rbj.net
The Christian Science Monitor (14 March 2012): Britannica ends its print edition "for the first time in its 244-year history". csmonitor.com
Rare Book School, on manuscript production after printing; Johannes Trithemius, In Praise of Scribes (1494). rarebookschool.org
Fortune (14 April 2023): Marc Randolph on Blockbuster declining to buy Netflix for $50 million in 2000. fortune.com
Brewers Association, national beer statistics: 9,578 craft breweries in 2025, 9,724 in all. brewersassociation.org
Fortune (3 May 2017): Wicked Weed is AB InBev's tenth craft brewery acquisition. fortune.com
Brewbound (7 August 2023): Anheuser-Busch sells eight craft brands to Tilray, including Blue Point, 10 Barrel and Breckenridge. brewbound.com
Jeff Alworth, Beervana (18 November 2024), citing Beer Business Daily: craft at 41% of SKUs and 5% of gross profits in one market. beervanablog.com
Constellation Brands, Form 8-K on the Ballast Point acquisition (16 November 2015), and Form 10-Q for the quarter to 31 May 2020: proceeds of $41.1 million from the sale. sec.gov and sec.gov
EBSCO Research Starters, the decline of the big three networks: more than 90% of US TV audiences from the 1950s to the 1970s. ebsco.com
Nielsen, The Gauge for January 2026: broadcast at 21.5% of US TV usage. nielsen.com
Publishers Weekly, citing Bowker: Amazon's platform handled 92% of US self-published print titles in 2018. publishersweekly.com
2026 capital spending guidance, January and February 2026: Amazon about $200 billion (Data Center Dynamics), Alphabet $175–185 billion (Constellation Research), Meta $115–135 billion (Meta investor relations). datacenterdynamics.com · constellationr.com · investor.atmeta.com
Thomson Reuters, Future of Professionals report (June 2025): five hours a week saved per legal professional, worth $20 billion to the US legal industry, as predicted by respondents. thomsonreuters.com
American Institute of Architects, 2024 Firm Survey Report (4 November 2024). aia.org
Wolters Kluwer, 2025 Future Ready Accountant report (8 October 2025): AI adoption from 9% in 2024 to 41% in 2025. wolterskluwer.com
Shopify, About (shopify.com), and Amazon Web Services, What is AWS (aws.amazon.com): both describe millions of customers.

First published in Uncomplicated, Nicolas's newsletter on LinkedIn. Figures updated since the LinkedIn edition.