Week in Review: AI’s Compression Effect — Solo Founders at 36%, Margins Doubling, and the One-Person Startup Era Begins

The Week That Was

This week crystallized AI’s dual nature: it’s simultaneously compressing startup formation (solo founders now 36% of companies), eating software from inside (margins doubling while growth halves), and challenging platform incumbents (AI browsers taking on Google’s 63% share). The compression is real. Whether the valuations reflect it correctly remains the central question.

Bank of America’s bubble comparison showed the AI rally has exceeded average duration while delivering half typical gains—leaving room to run or signaling a more moderate cycle. Meanwhile, structural shifts continued: Anthropic went fully vertical with 1 million TPUs, India bet $4.6B on capturing electronics manufacturing, and the Great Re-Sorting data revealed how dramatically pathways to top-5% income have narrowed by geography and profession.

—

The Big Themes

1. AI Compression of Everything

Solo founders rising from 24% to 36% of startups. Software margins expanding from 6% to 19% while growth halves. AI browsers challenging incumbents. The pattern: AI compresses the human and capital requirements for economic activity across every domain.

2. Platform Wars Intensify

OpenAI, Perplexity, and Microsoft launched AI-native browsers to challenge Chrome’s 63% share. The browser becomes the “operating system of the mind”—the interface through which AI agents perceive and act. Whoever controls it controls AI commerce.

3. Infrastructure Goes Vertical

Anthropic purchasing 1 million TPUs directly, deployed in facilities it controls. Crypto miners (TeraWulf, Hut8) pivoting to AI infrastructure. The model: own don’t rent, diversify away from NVIDIA dependency.

4. The Bubble Question Persists

Bank of America: 3 years duration (above 2.55 average), 131% gains (below 244% average). Tech narratives attach to visible events while macro forces drive markets. The question isn’t whether AI is real—it’s whether current prices already capture future value.

5. Geographic and Professional Re-Sorting

21% of top-earning millennials in California/Washington alone. Lawyers dropped from 32% to 22% reaching top-5% income. Tech and finance displaced medicine and law as wealth escalators. Two Americas with non-overlapping pathways.

—

🏆 Story of the Week

Solo Founders Rise to 36.3%: The One-Person Startup Era

Carta’s data shows a 53% increase in solo-founded startup share over six years, with sharp acceleration in 2024-2025 coinciding with AI tool mainstreaming. This is the week’s defining story because it captures AI’s compounding effect on economic activity at the most fundamental level: company formation.

If AI tools can substitute for co-founders (technical implementation, customer support, operational tasks), the minimum viable team shrinks to one. More startups form with less capital. Equity concentrates (no co-founder dilution). Speed increases (no coordination overhead).

The second-order effects ripple through venture economics, talent markets, and competitive dynamics. Solo founders competing against teams with AI leverage may have advantages previous generations lacked.

—

By the Numbers

—

Sector Breakdown

🤖 AI & Technology

AI Browsers Challenge Google’s 63% Share

OpenAI, Perplexity, and Microsoft frame browsers as AI operating systems—the interface for agents that browse, transact, and act. Google’s moats (share, Gemini 3, Search) are formidable, but whoever controls the browser controls AI commerce.

—

Anthropic Goes Fully Vertical: 1M TPUs

Anthropic purchasing chips directly, deploying in controlled facilities—owned infrastructure, not cloud rental. Crypto miners (TeraWulf, Hut8, Cipher) pivoting data center expertise to AI. Broadcom emerges as direct TPU supplier, fragmenting custom silicon away from NVIDIA’s grip.

—

🌍 Macro & Markets

AI Rally vs Historical Bubbles

Duration exceeds average, gains trail average. Either room to run toward 244% or a more moderate cycle. Unlike dotcom, NVIDIA’s $60B FCF provides real foundation—historical comparisons have limits when underlying economics differ.

—

The Great Re-Sorting: Pathways Narrow

Lawyers dropped from 32% to 22% reaching top-5% income. Tech/finance displaced medicine/law. San Francisco earnings +48.6% vs mid-tier metros flat. Profession, geography, and metro function as multiplicative filters—two Americas emerge.

—

🏢 Enterprise & Industry

Sportswear’s Simultaneous Crises

Nike (worst single-day stock drop), Lululemon (shares -44%, proxy fight), Under Armour ($2B market cap, M&A target). Cult brands like Tracksmith capture cultural energy. Incumbents built on scale face disruption when cultural relevance shifts faster than organizations adapt.

—

India’s $4.6B Electronics Component Bet

Targeting camera modules, display assemblies, metal enclosures—the components that determine whether India becomes assembly hub or manufacturing power. China hedge, not China replacement: offering multinationals diversification without requiring full decoupling.

—

What We Got Wrong

Previous assumption that AI tools would primarily benefit large teams with resources to implement proved incomplete. Solo founders are capturing disproportionate benefit—AI as co-founder substitute changes startup economics more fundamentally than AI as productivity tool for existing teams.

The bubble comparison framing also deserves scrutiny. Treating AI as another tech cycle assumes similar dynamics, but cash generation profiles (NVIDIA’s $60B FCF) and infrastructure concentration differ from historical precedents.

—

What to Watch Next Week

Browser market share data will signal whether AI browsers gain traction beyond early adopters. Enterprise AI renewal cycles continue—watch for churn data revealing true product-market fit versus early enthusiasm.

Anthropic’s TPU deployment progress and any infrastructure announcements from OpenAI will clarify the vertical integration trend. India’s implementation of electronics incentives faces execution risk—early signals matter.

The solo founder trend deserves tracking: if 36% becomes 40%+, we’re witnessing structural change in startup formation, not cyclical fluctuation.

—

The Framework Lens

This week’s events map to Disruptive Innovation theory in real-time. AI browsers are classic low-end disruption—initially worse at traditional browsing but better at agent tasks. Incumbents face the choice of cannibalizing search revenue to compete or defending legacy business while challengers improve.

The solo founder rise reflects economies of scale shifting from team coordination to AI leverage. Capital efficiency inverts: less money, fewer people, faster iteration.

Vertical integration (Anthropic’s TPU ownership) represents the counter-trend to platform dependency. When compute becomes strategic, ownership beats rental despite operational complexity.

—

The Bottom Line

This week revealed AI’s compression effect across every dimension: companies (solo founders), margins (half the people, same output), platforms (browsers as AI OS), and geography (prosperity concentrating further).

The strategic question isn’t whether AI is transformative—the data confirms it is. The question is who captures the value: the AI providers, the enterprises deploying it, or the workers whose productivity compounds. Current evidence suggests concentration: fewer founders needed, fewer employees needed, fewer geographies winning.

The winners will be those who recognize that AI compression is the trend, not AI adoption. Building for a world with fewer, more leveraged participants requires different strategies than building for broad-based growth.

—

This is the FourWeekMBA Weekly Roundup—the essential synthesis of what moved markets, shifted strategies, and revealed structural change. Subscribe to The Business Engineer for daily analysis.