Most important take away

Governments need to rebuild internal “dynamic capabilities” and pursue bold mission-oriented industrial policy (moonshots) instead of outsourcing core functions to consultants and merely fixing market failures. Without this shift, large publicly funded technological waves like AI will continue to socialize risks while privatizing rewards, leaving public problems (health, climate, housing) unsolved even as private rents balloon.

Summary

Actionable insights for investors and policymakers from Mazzucato’s framework:

• Mission-oriented industrial policy is back, but not all “industrial policy” is created equal. Watch whether spending (IRA, NextGenEU) is tied to genuine outcomes/conditionality or is just sectoral handouts. Companies that meet conditional terms (greener production, knowledge sharing, lower material content) tend to win durable government partnerships and subsidies.
• Conditional public lending shapes winners. Germany’s KFW required steel firms to reduce material content of production in exchange for loans, producing the “greenest steel in the world.” Investors should monitor public banks like BNDES (Brazil), KFW (Germany), and similar institutions because their conditionality drives sectoral transformation - companies aligned with green mandates are more likely to access cheap public capital.
• AI rents are unprecedented and should be watched as a structural risk and opportunity. The big AI firms are extracting “trillions, not billions” in excess returns rooted in publicly financed foundational research (DARPA, NSF, internet, GPS, touchscreens, Siri, LLM precursors). Mazzucato flags hemorrhaging of public-sector and academic talent into private AI labs as the most under-discussed change. Implication: regulatory pressure on AI rents, taxation, and disclosure regimes is a non-trivial tail risk for mega-cap AI beneficiaries (implicitly Big Tech: Amazon, Google/Alphabet, Microsoft, Meta, OpenAI ecosystem).
• AI-specific disclosures are coming. Mazzucato and Tim O’Reilly are working on an “Algorithmic Rents” framework analogous to climate/ESG disclosures. Expect new compliance overhead for AI firms - similar trajectory to climate disclosure adoption. Compliance and AI-governance vendors could benefit.
• “Bots fighting bots” is a real concern in healthcare and insurance. Startups simplifying hospital billing while insurers deploy AI to deny claims may cancel out, enriching only the bot makers. Be skeptical of AI healthcare plays that don’t fix the underlying system; favor those tied to systemic mission objectives.
• Consultants (“The Big Con”) are a value-extraction red flag. Heavy consultant dependence - McKinsey, Deloitte, PwC, Accenture - signals weak state/corporate capability. PwC’s Australian conflict-of-interest scandal and Deloitte’s UK COVID test-and-trace failure are cited examples. Investors should view excessive consulting reliance as a governance warning at corporates too (mergers, downsizing, buybacks justified by McKinsey).
• Procurement reform is leverage. NASA’s shift from cost-plus to outcomes-oriented procurement seeded camera phones, software, materials, and electronics. Today, ~30% of government budgets flow through procurement; outcome-based procurement is a forward indicator of which private firms get scaled.
• Public-private collaboration with conditionality (Oxford/AstraZeneca COVID vaccine model) is the template. Firms accepting patent pooling and price conditions in exchange for public research access can win durable contracts.
• Cities are an underrated arena. Barcelona’s insourcing of hackers under Ada Colau and the UK’s Government Digital Services (gov.uk) show that city/agency-level innovation can attract top talent away from the private sector - watch civic tech procurement.
• Geographies and entities to watch for thoughtful industrial strategy: Spain (current government’s economic thinking), Brazil (BNDES + ecological transition mission), Sweden (Vinnova, fossil-free welfare state mission), Germany (KFW), UK GDS-style digital agencies, Chile’s Corfo, Finland’s Sitra. Avoid assuming the US “doesn’t have” industrial policy - it always has, just covertly.

No specific public equity tickers are recommended. The implicit investment posture: be cautious on companies whose excess returns depend on under-taxed appropriation of public R&D (notably AI mega-caps); favor companies positioned to win conditional public capital and outcomes-based procurement (green steel, climate tech, public-health-aligned AI, civic tech).

Chapter Summaries

• Madrid and Mayors: Tracy and Joe set the scene at Bloomberg City Lab in Madrid, noting that mayors share solutions across cities in a way national leaders rarely do.
• Meet Mariana Mazzucato: The long-requested guest joins; she discusses advising Spain’s Prime Minister and launching the Global Council for a Common Good Economy, plus a Public Sector Capability Index with Bloomberg.
• The Entrepreneurial State Revisited: Mazzucato argues the post-2020 industrial policy revival isn’t automatically strategic - much of it is still subsidies that socialize risk and privatize reward. Real strategy means picking the willing, not picking winners, and orienting around bold problems (moonshots).
• State Capacity vs Capabilities: She distinguishes capacity (budget, headcount), administrative routines, and dynamic capabilities (agility, learning by doing, partnership skill) - the last being most lacking and most important.
• The Big Con (Consultants): Critique of consultant overreach since the 1980s downsizing era. Issues include conflicts of interest (PwC Australia, Eskom South Africa), lack of expertise, and the diffusion of accountability (“nobody got fired for hiring McKinsey”). NASA’s Ernest Brackett warned of capture by “brochuremanship.”
• When Consultants Make Sense: NYC trash redesign as a one-off task is reasonable to outsource; the key is government having internal capability to know who to hire and how to write terms of reference.
• Missions and Procurement: NASA’s outcomes-based procurement, working with 400,000 private-sector partners, generated cross-sector innovation. Mission framing applies equally to school lunches, pandemic prep, and trash collection.
• AI as a Different Beast: AI emerged from public investment (DARPA, NSF) but is now extracting unprecedented rents and draining academic/public talent. Governments must regulate, tax, and design conditionality early - and develop AI-specific disclosure regimes akin to climate disclosures.
• AI Bottlenecks and Health Systems: Without underlying system reform (e.g., a real health system), AI apps won’t fix anything. Risk of “bots fighting bots” in healthcare billing/insurance.
• Cities and Insourcing Talent: Barcelona’s Ada Colau hired hackers into city government; UK GDS won design awards and became a magnet for top engineers - a model for public-sector talent strategy.
• Politics, Dignity, and Populism: Stable mission-oriented agencies (DARPA-style five-year terms) buffer the electoral cycle. Real anti-populism comes from giving people dignity through co-designed policy (Camden adult social care, food cooperatives replacing food banks).
• Who’s Doing It Right: Brazil (missions at center of government, BNDES conditionality), Germany (KFW green steel), Sweden (Vinnova, fossil-free welfare state, school meals mission), UK GDS (gov.uk). Look for specific organizational examples, not perfect governments.
• Wrap-up: Tracy and Joe reflect that AI uniquely creates its own pitfalls and may need its own dedicated mission, not just be a tool serving other missions.