OpenAI proposes new AI doom scenario

OpenAI released a 13-page paper titled "Industrial Policy for the Intelligence Age" proposing policy shifts—including a national wealth fund, stronger social safety nets, and rebalancing taxes toward capital (e.g., higher capital-gains and corporate taxes)—to address large-scale AI-driven economic disruption. The paper cites existing disparities (top capital-gains rate ~20% vs top labor tax ~37%, and corporate tax cut to 21% from 35%) and urges exploring taxes on sustained AI-driven returns or automated labor. These ideas are currently politically contentious and unlikely to pass immediately but could gain traction if AI disruption materially reshapes economic and political coalitions, similar to policy responses after 2008 and 2020.

Analysis

Policy talk about capturing AI rents shifts the primary political battleground from growth-versus-tax orthodoxy to distribution-of-surplus; that will change capital allocation incentives across a multi-year horizon (24–60 months) rather than overnight. Firms that capture scarce compute and scale effects (chipmakers, hyperscalers, equipment suppliers) will see durable demand even if margins are partially re-taxed — effective profitability gets compressed, but revenue growth is stickier because replacing human labor is a one-way acceleration. Conversely, businesses whose valuations rely on persistent high multiples and buyout exit optionality (late-stage software, PE-backed platforms) have more policy risk: higher capital taxes and targeted levies reduce LBO activity and price-to-earnings uplift from buyouts, compressing terminal valuations by 10–30% in stressed scenarios.

Second-order supply-chain winners are domestic semiconductor-capex beneficiaries: onshoring subsidies and industrial policy would reroute multi-year capex to local fabs and equipment makers, favoring suppliers with US footprint or eligible for subsidies; expect capex reallocation materializing in 12–36 months with visibility improving after announced subsidy programs. A countervailing risk is that punitive, AI-specific taxes could slow enterprise AI adoption if vendors face pass-through costs or compliance burdens; adoption sensitivity will be highest in public-sector and regulated industries where procurement and compliance lag by 6–18 months. Political timing is path-dependent — a visible spike in unemployment among white-collar or gig sectors could catalyze fast legislative moves, compressing market repricing into a 6–18 month window rather than the textbook 2–5 years.

Positioning should therefore balance near-term secular AI demand with asymmetric tail hedges against distributional policy shocks. Prefer exposure to firms that receive direct capex (equipment, fab services) and hyperscalers that can internalize taxes through scale, while using long-dated, cheaper downside protection on high-multiple software and asset managers that would be hit most if capital taxation increases materially. Monitor three triggers for acceleration: (1) bipartisan hearings producing draft legislation, (2) major asset managers or sovereign funds signaling portfolio reallocation due to AI rents, (3) a labor-displacing event with >1M white-collar job impact called out in major studies — any of which shortens the policy timeline to <18 months.

Key Decisions for Investors

• Long LRCX (Lam Research) 12–36 month exposure: buy shares or 2x leveraged ETFs sized to net 3–5% portfolio. Rationale: direct beneficiary of onshore fab capex; upside if industrial policy redirects $50B+ capex domestically. Risk/Reward: downside if capex delays (Q/Q revenue swings) but asymmetric upside if subsidies accelerate—target 30–60% upside vs 20–30% downside in stress.
• Long NVDA and MSFT (35/65 split) as core AI exposure for 6–18 months, funded by trimming high-multiple SaaS names: NVDA for GPU-driven demand, MSFT for cloud + enterprise uptake. Hedge: buy 2–3 year OTM put protection on combined position (e.g., 2028 puts) sized to cap 20% drawdown. Risk/Reward: captures continued AI monetization with tail protection against retroactive capital taxes; expected return 25–50% vs capped loss to hedge cost.
• Pair trade (12–24 months): long ASML or LRCX vs short CRM (Salesforce) or another high-multiple enterprise SaaS with heavy private-market exit exposure. Rationale: ASML/LRCX benefit from capex redirection; CRM reliant on durable high multiples and M&A-driven upside. Risk/Reward: aim for 2:1 upside/downside; monitor earnings and policy headlines — stop-loss at 12% adverse move.
• Buy long-dated (2–4 year) put spreads on asset managers/PE-exposed tickers (e.g., KKR, BX) sized 1–2% of portfolio as insurance. Rationale: a capital-tax regime or wealth fund could reduce carried interest/returns and compress valuations. Risk/Reward: modest premium for meaningful downside protection if capital taxation headlines accelerate; expected payoff multiples 3–6x if policy shock occurs.