Franklin Templeton CEO on SpaceX, Anthropic, Circular Deals & the State of the Market

Summary

• Jenny Johnson’s constructive market view rests on earnings and cash flow, alongside caution on leverage and sticky inflation. She thinks roughly 70% of companies beat earnings projections and sees markets broadening beyond hyperscalers. The 2% productivity reading, she argues, reflects earlier technologies; meaningful AI contributions to earnings still lie ahead because adoption requires organized data and “change management.”
• Johnson favors shorter-term AI financing and would personally avoid longer-term data-center financing tied to today’s technology. Extrapolating current chip and energy requirements for 20 or 30 years overlooks the innovation that demand encourages. Today’s borrowers have substantial earnings, but circular financing and unclear revenue recognition require scrutiny: “Know your stuff as you’re making those investments.”
• Franklin Templeton’s SpaceX and Anthropic positions came through late-stage investing intended to recover the old IPO return opportunity. Companies staying private longer meant portfolio managers were missing their “IPO kickers.” Johnson says mutual funds or ETFs can hold up to 15% illiquid assets, but suggests 6%, 7% or 8% to leave room for market movements.
• Johnson is enthusiastic about SpaceX’s innovation, citing Starlink’s usefulness and Musk’s work in space and robotics. She says colonizing Mars is quite possible, while emphasizing “a lot of key-man risk with Elon.” She does not directly endorse Musk’s separate claim that SpaceX will become the most profitable company on Earth.
• Client consolidation underpins the 17 acquisitions in six years cited by O’Shea. The $4.5 billion Legg Mason deal balanced wealth and institutional channels at 50/50; Johnson estimates Canvas grew from about $2.5 billion to $30 billion in roughly four years. She acknowledges Lexington was “fully priced,” but secondary private-equity growth exceeded their models.
• Tokenization’s clearest benefit in Johnson’s account is lower servicing costs and faster access to earned cash. Franklin’s traditional money-market fund requires $500 to open an account; its blockchain version opens at $20 and probably could reach $5. Second-by-second interest calculation and daily cash payments illustrate why she believes blockchain “will replace the rails in financial services.”
• AI agents are already producing an early sales result, while their eventual contribution to model-provider revenue remains uncertain. A Microsoft-supported deployment delivered 23% more client visits and 11% higher sales in early results. Johnson says firms will need to consider agent frequency and whether some queries can use free models; O’Shea describes governance that routes requests between models and cites companies shifting to 99% open-weight or open-source usage.

Deep dive

1. Earnings are holding up before AI lifts productivity

• Johnson separates “noise” around AI, circular financing and revenue recognition from the broader economy. Inflation is “probably pretty sticky” and rates probably stay higher; today’s level is unusual against the last 25 years, less so historically. Leveraged businesses face pressure, while the US consumer remains strong. She thinks “something like 70%” of companies beat earnings projections and sees markets broadening beyond hyperscalers.
• Her geopolitical example is the $20,000 drone that can bring a tanker to a halt, make insurance hard to obtain and effectively stop oil moving through the Strait of Hormuz. More oil is getting through than expected, but she sees no quick resolution in the Middle East or Russia–Ukraine. US–China relations look more constructive, though she says each country is heading in its own direction and there are lines neither will cross. The US needs China for rare earths; China needs US consumers for its manufactured goods.
• AI is her “biggest growth lever,” but Johnson attributes the current 2% productivity reading to earlier digitization and cloud adoption. Running technology taught her that the challenges are data quality and organization—not the system integration people often assume—and “change management,” because “people don’t like change.” Meaningful AI contributions to earnings will take time; that remaining opportunity supports her optimism.
• Her jobs argument reaches back to a 1997 BusinessWeek cover predicting the internet would eliminate brokers: advisers now provide broader financial planning. Efficiency comes first, then employee creativity. Wearing basic Ray-Ban Meta glasses golfing, she wondered why they could not find her ball; her programmer son is working on an app. “We find new things that we can’t imagine initially.”

2. Long-term AI financing bets on today’s technology

• O’Shea’s challenge, citing Ken Griffin, is whether chip supply, demand and obsolescence undermine the buildout. Johnson recalls a Gartner curve—she is unsure of its exact name—of excitement, disillusionment and recovery: “So the question is where are we on that curve?” She sees “a frothiness in some areas,” yet distinguishes today’s large, cash-generating borrowers from internet-bubble businesses without revenue.
• Her financing preference is shorter term; she personally would avoid longer-term data-center financing. Extrapolating today’s chip and energy requirements for 20 or 30 years assumes insufficient innovation precisely when demand encourages breakthroughs. She thinks Griffin is “probably right on that piece of it.” Revenue recognition also leaves company comparisons unclear: “Know your stuff as you’re making those investments.” That supports her preference for active management.

3. Longer paths to IPO drove SpaceX and Anthropic exposure

• Companies staying private longer deprived portfolio managers of their “IPO kickers,” Johnson explains, prompting late-stage investments including SpaceX and Anthropic. She says mutual funds or ETFs can hold up to 15% illiquid assets but might run at 6%, 7% or 8% to avoid a market-driven breach. Client interest subsequently prompted a venture group that can invest alongside traditional portfolio managers.
• O’Shea asks whether SpaceX’s Starlink, launch economics and AI positioning support Musk’s extraordinary profitability claim. Johnson does not directly assess that forecast; she first cautions against fashionable labels, noting that Franklin Resources owed its name to the 1970s appeal of gold and “resources.” Her instruction is “stick to your knitting and say what you are.” Her concrete customer endorsement is Starlink, “a game changer” for internet access at her Montana property.
• Johnson credits Musk with challenging the system around space investment and driving cost efficiency. She says his Mars ambition may not arrive on schedule, but “he often delivers.” She sees colonizing Mars as quite possible and points to robotics as part of that vision, while emphasizing “a lot of key-man risk with Elon.” Her closing choice of Musk as the closest modern Benjamin Franklin carries the qualification that there is “no perfect match.”
• Johnson traces the pooled-fund idea to giving ordinary investors access to equity-market returns through expert advice, reasonable prices and diversification. Her father took over the Franklin Funds subsidiary with $2.5 million in AUM and $10,000 in revenue in 1955; it took until 1983 or 1984 to reach $1 billion. She sees a parallel in private markets and wants access to be fairer, while recognizing these are “absolutely illiquid vehicles.”

4. Acquisitions buy distribution as clients consolidate managers

• The $4.5 billion Legg Mason acquisition, which O’Shea says doubled AUM, filled core fixed-income and private-market gaps, including through Clarion Partners. Franklin’s 75% wealth/25% institutional channel mix met Legg Mason’s inverse, producing a 50/50 balance. The deal was announced five days after Johnson became CEO, after a year of work; COVID lockdowns followed three weeks later. She held to the strategic case despite falling markets. Selecting the “best athletes” across teams moved online, and the deal closed 60 days early.
• O’Shea cites 17 acquisitions in six years; Johnson says the logic is capability and distribution as clients choose fewer managers. Infrastructure is the capability she says remains incomplete. O’Shaughnessy Asset Management’s Canvas platform added tax-efficient separately managed accounts and technology usable over active strategies. She estimates Canvas went from about $2.5 billion at acquisition to $30 billion roughly four years later.
• Lexington Partners and Clarion Partners have more than doubled since acquisition. Johnson recalls Benefit Street Partners at roughly $26 billion when purchased; private credit now totals just under $100 billion, including subsequent acquisitions. Private credit contains distinct segments requiring origination capabilities. The latest asset-backed team acquisition filled a gap where clients were interested and spreads looked attractive.
• Deals usually used cash, sometimes with debt; part of the Putnam deal was paid with Franklin stock because Great-West Life wanted broader asset-manager exposure. Johnson estimates Putnam grew from roughly $136 billion to over $220 billion as Franklin opened previously inaccessible platforms. Lexington was “fully priced,” but growth in secondary private equity beat models. Private-market carry creates natural retention through payouts years later, while Franklin’s wealth distribution helps acquired teams reach a different buyer.

5. Tokenization changes servicing costs and cash access

• Johnson believes blockchain “will replace the rails in financial services,” slowly because it changes business models and transaction-fee economics. Three functions support her claim: one immediately updated source of truth reduces reconciliation; smart contracts embed agreed terms; and the system makes payments. Her example is collateral owed under terms sitting in a document: programming those terms into the transaction makes payment immediate.
• Franklin’s SEC-approved, natively on-chain tokenized money-market fund initially ran alongside its legacy shareholder-recordkeeping system, allowing a direct cost comparison. Traditional accounts need $500 to avoid other shareholders subsidizing servicing; the blockchain version opens at $20 and probably could reach $5. Johnson expects 24/7 trading, atomic settlement and innovation from programmable transactions, potentially combined with AI.
• Interest is calculated every second: a hedge fund posting the fund as collateral for 4 hours, 22 minutes and 32 seconds earns interest for that holding period. Cash is paid daily, versus monthly access in a traditional money-market fund. For corporate treasurers and portfolio managers, the benefit is access to earned cash they can redeploy.

6. AI agents lift sales while governance could temper model revenue

• O’Shea cites more than 1,000 agents for 800-plus investment professionals. Johnson separates the objectives: better research and performance in investing, growth in distribution, and efficiency in operations. Investment-side cost savings are “insignificant” beside better performance. Coding assistance can clear more projects from the backlog; greater capacity need not mean materially fewer people. “I think you just do more.”
• A Microsoft partnership built coordinating agents combining adviser holdings, product data, ratings and client context to help salespeople “meet the right clients and have the right conversations.” Microsoft supplied free resources because of the technical challenge. Early results showed a market leader visiting 23% more clients and an 11% increase in sales.
• The revenue uncertainty emerges after experimentation: Johnson says firms will need to ask whether agents need to run, how frequently, and whether some queries can be handled by free models. O’Shea describes a governance layer that could route requests to a top or free model, then cites emerging companies that were overloaded after 14 trillion tokens before shifting to 99% open-weight or open-source models. She suggests frontier models may still be used for biosciences and other frontier projects.