Ideas
Japan long end drives volatility
The long end of the Japanese government bond curve is selling off sharply—30-year yields rose 38 bps in two sessions—on proposals for snap elections and increased fiscal assistance. This is making Japan a source of global bond-market volatility and could reinforce the sell-America trade in US Treasuries.
All roads lead to America
He sees global capital and deal activity continuing to favor the US: all roads lead to America, capital formation and liquidity are flocking there, the M&A pipeline is the strongest he has seen across sectors, and AI is an overlay driving every company to rethink efficiency and deals. He expects the M&A and broader deal environment to persist barring a confidence shock from geopolitics.
Strong M&A cycle continues
The M&A pipeline is the strongest he has seen across sectors, with trillions of dollars of firepower incentivized to deploy and boards under pressure to buy growth if organic growth disappoints. Equity valuations and liquidity support continued deal activity, making the M&A cycle durable absent a geopolitical confidence shock.
Citi earnings power improving
Citi has moved past remediation into a high-performance, joined-up model across banking, markets, wealth and services; last year demonstrated earnings power with banking revenue up 35%, M&A up 85%, and management is leaning in to capture wallet share and cross-sell global relationships. AI is being embedded to improve controls and front-office productivity.
AI needs US power buildout
Energy is a key driver of the AI race; the White House action plan focuses on building enough US energy and data centers quickly to power AI compute. Regulatory challenges are being managed and demand can be met, but power must be built out without raising consumer electricity prices.
AI application software next wave
The digital transition is the great wave; AI code generation is the first killer use case, already a $15B business where Anthropic leads. The next wave will automate knowledge-worker workflows like Word, PowerPoint and Excel, which he thinks is 20x bigger, and he invests in asset-light AI application companies that use models rather than large language models themselves.
Private equity exits are back
Private equity is exiting a three-year winter: 2025 was the second-highest exit year for the industry, $1.3T was returned to LPs, and M&A markets are back. He expects 2026 to surpass 2025 as liquidity returns and the private-equity wheel turns again.
Diversify outside pricey US
The case for global diversification is the strongest ever: US equities trade at 23x earnings while emerging markets are at a 70% discount, fiscal and currency dynamics favor non-US assets, and for the first time in 45 years General Atlantic has more assets outside the US than in it. Policy unpredictability reinforces diversification as the only free lunch.
Critical Metals Greenland tension play
Critical Metals has its flagship mining project in Greenland and tends to rally when markets get more nervous about President Trump's ambitions for Greenland. It was up 2.5% as those geopolitical moves intensified, making it an event-driven Greenland exposure.
Lumen is AI-ready network play
Lumen is transforming from a traditional telecom into a digital network services company providing AI-ready connectivity. AI requires massive data movement between models, data centers and users; Lumen is expanding fiber and interconnecting hyperscale clouds and data centers while digitizing and programmizing the network, and networking sits on both the production and consumption sides of the AI economy.
AI buildout needs fiber/network
Fiber and network infrastructure is critical infrastructure for AI: data centers need compute and compute must be connected. The buildout of AI supply-side infrastructure will take decades because power, chips and networking all need to be constructed, and telecoms that build digital platforms on top of physical networks are advantaged.
Gita Gopinath
Professor of Economics, Harvard University; Former IMF First Deputy Managing Director
92:39
Long-end Treasury yields will rise
She thinks markets may be underestimating 2026 inflation from fiscal stimulus, continued AI spending, tariffs and dollar weakness. Regardless of the Fed, the long end of the US curve should rise: either the Fed hikes or it is seen as too dovish and inflation pushes long yields higher, a risk she says is not fully priced.
Gita Gopinath
Professor of Economics, Harvard University; Former IMF First Deputy Managing Director
95:24
AI trade is fragile
The AI trade is one of the most fragile parts of the market. Valuations across public and private AI companies imply they can all earn large profits, but competition and uncertain monetization make that inconsistent; a dot-com-style correction could wipe out about $35T in wealth. Investors have become more discerning, but the inconsistency remains.
US market takes off after geopolitics
The US economy is one of the strongest in the world and still has a good tailwind. The main thing stopping markets is geopolitical risk; when that settles, he expects the market to take off again.
AI data center buildout continues
The AI buildout is driven by forward demand, not Fed easing: AI companies feel behind the curve serving enterprise clients, so they will keep building data centers. These are long-lived expenditures financed at the long end, so long-term rates matter more than Fed funds, but demand remains the primary driver.
AI infrastructure demand is voracious
AI infrastructure demand is relentless and voracious across enterprises, governments and retail. The immediate constraint is bringing powered shells online fast enough; later constraints will shift to power, foundry chips and memory, but markets will address those infrastructure bottlenecks. This is a multiyear buildout.
CoreWeave AI demand and GPU life
CoreWeave is a key AI infrastructure provider with software that integrates Nvidia GPUs, and capital markets remain receptive to financing its buildout when structured properly. He rejects the claim that GPUs become obsolete after two to three years, pointing to sophisticated clients like Microsoft, Meta and Nvidia signing six-year compute contracts while knowing new technology is coming; the lawsuit alleging overrepresented capacity is baseless.
Chris Womack
Chairman, President, and Chief Executive Officer, Southern Company (The)
126:58
Southern benefits from data centers
Southern Company is seeing 8%-10% growth as data-center customers drive unprecedented demand. Its model is to price data-center contracts so those customers cover their costs and existing customers also benefit, allowing it to build gigawatts of new generation and provide stable returns.
Chris Womack
Chairman, President, and Chief Executive Officer, Southern Company (The)
129:43
US power demand needs buildout
US power demand from AI data centers requires an all-of-the-above buildout—natural gas, batteries and renewables—and renewables are now economic without many incentives. Utilities with the right regulatory structures and pricing models can grow while benefiting existing customers.
This Bloomberg Markets video, published January 20, 2026,
features Lisa Abramowicz, Vis Raghavan, Michael Kratsios, Martin Escobari, Dani Burger, Kate Johnson, Gita Gopinath, Gary Cohn, Mike Intrator, Chris Womack
discussing Japanese long-end government bonds, SPY, KCE, C, XLU, AI Application Software, PSP, non-US equities, EEM, Critical Metals, LUMN, Fiber/network infrastructure, US long-end Treasuries, AI equities, AI Data Centers, AIQ, CoreWeave, SO.
19 trade ideas extracted by AI with direction and confidence scoring.
Speakers:
Lisa Abramowicz,
Vis Raghavan,
Michael Kratsios,
Martin Escobari,
Dani Burger,
Kate Johnson,
Gita Gopinath,
Gary Cohn,
Mike Intrator,
Chris Womack
· Tickers:
Japanese long-end government bonds,
SPY,
KCE,
C,
XLU,
AI Application Software,
PSP,
non-US equities,
EEM,
Critical Metals,
LUMN,
Fiber/network infrastructure,
US long-end Treasuries,
AI equities,
AI Data Centers,
AIQ,
CoreWeave,
SO
