Top of the Morning: CIO Strategy Snapshot - Weeks when decades happen

Jason Draho · UBS Global Wealth Management

State-level data center curbs matter more than federal action; three new states hold 45% of capacity.

State and local governments can do a lot to curtail or slow data center build-outs and impose moratoriums. Newly elected Democratic governors in Texas, Georgia and Michigan could impose such restrictions, and those three states together represent about 45% of planned data center capacity, with Texas alone about 25%. What states do therefore probably matters more than what happens in Washington for the time being, and it bears directly on the AI investment thesis.

“They can do a lot to curtail, slow down data center build outs and impose moratoriums. ... Those three states combined represent about 45% of planned data center capacity build out. Texas alone is about 25%.”at 6:23

Jason Draho · UBS Global Wealth Management

State data center rules aim at self-reliance, not stopping buildout; Texas will let permitted projects finish.

Investors should not misread state and local efforts to regulate data center development as attempts to stop it entirely. Even among most Democrats, the policy goal is to make data centers more self-reliant on electricity generation, grid infrastructure and water consumption, so communities get something in return. Texas Governor-elect Gina Hinojosa has said all currently permitted data centers will be allowed to complete construction, which should cover most of what is set to be built in 2027 and 2028. The Texas Commission on Environmental Quality, a six-member appointed board, regulates data center permits, and the governor can request but not require follow-through.

“let's not misinterpret efforts at the state or local level to regulate data center development with attempts to stop it entirely. I think even among most Democrats, the policy objective in most of these cases is to make data centers more self reliant”at 7:11

Jason Draho · UBS Global Wealth Management

New state rules will slow AI investment only marginally, mainly hitting 2028-2029 rather than 2027.

New regulations and restrictions may slow AI investment, but probably only marginally, and they are more likely to affect the story in 2028 or 2029 than in 2027. A real slowdown or bust is more likely to come from bottlenecks in energy access, capital access, semiconductor access, or simply insufficient compute demand. Regulation is another factor but not a game changer. If it slows things, that could smooth the spending path, extend the cycle, reduce demand for capital and ease financing constraints.

“these new regulations and restrictions may slow AI investment, but ultimately kind of perhaps only marginally, and it's more likely to impact the story in 2028 or 2029, as opposed to 2027.”at 8:28

Jason Draho · UBS Global Wealth Management

UBS expects AI-related spending of $1 trillion in 2026, rising to $1.4 trillion in 2027.

The UBS Chief Investment Office still expects about $1 trillion of AI-related spending in 2026, rising to about $1.4 trillion in 2027. This forecast sits alongside a stable labour market, resilient consumer spending and continuing AI capex growth, which together support the overall growth outlook.

“Overall, we and CIO still expect a trillion dollars of AI related spending in 2026, and that's gonna rise to [$1.4 trillion] in 2027.”at 9:14

Jason Draho · UBS Global Wealth Management

Thursday's sharp sell-off was driven by positioning and deleveraging, not fundamental news.

No major economic policy or corporate news explains a 4%+ S&P 500 drop. Retail investors sold heavily at the open, which is unusual since they have recently bought such dips, and heavy selling of levered ETFs amplified the moves. Investment-grade credit spreads widened only 2 basis points, versus the roughly 7-8 basis points the equity move would normally imply, and spreads were slightly tighter for the week, so there was no sign of a fundamental AI financing scare. The VIX rose about 60% in one day versus roughly 40% implied by the equity move, with downside volatility surging as investors bought protection. This points to technical amplification rather than fundamentals.

“For Thursday specifically, I'd say it looks to be more of a positioning led sell off or deleveraging more so than fundamentally driven.”at 1:21

Jason Draho · UBS Global Wealth Management

Most systematic selling is done, though some retail and year-end risks linger.

Most indicators suggest a lot of capitulation has already occurred, though a little more is possible. Modelling mechanical systematic strategies (momentum, volatility targeting, risk parity) suggests most of their selling is probably finished. Retail selling is unusual for them, so they are likely done too, but this is not certain. Hedge funds cleaning up books into year-end may be reluctant to take risk and buy, even after the sell-off, because clearing a book does not mean they want to buy.

“And I'd say, most indicators would suggest a lot of capitulation, although there could still be a little bit more.”at 3:28

Jason Draho · UBS Global Wealth Management

Medium-term outlook is positive, but near-term markets will stay volatile until policy clarity emerges.

The medium-term outlook remains positive, but markets are likely to be volatile in the near term. They need more policy clarity from Washington and from state and local governments on data centres to regain confidence that AI investment will continue as expected and that returns will materialise. Some investors may be reluctant to buy after the sell-off, not because they think the AI trade is dead but because of timing and uncertainty. The macro backdrop should remain supportive over the next twelve months.

“I think the medium term outlook remains positive, but in the near term, the markets are likely to be volatile as they need more policy clarity on both from DC, but also at the state and local level regarding data centers.”at 11:55

Jason Draho · UBS Global Wealth Management

Use volatility to add to industrials, which were hit by data-centre fears and can rebound.

UBS recommends investors use market volatility to add diversified exposure to higher-quality companies and to areas less tied to the AI trade, including industrials. Industrials got hit hard last week because of data centre build-out fears, but if the macro plays out as UBS expects, those stocks can bounce back.

“areas of the market that are less tied to the EI trade, including industrials, which I think got hit hard last week because of data center build out. But again, if that's kind of flows through as we expect those stocks can bounce back.”at 13:09

Jason Draho · UBS Global Wealth Management

Rescinding Medicaid and SNAP cuts would add about 20bp to 2027 GDP, easing fiscal contraction.

Democrats in Congress will likely push to rescind spending cuts from the One Big Beautiful Bill, including Medicaid and food stamp (SNAP) cuts set to take effect by 2027. Democrats have publicly stated that if they succeed, it could add about 20 basis points to GDP growth next year relative to the current baseline. Current law implies some fiscal contraction next year, so this would mean less contraction. That is an important factor for the macro growth outlook, and the debt ceiling is expected to be reached around Q1 next year.

“if they're successful, it could add about 20 basis points of GDP growth next year relative to the current baseline. So don't view necessarily a stimulus, current law suggests some fiscal contraction next year.”at 4:52

Jason Draho · UBS Global Wealth Management

Trump's trade and foreign policy tails are fatter because he could moderate or double down.

President Trump still has considerable discretion over trade regulation and foreign policy, and could either moderate existing policies or double down on them. Nobody knows how he will respond to the change in Washington's political balance, so the tails of the Trump policy distribution are probably a little fatter because outcomes could go either direction.

“we could say that the tails of the Trump policy distribution are probably a little bit fatter just because it could go kind of either direction.”at 5:32

Jason Draho · UBS Global Wealth Management

A Democratic Congress will change federal AI policy little; Trump governs through executive actions.

Democrats controlling Congress does not change much at the federal level because they have limited policy tools. President Trump can block legislation, and most action will come through administrative actions and orders. Federal AI policy would change significantly only after a major event, such as a serious security breach that turned public opinion sharply against AI. Federal changes are therefore unlikely, and the real focus is state and local policy.

“The reality is that the Democrats controlling Congress, it doesn't change much because there's somewhat limited they can do from a policy perspective. President Trump can beat any legislation... most legislation is gonna come kind of administrative actions and orders.”at 5:48

Jason Draho · UBS Global Wealth Management

October payrolls rose 165,000, well above expectations, with the labour market stable.

Non-farm payrolls rose 165,000 in October, well above expectations. August and September were revised higher, lifting the three-month average to 137,000, versus 51,000 before that for the quarter through Q3. Together with initial jobless claims and job openings, this suggests the labour market is stable, which supports confidence in the growth outlook.

“Just looking at the payrolls report for October, the non farm payrolls increased 165,000. That was well above expectations.”at 9:43

Jason Draho · UBS Global Wealth Management

US growth should end 2026 near 2.5% and moderate to about 2% in 2027.

Q3 first estimated GDP growth was 3.4%. Stable labour markets, resilient consumer spending and continuing AI capex support solid momentum into year-end, with Q4 2026 growth versus last year around 2.5%. Growth should moderate in 2027 to around 2%, roughly trend level growth, which is still solid but slower.

“We expect to end this year on a solid momentum where growth in the Q4 of this year versus last year is going be around 2.5%. Ultimately, we expect it to moderate in 2027 to still around 2%”at 10:31

Jason Draho · UBS Global Wealth Management

Disinflation will continue in the US, but at a more gradual pace.

Inflation remains above target but is turning lower slowly. UBS expects disinflation to continue at a more gradual pace, which together with solid growth is why it still expects the Fed to keep hiking rates.

“Likewise, we'd expect disinflation to continue, but at a more gradual pace”at 10:51

Jason Draho · UBS Global Wealth Management

UBS expects three Fed rate hikes, the last one in March, and thinks markets have swung too far.

Solid growth and inflation that is above target but slowly turning lower mean UBS still expects the Fed to hike once in December, once more, and one final time in March, three hikes in total, because the economic data would warrant it. Market pricing has shifted: about 80% chance of a December hike, a full hike priced by end of January, about 1.4 hikes by March and nearly two by September. Earlier the market had been more aggressive than UBS expected, but after AI data centre concerns it has swung the other way and may have overshot.

“Hiking rates once in December, once more, one final time in March. So three hikes in total because the economic data would sort of warrant it.”at 11:00

Jason Draho · UBS Global Wealth Management

S&P 500 earnings should grow about 15% next year, though markets already worry about 2028.

UBS expects S&P 500 earnings to rise another 15% or thereabouts next year, and the macro should remain supportive over the next twelve months. The market already seems to know this and is more worried about 2028. Whether AI investment pays off over 18 to 24 months is an open question, so the longer-term picture is less certain than the next year.

“We expect earnings growth next year to be up another 15% for the S and P 500 or thereabouts.”at 12:40