Wall Street Bets on If, When AI Bubble Will Pop

Coming up, tech is wobbling as AI jitters hit the last full trading week of the year. Are investors pressing bets or easing off ahead of 2026? We’re also tracking big-ticket deal talks at ServiceNow and Intel, plus a crypto stablecoin giant trying to buy Juventus. Quick market check: the Nasdaq 100 is modestly lower, crypto’s under pressure, ServiceNow pops on M&A chatter, Intel rises on an AI chip target, and Tesla flirts with a fresh record on a broker upgrade.

Ryan, there’s a lot of chatter about stretched AI valuations. Is this just year‑end nerves or something deeper about returns on all this spending?

The caution has been simmering for months and it’s resurfacing. Investors are questioning valuations, growth timelines, and when tangible returns show up. Year‑end book‑closing adds to the reassessment of how far AI winners can run into 2026.

Where’s the pain most obvious?

Names tied closely to OpenAI have flipped from market darlings to underperformers as doubts grow about delivering on big partnerships. Oracle and even some mega caps have cooled, while companies aligned with Alphabet and Gemini are holding up better.

Natalie, you’ve called 2026 the prove‑it year for AI. What needs to change from 2025’s picks‑and‑shovels phase?

Adoption is broad—more than half of U.S. workers report using AI—but productivity hasn’t moved in the data. To validate valuations, we need depth of deployment and measurable efficiency gains, and that will likely show up in earnings before it hits government stats.

AI capex has propped up growth this year. How exposed is the broader economy if that spending cools or the payoff disappoints?

AI investment has been a macro anchor in 2025 as other areas softened. A meaningful pullback would ripple beyond tech, with wide economic consequences if promised productivity doesn’t materialize.

And the U.S.–China tech rift—how is that feeding through?

Despite trade uncertainty, U.S. firms have kept delivering. But deeper decoupling would dent revenue and R&D. Depending on how competition shifts, it could still create some advantages, yet the downside risks are broad.

Tesla’s rally comes as private‑market froth pops up around related assets like SpaceX. How do you parse that?

With cash yields less compelling at this stage of the rate cycle, we need to see whether high valuations reflect real infrastructure investment or a reach for alternatives. The bar for proof is high.

On to Juventus. Tether lobbed a more than one‑billion‑euro bid and got swatted away. Tomaso, what’s the playbook here?

It’s new money versus old. Tether’s founders are Turin natives and lifelong fans; they bought around ten percent and a board seat, then moved for control. The Agnelli family refused and reaffirmed a century‑long commitment. Shares jumped anyway, while the family juggles shifts across its portfolio, stirring political debate.

Netflix is pitching its Warner Bros. Discovery plan to employees. Lucas, what did leadership try to answer?

They laid out the logic, addressed approval odds, how it affects Hollywood, and the theatrical strategy. It read as much like a public case as an internal memo, and day‑to‑day operations remain steady.

And the AI divide in Hollywood?

The industry is split. Some creators want to integrate AI; others are adamantly opposed. The reaction to Disney’s tie‑up with OpenAI showed strong union pushback.

President Trump’s new one‑hundred‑thousand‑dollar H‑1B fee is drawing lawsuits. Hiba, do states have a case?

Yes. They argue it will worsen shortages in critical roles like physicians and researchers, hurting state economies. The H‑1B process already has safeguards to prevent displacement, and punitive fees risk undercutting U.S. innovation.

There’s also conservative criticism about allowing more legal immigration. What are companies weighing now?

Firms are weighing offshoring if talent won’t come here. Other countries are opening doors to high‑skilled workers, and that competition could shift where innovation happens. Both big tech and thousands of smaller employers rely on H‑1Bs.

Europe says China may grant broader rare‑earth licenses. What’s changing, and why does it matter?

We’re hearing early reports of general licenses, which would reduce intrusive data requirements and streamline access.

The EU and U.S. share worry over China’s dominance in mining and refining inputs used in everything from chips to defense. General licenses for repeat shipments to vetted buyers would ease bottlenecks, aligning with the recent trade truce signals that autos, among others, are watching closely.

AI data centers are straining the grid, and nuclear’s comeback has been slow. Will, where do things stand?

This year, capacity actually slipped as a few plants shut and only two started up. But more than sixty reactors are under construction, with a wave due from 2026 onward. The bigger choke point is grid connections, which take years; even Three Mile Island’s restart in 2027 hinged on lining up a faster hookup.

Chai Discovery just raised one hundred thirty million dollars to accelerate AI‑driven drug design. Josh, why raise again so fast, and where’s the money going?

We’ve hit research milestones that let us design molecules we didn’t think possible. We’re pouring capital into compute, scaling lab validation, building software tools, and hiring to deploy these models at scale.

Elena, what convinced you to double down now?

Last year proved AI can discover; next year is about deploying inside pharma. Teams that adopt will leap ahead, and some medicines will only be found with this technology.

Josh, your latest antibody model drew attention. How big was the jump?

We aimed for a one‑percent hit rate in the lab and reached roughly fifteen to twenty percent on this task. That moves it from curiosity to a tool for real discovery.

Elena, help non‑experts understand the step change.

Traditionally you hunt for a needle in a haystack. Chai shows most designs meet core drug‑like criteria, then go through the usual lab and clinical steps. The headline for patients is medicines with known biology and previously impossible chemistry starting to appear.

Quick M&A sweep: iRobot files for bankruptcy; McKinsey weighs cuts; ServiceNow is in talks to buy Armis, and Intel is eyeing SambaNova. Ryan, why do these deals make sense now?

ServiceNow is bulking up security to meet AI‑era threats, part of a broader cyber land grab after several large deals this year. For Intel, SambaNova would give a clearer AI chip platform and reconnect the CEO with a company he helped back, putting Intel deeper into the enterprise AI conversation—even if the headline pressure hits near term.

Before we go, Tesla is the midday standout, up more than four percent and inching toward a fresh record. That does it for Bloomberg Tech. Catch the podcast on the Terminal, Apple, Spotify, and iHeart. I’m live from London today and back in New York tomorrow.