Skip to content
AI Jun 24, 2026 6 min read

Meta Bets on Prediction Markets ('Arena'); Qualcomm Buys Modular for ~$4B

Two AI-era power moves landed in late June 2026: Mark Zuckerberg directed Meta to build a prediction markets app codenamed 'Arena,' aiming at billions of users and rivals like Polymarket and Kalshi. Separately, Qualcomm is acquiring AI software startup Modular — maker of the Mojo language and MAX engine — in an all-stock deal worth nearly $4 billion to challenge Nvidia's CUDA moat.

D

DevCraftly Team

DevCraftly

Share
Meta Bets on Prediction Markets ('Arena'); Qualcomm Buys Modular for ~$4B
Meta Bets on Prediction Markets ('Arena'); Qualcomm Buys Modular for ~$4B

Two of tech’s biggest names made distinct AI-era bets in late June 2026. Mark Zuckerberg directed Meta to build a prediction markets app, internally codenamed “Arena,” pushing the social giant into a fast-growing, contested corner of fintech. Separately, Qualcomm confirmed it is acquiring AI software startup Modular — maker of the Mojo language and MAX inference engine — in an all-stock deal valued at nearly $4 billion, a direct shot at Nvidia’s CUDA software moat.

Fast-moving story. Product status, valuations and deal terms below are point-in-time around June 23–24, 2026. Arena is unreleased and may change or be shelved; the Modular deal is still expected to close later in 2026. Verify current details before relying on specifics.

At a glance

Detail
Meta’s projectPrediction markets app, codename “Arena”
DirectionOrdered by Mark Zuckerberg; a top internal priority
ReachTargets Meta’s massive user base (~3.56B daily users)
RivalsPolymarket, Kalshi (combined ~$24B monthly volume in April 2026)
Qualcomm dealAcquiring Modularall-stock, ~$4 billion
Modular’s techMojo language + MAX engine (write once, run on any chip)
Strategic targetLoosen Nvidia’s CUDA lock-in

Meta moves into prediction markets

According to reporting, Zuckerberg has made Arena a priority and directed staff to build it. The app would target Meta’s enormous distribution — on the order of 3.56 billion daily users — and run independently from Facebook, Instagram, WhatsApp and Messenger.

Early builds are said to use a points system, with real-money wagering not ruled out for the future. That distribution is the whole story: incumbents Polymarket and Kalshi saw combined monthly global volume hit roughly $24 billion in April 2026, and a Meta entrant would arrive with unmatched reach from day one.

The caveat matters, though. Insiders describe Arena as both experimental and a top priority, and have cautioned it may not ship. Treat it as a serious signal of intent rather than a launched product.

Qualcomm buys Modular to attack the software moat

The same week, Qualcomm confirmed it is buying Modular in an all-stock deal worth nearly $4 billion as it pushes beyond its core smartphone business into AI and data centers.

What Modular actually makes is the point. Its software — most notably the Mojo programming language and the MAX inference engine — lets developers write AI code once and run it across chips from different makers (Nvidia, AMD, Intel, Qualcomm) without costly rewrites. That puts Qualcomm directly against CUDA, the platform that has helped lock millions of developers into Nvidia’s hardware.

A few details sharpen the picture:

  • Structure: Qualcomm expects to issue up to 19.2 million shares of common stock to Modular’s equity holders.
  • Valuation jump: Modular was valued near $1.6 billion in a September 2025 funding round — so a ~$4 billion price is a 150%+ markup in under nine months.
  • Timeline: the transaction is expected to close in the second half of 2026.

Why it matters

1. Distribution is Meta’s weapon. Prediction markets live or die on liquidity and users. Meta doesn’t need a better product than Polymarket or Kalshi — it needs to bolt a familiar one onto billions of existing accounts. That’s a structural threat to incumbents.

2. The CUDA moat is the real battlefield. Nvidia’s lead is as much software as silicon. By buying the team behind Mojo and MAX, Qualcomm is trying to make AI code portable across hardware — which, if it works, erodes the switching costs that keep developers on Nvidia.

3. Two paths to the same goal: own the layer that locks people in. Meta wants to own the app and the audience; Qualcomm wants to own the developer software layer. Both are bets that whoever controls the lock-in point — distribution or developer tooling — captures the value.

Bottom line

Late June 2026 produced two contrasting AI-era plays. Meta’s “Arena” is an ambitious but unconfirmed push to weaponize its billions of users against Polymarket and Kalshi in prediction markets — promising if it ships, uncertain until it does. Qualcomm’s ~$4 billion, all-stock acquisition of Modular is a concrete strategic strike at Nvidia’s CUDA moat, buying the write-once-run-anywhere software that could make chip choice a lot less sticky. Different sectors, same instinct: in AI, the durable advantage is owning the layer everyone else has to build on.


Sources: reporting from June 23–24, 2026, including The New York Times, CNBC, Bloomberg and TechCrunch on Meta’s “Arena,” and Bloomberg, Reuters and Yahoo Finance on the Qualcomm–Modular deal. Product status, valuations and deal terms are point-in-time and subject to change; verify current details before relying on them.

#meta #prediction-markets #qualcomm #modular #ai #nvidia #cuda #industry
Keep reading
Get in touch

Have a project or an idea?

We don't just write about software — we build it. Tell us what you're working on and we'll get back within 1–2 business days.