Defense contractors discover their inner VC
The most striking number of the day isn't a valuation — it's a behavior change. Traditional defense contractors are backing startups at a pace never seen before, according to a Financial Times report citing Dealroom data. Defense giants including Lockheed Martin are among the participants in that surge.
It's worth pausing on how unusual that is. The prime contractors built their business on decade-long procurement cycles, cost-plus contracts, and engineering programs measured in generations. Writing venture checks is the opposite posture: fast, speculative, and comfortable with failure. When incumbents start behaving like the people trying to disrupt them, it usually means they've concluded the technology curve is moving faster than their own internal R&D can. Watch where that money lands — it's the clearest available signal of what the primes think the next decade of defense technology actually looks like.
Washington's fight over open AI models gets two heavyweight sponsors
The day's biggest policy story arrived in three pieces that point the same direction.
Nvidia and Microsoft are reportedly making the same case to U.S. policymakers: don't clamp down on openly available AI models, per a report from Israeli business outlet Calcalist. Separately, TheStreet reports that Microsoft "just took sides in AI policy fight" — no longer a bystander in one of the loudest arguments in tech policy.
And then the flourish: Nvidia chief executive Jensen Huang joined X, and used the debut post to voice support for open-weight artificial intelligence models, according to a TweakTown report surfaced via Google News. That's not an accident of timing. A CEO who has spent years communicating through keynotes and earnings calls does not open a social account to say something incidental. The first post is the message.
The common thread: the two companies with the most to gain from a large, permissionless ecosystem of AI developers — one selling the chips, one selling the cloud and tooling — are now lobbying in public and in Washington against restrictions on open-weight models. Whatever the merits, the argument now has the industry's two largest balance sheets behind it.
Meanwhile, Congress floats a kill switch
Running against that current: U.S. lawmakers have introduced legislation dubbed the AI Kill Switch Act, a proposal aimed at giving the government a mechanism to rein in AI systems on safety grounds.
The name does a lot of work, and the details matter more than the branding. But note the structural tension with the open-weights push. A kill switch presumes there is something to switch off — a system with an operator, a chokepoint, an accountable party. Openly released model weights are, by design, the case where no such chokepoint exists. Any serious version of this bill has to answer that question, and the answer will tell you how much of the open-weight ecosystem the government intends to leave alone.
The lobbying is now electoral
The policy fight isn't staying inside hearing rooms. Big Tech companies and AI industry leaders are pouring millions of dollars into political spending ahead of the U.S. midterms, according to a report highlighted by the Financial Express.
Read the day's items together and a strategy emerges rather than a set of coincidences: executives arguing publicly, companies lobbying in Washington, and industry money moving into campaigns — all while Congress drafts restrictions. AI's other arms race is being fought with campaign contributions.
One quiet story with a long fuse
Finally, a reminder that not every threat arrives on a news cycle. Quantum computers, still early in their development, could eventually grow powerful enough to break the cryptography securing digital assets. Cryptonews.net reports that cryptocurrency exchange Coinbase is already preparing for that possibility.
That's the right instinct on a problem with this shape. Cryptographic migrations take years, and the deadline is set by an adversary's progress, not your roadmap. Preparing early looks premature right up until the moment it looks obvious.