The Cracks Are Showing—And It's Not Just Fusion

Something’s breaking.

Not in the dramatic, everything-collapses way. More like the way a dam starts leaking before it fails—small fissures that tell you the pressure inside is real. The past few weeks have handed us a masterclass in what happens when the tech industry’s confidence outpaces its ability to defend what it’s built.

Let’s map this out. You’ve got quantum computing advancing fast enough that the National Security apparatus is now openly calling it a “Q-Day danger zone.” You’ve got Russia’s military hacking thousands of consumer routers. Iran-linked operators disrupting US critical infrastructure. A $15 million cryptocurrency heist blamed on “unfriendly states.” And that’s just the defensive failures—the stuff we found out about.

Meanwhile, on the optimism front: fusion startups are starting to fight with each other over fundamentals. Broadcom’s reputation is so bad that enterprises are spending real money to rip out their hardware. A legendary venture capitalist just announced he has cancer. Even SpaceX’s near-monopoly on launch is being challenged by Blue Origin actually pulling off a successful reuse of a rocket built to compete with Starship.

These aren’t unrelated news items. They’re symptoms.

The Quantum Sword Is Getting Sharper

Here’s the thing about quantum computing that most people miss: the danger isn’t theoretical anymore. “Recent advances push Big Tech closer to the Q-Day danger zone” isn’t a future-focused headline. It’s happening now.

This matters because the cryptography protecting literally everything—your banking infrastructure, military communications, the encrypted backups of every startup’s source code—assumes quantum computers don’t exist yet. The moment they do, all of that security evaporates. Not slowly. Instantly.

The timeline used to be 10-15 years. Now people who actually work on this stuff are getting nervous. That’s not hype. That’s the sound of an assumption dying.

What makes this worse is that whoever gets there first—whether it’s Google, China, or some startup funded by people we don’t know about—will have a brief window where they can decrypt everything that’s been collected over the past decade. And you can bet adversarial nation-states have been harvesting encrypted data specifically for the day quantum computers arrive.

Detailed close-up of a cracked concrete wall texture, perfect for backgrounds and design elements. Photo by cottonbro studio / Pexels

The Consumer Router Problem Is a Symptom of Worse Things

“Thousands of consumer routers hacked by Russia’s military.” Read that again.

Not sophisticated attacks on defense contractors. Not espionage against tech companies. Thousands of consumer routers. The devices sitting in people’s homes. Russia’s military thought this was worth their time.

Why? Because once you own the router, you own everything flowing through it. You own the traffic from their smart home devices, their security cameras, their work laptops that connect to corporate VPNs. You own the metadata. You own the ability to watch what they’re watching, where they’re connecting, what they’re doing.

This is different from a data breach. This is infrastructure possession. And if Russia’s military is doing it at scale, we should assume others are too.

The real tell: we only know about this one. How many hacks of consumer infrastructure have we not found out about? How many routers are compromised right now, in right this moment, running in your neighbor’s house?

The answer is almost certainly “more than anyone wants to admit.”

Fusion Is Breaking Under Its Own Weight

This one’s almost funny if it wasn’t so predictable.

Fusion startups have been getting insane amounts of capital. Commonwealth Fusion Systems, TAE Technologies, Helion—all riding the wave of belief that they’ve finally cracked it. And maybe they have. But the industry is now having disagreements serious enough that investors are worried about “deepening fissures” if they’re not settled.

Translation: nobody agrees on the physics anymore, and different companies are betting their billions on different approaches.

This happens right before money stops flowing. It’s the venture capital version of when people start whispering in a meeting instead of talking out loud. The confidence wasn’t real. It was collective delusion powered by hype and FOMO.

I’m genuinely uncertain whether fusion is actually close or whether we’ve just gotten really good at making it sound close. But I’m pretty sure: the moment one of these startups misses a major milestone by 18 months, the entire category gets marked down in venture portfolios. And that happens faster than the actual physics changes.

Close-up of hands holding a smartphone displaying 'Announcing Grok 3' on a dark background. Photo by UMA media / Pexels

The Broadcom Exodus Tells You What Enterprise Customers Really Think

“Negative” views of Broadcom are driving thousands of VMware migrations. Thousands.

This doesn’t sound like a tech story. It sounds like a boring enterprise infrastructure story. Which is exactly why it matters.

Enterprise customers don’t switch hardware suppliers lightly. The switching costs are enormous. The risk is enormous. If you’re ripping out Broadcom and installing someone else’s gear, you’ve crossed a threshold of distrust. You’re saying: “I trust this competitor more than the current vendor, even though it’s going to hurt.”

That’s what an actual loss of confidence looks like. Not complaints on Twitter. Not angry quarterly earnings calls. Actual capital expenditure being redirected.

My read: Broadcom’s acquisition of VMware was supposed to tighten the company’s grip on enterprise infrastructure. Instead, it’s triggered the exact opposite reaction. Companies are now actively building supplier diversity into their plans. And once you start that process, it compounds.

The Venture Capital Vulnerability Nobody Wants to Talk About

Ron Conway has cancer. That’s personal, and I’m genuinely sorry. But it’s also a signal.

Conway represents a type of venture capitalist that doesn’t really exist as much anymore—someone who built genuine relationships with founders and had actual pattern recognition about markets. He was SV Angel. He was one of the people who could see around corners.

The reason this matters isn’t morbid speculation about succession. It’s that the venture capital industry has been running increasingly on autopilot, with capital flowing to whoever has the best story and the best connections. When the people with actually good instincts step back, the industry’s ability to distinguish between genuine innovation and well-packaged hype gets worse.

And if you look at what’s being funded right now—and what’s not—you can see the cracks already forming.

What’s Actually Happening

The tech industry spent the last 15 years assuming abundance. Abundant capital. Abundant compute. Abundant cybersecurity theater that made things look safe without making them actually safe. Abundant confidence that whatever breaks can just be fixed with more money and more engineers.

That’s ending. Not because we’ve run out of money or engineers. Because the actual defensive infrastructure—both cyber and physical—can’t keep up with the offensive capabilities anymore. Russia can hack thousands of routers faster than router manufacturers can patch them. Quantum computers are getting closer to breaking all our encryption, and we don’t have a timeline for replacing it. Enterprise customers are losing faith in major vendors. Fusion startups are starting to squabble instead of collaborate. And the venture capitalists who were supposed to be thinking ahead are dealing with their own mortality.

This isn’t a crash. It’s a reset. And resets are usually more painful than crashes because they take longer.

Glowing digital globe display at night in Dubai Expo, showcasing illuminated continents. Photo by Denys Gromov / Pexels

What I’m Watching

Cerebras IPO pricing vs. actual demand (coming weeks): If this AI chip startup that just inked deals with AWS and OpenAI can’t command a premium IPO despite those partnerships, it signals that the market is starting to separate real AI infrastructure wins from hype. Watch the valuation vs. comparable hardware companies.
How many Fortune 500 companies announce Broadcom alternative plans in Q2 2025: If it’s more than 3-5, we’re looking at a genuine enterprise infrastructure shift, not isolated complaints. This is the leading indicator for whether Broadcom’s acquisition actually worked.
Timeline updates from Commonwealth Fusion Systems or TAE on major milestones: Fusion’s credibility window is closing. If either of these companies misses a promised achievement by more than 12 months in the next year, capital will begin rotating out systematically.
The actual deployment rate of post-quantum cryptography by major cloud providers: This is where quantum’s “imminent threat” gets real. If AWS, Google Cloud, and Azure don’t have production post-quantum encryption options available by late 2025, we’re running blind into Q-Day.