Deals, Breaches, and the Eve of Nvidia's Biggest Moment: Your February 24th Tech Roundup

 If you stopped by on Sunday, you already know the pace of this week was not going to slow down. The February 22nd post laid out a lot of threads: OpenAI entering the hardware game, Apple gearing up for a product blitz, Grok 4.2 going public, and Nvidia's earnings on the horizon. Well, Tuesday has arrived, and this is one of those days where the news does not ease you in gently. It lands all at once. We have a semiconductor deal that could reshape Nvidia's stranglehold on the AI chip market, a government-AI standoff that has taken a new turn, a data breach affecting tens of millions of Americans that deserves far more attention than it is getting, and a fresh wave of anxiety in financial markets about what all of this AI spending is actually worth. Let us get into all of it.

Meta and AMD Just Shook Up the Chip Market

The headline that moved markets this morning comes from a partnership that nobody expected to be this big. Meta and Advanced Micro Devices have announced a deal valued at more than one hundred billion dollars, in which AMD will supply Meta with six gigawatts of AI computing power over the coming years. To put that in perspective, six gigawatts of computing capacity is an almost incomprehensible amount of raw processing power. The first gigawatt of deployment is expected to begin in the second half of 2026, built around AMD's custom Instinct GPUs on their MI450 architecture paired with AMD's EPYC processors.

What makes this deal particularly interesting is its structure. AMD is not just selling chips. The company has issued Meta a performance-based warrant for up to 160 million AMD shares, roughly ten percent of the company, vesting as specific GPU shipment milestones are hit. In other words, the more Meta buys, the more AMD equity Meta earns. It is an unusual incentive structure that effectively makes Meta a co-stakeholder in AMD's success, and it mirrors a similar arrangement AMD struck with OpenAI back in October. That the two largest AI spenders on the planet are both cutting equity-linked chip deals with AMD is a signal worth paying attention to.

For the average person, what does this mean? It means competition is entering a market that Nvidia has dominated almost entirely. Nvidia's shares have barely moved in 2026 while the broader Nasdaq has pulled back more than two percent. Part of that is because Wall Street has been pricing in near-infinite demand for Nvidia GPUs, and deals like this one between Meta and AMD introduce at least the possibility that some of that demand gets redirected. If AMD can deliver on the performance promises embedded in this contract, consumers could ultimately benefit from more competitive pricing on AI-powered products and services down the road. The cost of running AI is, right now, astronomical, and anything that drives those costs down eventually shows up in the tools and services you use every day.

Grok Is Going to War, and Anthropic Is Pushing Back

The story that has been building in this space for the past week took a sharp new turn today. Elon Musk's xAI has reached an agreement with the Pentagon to deploy Grok in classified military systems, including intelligence analysis, weapons development, and battlefield operations. This is significant on its own, but it becomes even more significant when you understand the context around Anthropic, which has been covered here since the February 20th post.

The key issue is what xAI agreed to that Anthropic has refused. The Pentagon requires AI companies operating in classified environments to accept an open-ended standard allowing use for any lawful purpose. Anthropic has held firm on carve-outs that would prevent its AI from being used in mass surveillance of American citizens or deployed in fully autonomous weapons systems operating without human oversight. The Pentagon initially threatened to label Anthropic a supply chain risk, a designation typically reserved for companies with ties to hostile foreign governments. Anthropic CEO Dario Amodei was summoned to meet directly with Defense Secretary Pete Hegseth earlier this week to try to resolve the standoff.

The broader implication here is significant for anyone thinking about where AI safety fits in a world where government contracts are the prize. xAI stepped in because it was willing to drop the guardrails Anthropic insists on keeping. Google is reportedly close to its own classified agreement on similar terms, and OpenAI is reportedly in discussions as well. What this means for you is that the AI safety debate is no longer purely academic or philosophical. It is now a competitive business question that is being settled by which company is willing to sign on the dotted line first. The outcome of the Anthropic-Pentagon standoff will likely set a precedent for how AI companies are expected to operate within government and defense contracts for years to come.

The Data Breach That Has Quietly Hit 25 Million Americans

There is a story that deserves significantly more mainstream attention than it has received, and this seems like the right place to make sure it lands. Conduent, a technology company that processes government benefit systems, unemployment payments, and employee services for some of the largest corporations in the country, suffered a ransomware attack in late 2024 that was not discovered until January 2025. The breach has now been confirmed to affect at least 25 million Americans across multiple states, including roughly 15 million in Texas alone and over 10 million in Oregon.

The stolen data includes names, dates of birth, Social Security numbers, addresses, medical information, and health insurance details. That combination is particularly dangerous because it creates a full identity profile that can be used for medical fraud, targeted financial scams, and identity theft that can take years to untangle. The ransomware gang responsible, known as SafePay, claims to have exfiltrated more than eight terabytes of data during the breach, which remained undetected for nearly three months before Conduent discovered the intrusion.

If you live in Texas or Oregon, or if you have ever received government benefits or worked for a major corporation that uses Conduent's systems, you should assume your data may be compromised. The practical steps are straightforward: place a credit freeze with all three major credit bureaus, sign up for identity monitoring if you have not already, and be extra cautious about any unsolicited calls or emails that claim to be from government agencies or financial institutions over the coming months. The combination of Social Security numbers with medical and insurance data makes this breach particularly dangerous for targeted scams. This is one of those stories where the most useful thing I can do is make sure you know about it.

Nvidia Reports Tomorrow, and the Stakes Could Not Be Higher

As promised in the Sunday post, the Nvidia earnings call is tomorrow, February 25th, and I want to use today to frame what we are looking for. Wall Street consensus estimates are pointing to approximately 66 billion dollars in revenue for the quarter, which would represent roughly 68 percent year-over-year growth. Analysts at Morgan Stanley have gone on record saying they expect strong results, and RBC analysts believe Nvidia could forecast the coming quarter as much as three to ten percent above current Street estimates. Analysts have now revised Nvidia earnings expectations upward three times in the past 30 days alone. That kind of revision momentum is unusual even for Nvidia.

But here is the tension that is worth sitting with today. A new analysis by Bridgewater Associates, one of the world's most respected macro investment firms, estimates that the four largest hyperscalers, Alphabet, Amazon, Meta, and Microsoft, are on track to collectively spend about 650 billion dollars on AI infrastructure in 2026 alone. That figure is up dramatically from 410 billion dollars in 2025. Bridgewater's co-chief investment officer Greg Jensen described the current moment as a more perilous phase of the AI buildout, one characterized by rapidly increasing capital expenditure and escalating dependence on external funding. The concern is that if the products and revenue growth do not eventually match the infrastructure investment, the entire edifice becomes fragile. All of that spending eventually has to translate into something consumers and businesses value enough to pay for. Nvidia's earnings call tomorrow will either reinforce the case that demand is real and durable, or it will introduce the first hints of doubt.

A Note on Bitcoin, the Markets, and the Bigger Picture

The broader financial backdrop today is worth acknowledging. Bitcoin has extended its February decline, now down more than 19 percent this month, putting it on pace for its worst monthly performance since the crypto collapse of June 2022. The token dipped to around 62,000 dollars in Asian trading hours today, a far cry from its all-time high of 126,200 dollars. The Nasdaq is also down more than 2.5 percent for the year, with software stocks hit particularly hard. The anxiety in markets is being driven by a combination of continued tariff uncertainty under the Trump administration and growing investor unease about whether the massive AI buildout is translating into proportional earnings growth.

For anyone holding crypto or tech equities, the lesson right now is that market confidence in AI is not unconditional. It is tethered to actual earnings results, and the nervousness ahead of Nvidia's report reflects just how much expectation has been baked in. If you have been watching your retirement account or investment portfolio take a hit this month, the root cause is the same question hovering over all of tech: is the AI revolution generating real economic value fast enough to justify the capital being poured into it? Tomorrow night, Jensen Huang will give us the clearest data point we have had all year to answer that question.

What a Tuesday It Has Been

Step back for a moment and look at what just happened in a single day. A hundred-billion-dollar chip deal reshuffled the AI hardware landscape. A military AI contract drew a hard line between companies willing to remove safety guardrails and those that are not. A data breach affecting one in thirteen Americans quietly continued expanding in scope. And all of it happened while markets nervously counted down to the most consequential earnings report the tech industry has seen in years. This is what normal looks like right now. If you are not tracking this space, you are missing the decisions and events that will shape the products you use, the jobs that exist in five years, and the infrastructure that underlies everything from your government benefits to your medical records. I will be back tomorrow with the full Nvidia breakdown, the Apple product launch approaching on March 4th, and whatever else this week insists on throwing at us. Thanks for reading. See you then.

Sources:

https://www.wsj.com/tech/ai/meta-and-amd-agree-to-ai-chips-deal-worth-more-than-100-billion-9c7fd06b

https://www.tomshardware.com/tech-industry/artificial-intelligence/amd-meta-100-billion-deal

https://www.axios.com/2026/02/23/ai-defense-department-deal-musk-xai-grok

https://www.teslarati.com/xais-grok-approved-for-pentagon-classified-systems-report/

https://www.nytimes.com/2026/02/23/us/politics/pentagon-anthropic-ai.html

https://techcrunch.com/2026/02/24/conduent-data-breach-grows-affecting-at-least-25m-people/

https://www.foxnews.com/tech/conduent-data-breach-hits-millions-across-multiple-states

https://www.kiplinger.com/investing/live/nvidia-earnings-live-updates-and-commentary-february-2026

https://www.cnbc.com/2026/02/24/nvidia-earnings-collide-with-wall-street-skepticism-over-ai-spending.html

https://www.reuters.com/business/big-tech-invest-about-650-billion-ai-2026-bridgewater-says-2026-02-23/

https://www.bloomberg.com/news/articles/2026-02-24/bitcoin-btc-heads-for-worst-month-since-june-2022-crypto-winter

https://www.reuters.com/business/nvidia-results-are-ai-markets-biggest-test-amid-competitive-worries-2026-02-24/