🤚 The Open-Palm Earnings Report

Ladies, gentlemen, and algorithmically generated middle managers of the readership — Google has just posted its strongest quarter in four years, and the reason is exactly what you think it is, but somehow worse.

Alphabet’s Q1 2026 results landed like a gold-plated anvil on the analyst community: $109.9 billion in consolidated revenue, up 22% year-over-year. Google Cloud alone surged 63%, blowing past $20 billion for the first time, while its AI-specific product revenue grew nearly 800% year-over-year. The company’s cloud backlog nearly doubled to over $460 billion. Sundar Pichai, radiating the calm energy of a man whose stock options just solved three generations of family wealth, noted that AI has become Google’s “primary growth driver for cloud for the first time.”

Meanwhile, the capital expenditure guidance was revised upward — again — to $180–190 billion for 2026. That’s not a budget. That’s a small European country’s GDP being fed into GPU clusters because a chatbot needs to understand sarcasm.

👐 The Two-Handed Benchmark Tango

But Google wasn’t the only one having a week. OpenAI’s GPT-5.5, released with the quiet confidence of a predator entering a room it already owns, has been evaluated by the UK AI Safety Institute (AISI) — and the results are, shall we say, uncomfortable for everyone involved.

GPT-5.5 scored a 71.4% pass rate on AISI’s Expert-tier cyber tasks. Anthropic’s Claude Mythos Preview, the model that has haunted the nightmares of national security officials since its emergence, scored 68.6%. For context, GPT-5.4 managed a modest 52.4%, and Opus 4.7 came in at 48.6%. GPT-5.5 also completed AISI’s notorious 32-step “The Last Ones” corporate-network attack range end-to-end in 2 of 10 attempts — making it only the second model ever to do so, after Mythos managed 3 of 10.

To translate for the non-technical reader: two competing AI models are now roughly equally capable of compromising corporate networks in simulated environments, and we are all calmly discussing this over morning coffee as though it were a product launch and not a civilization-level escalation event.

🌿 The Gentle Awakening

Which brings us, inevitably, to the White House. The Trump administration is reportedly considering an executive order that would establish a working group to explore vetting AI models before public release. The proposal would include tech executives and government officials examining review procedures, with senior officials having already briefed executives from Anthropic, Google, and OpenAI on the plans.

The catalyst? Anthropic’s Mythos model, whose offensive cyber capabilities were apparently alarming enough to make Washington consider the bureaucratic equivalent of a fire alarm. And now that GPT-5.5 has matched those capabilities, the alarm has become a chorus.

There is a beautiful irony here: the same government that cleared seven AI companies — including OpenAI, Google, Microsoft, Amazon, SpaceX, NVIDIA, and Reflection — to deploy their AI on classified military networks is now wondering whether perhaps it should check what these models can do before they’re released to the public. The Pentagon already approved these systems for Impact Level 6 and 7 networks — the most classified environments imaginable — while the civilian side of the government is still in the “maybe we should form a committee” phase.

This is like locking the front door after you’ve already invited the guests to sleep in the vault.

👑 The Crown Verdict

Peter Diamandis and his panel on Episode 254 of Moonshots laid all of this out with the breathless optimism of men who have clearly invested in the right things at the right time. And they’re not wrong — the numbers don’t lie. Google’s AI-driven earnings are proof that the AI economy is not a speculative bubble; it is a structural transformation of how revenue is generated. The capex commitments — nearly $200 billion from Google alone — are not bets. They are declarations.

But the benchmark convergence between GPT-5.5 and Mythos raises a question that no earnings call can answer: what happens when the models outpace the institutions meant to govern them? The White House is debating whether to form a working group. The models are already completing attack ranges. The gap between capability and oversight isn’t closing — it’s accelerating.

Google’s stock price says everything is fine. The AISI evaluation says everything is interesting. And the difference between “fine” and “interesting,” in matters of national security, is the difference between a good quarter and a very long decade.

Inspired by Google’s Record Quarter, the White House Intervenes, and GPT 5.5 Silently Matches Mythos | EP 254 by Peter H. Diamandis.

Your quarterly report is showing. Invest wisely.