DX Today AI Daily Brief - Wednesday, May 13, 2026

DX Today AI Daily Brief - Wednesday, May 13, 2026

On today's briefing: Google and SpaceX are reportedly in early talks to put A.I. data centers into orbit

Transcript

SPEAKER_05 0:00

It's Wednesday, May 13th, 2026. You're listening to the DX Today AI Daily Brief. Today, Google and SpaceX are in talks to put AI data centers into orbit. Isomorphic Labs raises$2.1 billion to scale its AI drug design engine, and Colorado lawmakers vote to gut the nation's first state AI statute. Let's get into it.

SPEAKER_00 0:26

Alphabet's Google is in early stage discussions with Elon Musk's SpaceX to deploy data centers in low Earth orbit dedicated to AI workloads, according to a Wall Street Journal report surfacing overnight. The proposal would lift power-hungry compute clusters above the atmosphere, where solar energy is abundant and cooling becomes a thermal radiation problem rather than a water problem. Google has been pursuing a parallel internal effort called Project Suncatcher since late 2025, and SpaceX has separately filed with regulators to launch up to one million orbital data center satellites. No commercial agreement has been signed. Skeptics, including some SpaceX investors, warn the unproven technology faces enormous launch costs and may never reach commercial viability.

SPEAKER_02 1:15

From orbit to the lab bench, Isomorphic Labs, the Alphabet Drug Discovery spinout led by DeepMind Chief Demis Hassabiz, announced a$2.1 billion Series B round on May 12th. Thrive Capital led the financing for the second time, with participation from the United Kingdom's Sovereign AI fund, Abu Dhabi's MGX, Alphabet's GV and Capital G arms, and Singapore's Temesec. Total outside capital raised now stands at roughly$2.6 billion. The company says the proceeds will fund global expansion, deepen its ISO DDE AI drug design engine, and advance its proprietary therapeutic pipeline. ISOMorphic is built on the Alpha Fold protein structure work that won DeepMinds John Jumper and Demise Hassabi's a share of the 2024 Nobel Prize in Chemistry.

SPEAKER_04 2:17

Now to Denver. Colorado lawmakers on May 12th passed a sharply pared-back rewrite of the state's first in-the-nation AI regulation law, ending a two-year political fight with little fanfare. The original 2024 statute would have required companies and public agencies to disclose how AI systems shape decisions in hiring, lending, and housing, and to give consumers an avenue to challenge the outputs. The compromise version stripped those obligations down to a simple notification rule. Businesses must tell consumers when AI is being used. Governor Jared Polis pushed for the rewrite, citing concerns about administrative burden on Colorado employers. While consumer advocates said the watered-down statute now offers little protection against algorithmic discrimination. Next, an OpenAI security launch.

SPEAKER_01 3:11

OpenAI on May 12th unveiled Daybreak, an autonomous cyber defense platform that the company says can find, validate, and remediate software vulnerabilities without human intervention. OpenAI says Codex, the engineering AI that underpins Daybreak, has already submitted patches for more than 3,000 critical issues across more than 1,000 open source projects. The launch sits alongside Google's warning earlier this week that it thwarted an attacker using AI to plan a mass exploitation operation against a zero-day floor. Daybreak is OpenAI's first dedicated security product, and it follows the firm's announcement of an enterprise consulting venture earlier in May. Pricing and general availability were not disclosed.

SPEAKER_03 4:00

Microsoft on May 12th detailed a new AI-assisted pipeline that generates realistic synthetic security telemetry on demand, drawn from adversary tradecraft cataloged in the MITRE ATT and CK matrix. The company says the system compresses the time to detection cycle for emerging threats from weeks to hours by giving defenders a continuous stream of fresh, labeled attack data to train and test detection rules against. The announcement came the same day Microsoft made third-party plugins generally available for Agent 365, the company's enterprise agent management platform, and hosted a live Ask Microsoft Anything session with the Agent 365 engineering team. Both moves accelerate Microsoft's push to position itself as the default control plane for managing AI agents in regulated industries.

SPEAKER_05 4:53

Hardware Partnership Next.

SPEAKER_00 4:55

Nvidia and Corning on May 12th announced a multi-year partnership to dramatically expand United States-based manufacturing of advanced optical connectivity components for AI data centers. Under the agreement, Corning will increase domestic production capacity tenfold and build three new manufacturing facilities in North Carolina and Texas, creating more than 3,000 new positions. The components used to link racks of AI accelerators with high bandwidth, low latency fiber, have become a critical bottleneck as model training clusters scale into the gigawatt range.

SPEAKER_02 5:43

Now a European deal. The German workflow automation firm N8N said on May 12th that SAP has taken a strategic stake in the company, valuing it at$5.2 billion, more than double its valuation from a funding round less than a year ago. The two firms also signed a multi-year commercial agreement embedding NLAN's open source orchestration platform directly into Joule Studio, SAP's agent-building environment within the SAP business AI platform. The tie-up is one of the largest enterprise AI infrastructure deals announced in Europe this year and reflects SAP's push to give its enterprise customers a vendor-neutral foundation for building agentic workflows that span SAP and non-SAP systems.

SPEAKER_04 6:40

Consulting BetNEX. Cap Gemini on May 12th disclosed an investment in OpenAI's deployment company, the$10 billion joint venture OpenAI finalized earlier in May to help enterprises build and roll out AI systems. Financial terms of the Cap Gemini commitment were not disclosed, but the French services giant said the deal reinforces its position at the center of enterprise AI deployment globally and complements its existing OpenAI partnership. The investment adds Cap Gemini to a consortium of 19 high-profile investors that raised more than$4 billion for the deployment company at a$10 billion pre-money valuation. Rival Anthropic launched a similar joint venture with Blackstone, Hellman, and Friedman, and Goldman Sachs earlier this month. Voice Startup Next.

SPEAKER_01 7:30

Whisperai, the developer of the voice dictation tool, WhisperFlow, is in talks to raise a new funding round that would more than double its valuation to roughly$2 billion, according to a Bloomberg report dated May 12th. The proposed round, said to be roughly$260 million, comes after the application crossed about 2.5 million global downloads between late 2025 and early 2026, and after enterprise adoption, expanded across several Fortune 500 companies. WhisperFlow lets users dictate text inside any application by tapping a shortcut, then uses on-device and cloud AI to clean grammar, formality, and tone. Investors named in the talks were not disclosed.

SPEAKER_05 8:17

Infrastructure Financing Next.

SPEAKER_03 8:19

Argentum AI on May 12th said it is expanding its focus on institutional grade AI infrastructure financing, packaging integrated capital, GPU procurement, and power-backed data center development for hyperscale buyers. The firm, founded by technology investor Andrew Sabko, says it is targeting the financing gap created by the surge in compute demand that has outpaced traditional bank balance sheets. Argentum's pitch positions AI compute capacity as a financial asset class, similar to how solar and wind projects are funded today, with long-duration power purchase and capacity contracts underwriting the equipment. The announcement comes as analysts at PitchBook estimate that three deals alone accounted for 67% of all AI capital raised in the first quarter of 2026.

SPEAKER_05 9:14

Courtroom drama next.

SPEAKER_00 9:37

An account corroborated by earlier testimony from Shivon Zeelis, the mother of several of Musk's children. The trial has already produced damaging admissions on Musk's side, including a concession that his XAI lab has used distillation from OpenAI's models in training GROC. A verdict is not expected for several weeks.

SPEAKER_02 9:58

Security guidance to close. And finally, Anthropic on May 12th convened a virtual briefing aimed squarely at Chief Information Security Officers titled Secure the Advantage, a CISO Guide to Agentic AI. The session laid out Anthropic's framework for deploying autonomous agents in regulated enterprises with emphasis on identity, scoped permissions, audit logging, and human-in-the-loop checkpoints for irreversible actions. The briefing is part of a broader Anthropic push, alongside the company's Wall Street Services joint venture, announced earlier this month to position Claude as the safer choice for enterprise agentic workloads. Anthropic also confirmed it is in talks to acquire an unnamed developer tool startup whose product is used inside both OpenAI and Google.

SPEAKER_05 10:56

That's your briefing for Wednesday, May the 13th, 2026. For DX today, stay curious.