Tech Giants Unite: Meta, Microsoft, and Nvidia Ask Washington to Protect Open AI.

Inside the Battle Over Open-Weight AI: Why Silicon Valley Is Flipping the Safety Argument:

Meta, Microsoft, Nvidia, IBM, and Others Back Open-Weight AI:

Two dozen rivals rarely found on the same letterhead just asked Washington to protect open-weight models — and made the case that closed AI is the bigger security risk.

24+: Companies and organizations signing the open letter

3: Core arguments the letter builds its case on

0: Recall mechanisms once model weights are public

1: A Rare Coalition Signs Onto One Letter:

Direct competitors and organizations with almost nothing else in common just agreed on one thing: open-weight AI needs protecting.

An open letter published today carries signatures from Meta, Microsoft, Nvidia, IBM, Dell Technologies, CrowdStrike, Palantir, ServiceNow, Hugging Face, Perplexity, Mistral, Andreessen Horowitz, Y Combinator, the Linux Foundation, Mozilla, and others — roughly two dozen signatories spanning commercial rivals and groups with barely overlapping business models.

The letter urges US policymakers to protect open-weight AI, drawing a direct parallel to the open-source software movement of the 1980s, and arrives ahead of anticipated AI policy action in Washington.

Open-weight models publish their trained parameters for anyone to download, inspect, modify, and run on their own hardware. That's the opposite of closed models like the frontier products from OpenAI or Anthropic, which stay locked behind API access with the underlying weights never leaving the vendor's infrastructure.

2: The Core Argument: Open Weights Spread AI Capability:

The signatories frame open weights as the mechanism that gets AI capability out of a handful of labs and into everyday business workflows.

● Lower cost of entry — startups and public institutions that can't afford to train frontier models from scratch, or pay per-token fees at frontier prices for routine tasks, can still participate.

● More competition across the stack — from chips to cloud infrastructure to applications, which the letter argues keeps costs down and prevents value capture by a small number of providers.

● No vendor lock-in — organizations running open-weight models control their own data and can adapt models to internal requirements without depending on one vendor's roadmap or pricing decisions.

The letter frames this as capability reaching "factories, hospitals, farms, classrooms, and main street businesses" rather than staying concentrated among a few well-capitalized labs.

3: Turning the Security Argument Inside Out:

The letter's sharpest section takes the usual worry about open models and flips it around. The signatories concede the real risk: once weights are released, they're beyond the original developer's control. Modified versions are hard to trace, safety guardrails can be stripped out of fine-tuned copies, and there's no way to recall a model once it's circulating.

Their answer isn't prohibition, though — it's a comparison to cybersecurity. Defenders facing AI-equipped attackers need access to models with comparable capability to detect and simulate threats, something closed, permission-gated systems don't easily provide.

Concentrating advanced AI capability behind a small number of closed providers creates single points of failure rather than removing them — closed models aren't inherently safer just because they're harder to inspect. — Core argument of the joint industry letter

The letter draws this back to the decades-old "open-source is more secure than obscurity" debate in software security, though it stops short of citing specific vulnerability-discovery data or incident figures to support the claim as applied to AI systems specifically.

4: Carving Out Space for Distillation:

The letter also stakes out a position on one of the field's more contentious techniques. Distillation — training or improving one model using another model's outputs — is standard practice in machine learning research, used for evaluation, validation, and transferring capability between models of different sizes.

The signatories draw a line between distillation as a legitimate technique and what they call unlawful efforts to extract value from closed models, arguing the former shouldn't get caught up in restrictions aimed at the latter. The framing reads as a direct response to disputes that flared after Chinese models like DeepSeek and Kimi emerged, when several US labs suggested rivals had distilled outputs from their closed systems without authorization.

The letter's position: handle misappropriation through targeted legal and commercial mechanisms, not blanket restrictions on a technique the entire field depends on.

5: What This Signals for the Policy Fight Ahead:

This is a positioning document, not a settled outcome — and it's worth reading as one. The letter arrives without a specific legislative or regulatory proposal attached. It calls on lawmakers to expand compute access for startups and researchers, fund shared training datasets and evaluation frameworks, and avoid what it calls premature restrictions on open models.

That's less a policy conclusion than a signal of where major infrastructure and chip providers want the regulatory conversation to land: Nvidia, IBM, and Dell all have direct commercial reasons to want open-weight ecosystems to flourish, since a wider range of deployable models sells more compute and services regardless of which lab produced the weights.

For procurement teams weighing open-weight against closed-model deployments, the takeaway is that the policy environment favoring one approach over the other remains unresolved — and any future restriction on distillation or open releases could shift the economics of self-hosted AI within a single legislative cycle.

Open or Closed, Your Business Shouldn't Have to Pick a Side:

This debate is playing out at the infrastructure level, between the companies that build models. For most businesses, the more urgent question isn't open weights versus closed APIs — it's how to deploy AI now, without getting locked into one vendor's roadmap or pricing decisions while the policy fight plays out.

Agent+, Otherworlds AI's Business AI Platform, gives you that flexibility today — governed AI workflows for $297/month, powered by Google Opal automated workflows, built to adapt as the underlying model landscape shifts.

Need a build tailored to your own infrastructure and data requirements? Our custom enterprise AI builds go further.

See what Agent+ can do for your business at otherworldsai.com