Fastly

Get more powerful websites and applications with Fastly’s edge cloud platform. We help places like Reddit, Epic Games, the NYT, and GitHub do so every day.

[Fastly @fastly]

Can AI agents buy from your store today? Or are they getting blocked at the door?

See where your store stands with our agentic commerce readiness quiz and receive a free 1:1 assessment with a Fastly expert.

👉Take assessment: https://t.co/8hFSRXBlg3

#AgenticCommerce

[Fastly @fastly]

What to do now: find your mod_php/php-fpm servers, patch with urgency (8.5.12, 8.4.27, 8.3.36, 8.2.35; older versions via your distributor), and NGWAF customers should apply the Virtual Patch.

Full analysis from Kelly Shortridge: https://t.co/PDLKaEWKWS

[Fastly @fastly]

php2shell (CVE-2026-109618 & CVE-2026-109619) was disclosed today. Every maintained PHP version is affected, and there's evidence of in-the-wild attacks before disclosure.

https://t.co/PDLKaEWKWS

[Fastly @fastly]

The root cause: when PHP hits a memory limit, timeout, or dropped connection, it abandons work mid-operation and leaves inconsistent state. Unauthenticated HTTP requests can trigger it. Patches fix the known bugs, but the mechanism remains, so expect variants.

[Fastly @fastly]

Anyone can claim to be GPTBot. In one day, 6.3M requests from ~10,200 IPs failed bot verification, with nearly 1,800 IPs probing for .env files.

🛡️ See how Bot Management helps stop the threat: https://t.co/ZIFqMSgeFG

#Cybersecurity #BotManagement

[Fastly @fastly]

58.98ms response times. 70M pageviews in a single day. ~100M malicious requests blocked monthly.

See how Le Monde uses Fastly Compute, Next-Gen WAF, and AI Bot Management with ContentGuard to personalize news at the edge and protect its 3M article archive from unlicensed AI scrapers.

Read the full story 👉 https://t.co/79LlvzfQc0

#CustomerStory

[Fastly @fastly]

DDoS protection shouldn't mean added latency, mitigation delays, or more hardware to manage.

Fastly Network Protect delivers always-on Layer 3 & 4 defense through our global network, with inline protection and 622 Tbps of total network capacity.

Defense without detours: https://t.co/wkTLPiW5Gb

[Fastly @fastly]

📣Fastly named a Leader in IDC’s 2026 Worldwide WAAP Platforms Vendor Assessment.

What made Fastly stand out?

• Context-aware detection engine
• Adaptive Threat Engine for DDoS mitigation
• High customer retention and satisfaction metrics

https://t.co/QeH75a6oR7

[Fastly @fastly]

AI Runtime Control is a unified control plane that sits between your apps and LLM providers. Cap team AI spending, set up failover AI providers, track per-engineer AI usage, and get protection from LLM-specific attacks with AI Firewall.

Learn more: https://t.co/TcOf0XC5PB

[Fastly @fastly]

No product pitches—just real solutions for engineering and cybersecurity leaders solving mission-critical problems at scale.

See you at Xcelerate NYC?

📅 Oct 14, 2026 | 9 AM – 5 PM
📍 Convene | 360 Madison Ave, NYC
👉 Reserve your seat: https://t.co/cSah4hfU4f

[Fastly @fastly]

Not all bots are the enemy. Some are indexing you for search. Some are feeding your content into AI answers. Some are just scraping you for parts.

The hard part isn't picking a side — it's telling them apart, fast enough to matter.

We wrote about how Fastly Bot Management and ContentGuard handle that, at the edge, in a few clicks: https://t.co/Dc4Lf1hksM

[Fastly @fastly]

Fastly’s 2025 Sustainability Report highlights strong progress:

🌱 100% renewable energy commitment delivered
🌱 94% YoY reduction in market-based Scope 1 & 2 emissions
🌱 Third-party assurance on reported data

Learn more: 🔗 https://t.co/5mAjoEKCXP

[Fastly @fastly]

2:15 AM. Errors spiking. Attack? Outage? Industry-wide botnet?

Fastly AEDA (built with Google Cloud's Gemini Enterprise) figures it out for you — autonomous root-cause analysis at the edge, before traffic hits your origin.

MTTR: hours → seconds. Early access is open now. 🧵

https://t.co/BAZYz8hQJI

[Fastly @fastly]

Agents are using your digital front door for customers you never see. API Security holds AI agents to the API contract you already published. Log it or block it, service by service.

Learn more: https://t.co/vHiCw81g0T

[Fastly @fastly]

The attack arrives as ordinary content and executes in the model downstream.

Stop it with AI Firewall: inline detection of prompt injection at the edge, so malicious prompts are blocked before they reach the models they call.

https://t.co/vHiCw81g0T

[Fastly @fastly]

Why your AI strategy needs a control plane: Jonathan Speek and Terri Allegretto introduce AI Runtime Control. One endpoint across your models, virtual keys instead of raw credentials, spend attributed per key, automatic failover.

Read: https://t.co/LsC2yNejp7

[Fastly @fastly]

AI Runtime Control: one gateway across your models. Virtual keys instead of raw credentials. Cost and tokens attributed per key, per session, per person. Bring your own keys. Routing serves your economics, not ours.

Learn more: https://t.co/yRTkGtVSHs

[Fastly @fastly]

How to govern, protect, and control every AI request, in one read.

👀 Read: https://t.co/MDtzN9yNAR

[Fastly @fastly]

Which model did that call go to, what did it cost, and who authorized it? Austin Spires on AI Runtime Control: one gateway across your models, virtual keys, per-key cost attribution, searchable logs.

👉 Watch on demand: https://t.co/oI22e3DJaJ

[Fastly @fastly]

AI traffic is growing 6.5x faster than human traffic. 93% of organizations are exceeding their AI budgets (McKinsey). The question that still gets asked: what did your AI do, and was it authorized to do it?

Answer it with Fastly for AI: every request, under your control.

See how: https://t.co/A0z2mL7FB8