Agents, Chips, and a Defense Shockwave

A day of strategic pivots: big retailers embrace AI agents, chipmakers prepare for a hardware boom, and Washington’s Anthropic ban ripples through the defense sector.

What happened
After President Trump ordered U.S. agencies to phase out Anthropic’s AI tools, defense contractors like Lockheed Martin announced they would scrub the company’s technology from their supply chains. Legal experts noted that the Pentagon’s supply‑chain‑risk authority might not justify banning contractors from using a private vendor, but firms are complying anyway to stay in line for lucrative government work. Lockheed said it expects minimal impact because it doesn’t rely on a single AI vendor, but the order could still force contractors to rebuild systems that rely on Claude models.

Why it matters
The Pentagon spends over a trillion dollars a year, so suppliers are quick to align with its preferences. This purge shows how a single executive order can cascade through the defense ecosystem, compelling private firms to dump a cutting‑edge AI vendor even amid questions about the ban’s legal basis. It also highlights how deeply Anthropic’s models were embedded in defense workflows — unwinding them could delay projects and increase costs.

What’s next
Anthropic plans to challenge the ban, arguing the government lacks authority to bar contractors from using its tools. Contractors will scramble to replace Claude with alternative models, giving rival AI suppliers an opening. Courts will ultimately decide whether the administration overreached, but until then, the industry’s shift away from Anthropic will accelerate.

What happened
On the company’s Q4 2026 earnings call, Best Buy CEO Corie Barry said the retailer wants to serve shoppers “as agentic commerce matures,” both on and off its platforms. She explained that Best Buy is redesigning BestBuy.com to be more agent‑friendly so AI agents can browse and discover products on customers’ behalf. The company is working with major AI partners like Google and OpenAI to enhance its digital experience; it also plans to integrate ChatGPT into its site to help customers search and learn about products.

Why it matters
Retailers are racing to make their sites legible to autonomous agents that will shop for consumers. By investing in “agentic commerce,” Best Buy hopes to capture early adopters as AI‑powered assistants become mainstream. The move also signals that e‑commerce giants see a near‑future where browsing and purchasing decisions are delegated to software agents.

What’s next
Expect more retailers to advertise their sites as “agent‑friendly” and forge partnerships with AI providers. As agentic commerce grows, companies will need to build infrastructure that balances automation with privacy and security. Successful platforms will be those that let agents navigate seamlessly while still giving humans transparency and control.

What happened
Chipmaker Broadcom told investors it expects AI‑chip revenue to surpass $100 billion by 2027 as demand for custom accelerators from cloud and AI companies surges. AI revenue in its fiscal first quarter doubled to $8.4 billion, and CEO Hock Tan said big tech customers plan to spend more than $630 billion on AI infrastructure this year. Broadcom designs processors for Google and OpenAI and plans to deliver one gigawatt of AI accelerators to Anthropic in 2026 and triple that in 2027.

Why it matters
As generative and physical AI workloads explode, specialized chips become the bottleneck. Broadcom’s bullish forecast underscores how demand is shifting from general‑purpose GPUs to custom silicon tailored to run large models efficiently. Supplying gigawatts of accelerators to Anthropic and other AI labs shows that hardware is now a strategic differentiator.

What’s next
Expect intensifying competition among chipmakers to secure design wins with leading AI labs and hyperscalers. Delivering on such massive orders will require new foundry capacity and supply‑chain resilience. Meanwhile, governments may scrutinize the geopolitical implications of AI‑chip supply given the technology’s strategic importance.

What Happened
A new Forbes analysis, “ROI On Mental Health Investments Recalculated Due To Low-Cost At-Scale Generative AI Psychological Guidance,” spotlights a seismic shift: at-scale generative AI is dramatically lowering the cost threshold for mental health care, making scalable, AI-powered guidance a reality. The report argues that the economics of mental health have fundamentally changed, with AI-powered chatbots and large language models slashing the cost of delivering psychological support—so much so that ROI calculations for mental health investments are being rewritten overnight.

Why It Matters
The mental health care access gap is massive—and growing. Globally, there are just 13 mental health workers per 100,000 people, and the shortage is projected to hit 10 million by 2030. In the U.S. alone, nearly 4.5 million more providers are needed, while 85% of people with mental health conditions go untreated due to cost, stigma, or lack of access. Traditional therapy? It’ll set you back $100–$500 per hour without insurance. Enter generative AI: platforms like Wysa (FDA Breakthrough Device, $19.99/month), Woebot (FDA Breakthrough Device for postpartum depression), and Flourish (validated by a 2025 Harvard RCT) are delivering always-on, evidence-based support at a fraction of the cost. The clinical evidence is mounting—RAND/JAMA Network Open found 92.7% of young users found AI mental health advice helpful, and nearly two-thirds of adults reported moderate to major improvement in wellbeing. The global AI mental health market is projected to skyrocket from $1.13B in 2023 to over $8B by 2026.

What’s Next
AI is moving from novelty to necessity in healthcare. Expect broader adoption as health systems integrate AI tools into core workflows, especially for populations underserved by traditional care. But the regulatory landscape is still catching up: while Wysa and Woebot have FDA Breakthrough Device status, no AI therapy app has full FDA approval yet, and CMS reimbursement remains a bottleneck. California’s new Companion Chatbots Act is setting the pace on safety and crisis protocols for AI mental health tools.

AI’s next phase is becoming strategic infrastructure. From defense policy to retail platforms to chip supply, the real competition is shifting from models to the systems that deploy and power them.