Microsoft’s plan to challenge Nvidia’s dominance in AI hardware has encountered a significant setback. A new report from The Information reveals that the company’s much-hyped AI accelerator chip, codenamed Braga, has been delayed by at least six months and is now expected to enter mass production in 2026.

More troubling for Microsoft, the chip is projected to lag far behind Nvidia’s Blackwell GPU in performance, undermining its ambition to compete on silicon and reduce dependency on external suppliers.

The report portrays a project struggling with internal disarray. Braga, part of Microsoft’s larger effort to produce custom silicon for its Azure cloud and AI infrastructure, was supposed to start rolling out in data centers in 2025. However, unexpected design revisions, pressure from OpenAI for added features, and a wave of employee exits have reportedly slowed progress. One-fifth of some project teams have left, creating a vacuum of engineering talent at a critical phase.

Sources close to the project say that the chip’s instability during simulation testing is one reason the launch had to be pushed back. Yet, instead of revising deadlines, Microsoft pressed on, leading to burnout and a disjointed development cycle.

The delay has opened a wider gap between Microsoft and Nvidia, which has already launched its next-generation Blackwell chips that are redefining the industry benchmark for AI performance. Microsoft now finds itself on the back foot, struggling to catch up in a fast-moving industry where AI model development and infrastructure evolution happen in months, not years.

Microsoft has invested in three chips so far—Braga, Braga-R, and Clea—targeted for data center deployment between 2025 and 2027. But with Braga’s delay, there is growing skepticism over whether the tech giant can meet its hardware roadmap. A separate chip initially meant for training AI models was reportedly scrapped earlier this year, narrowing Microsoft’s focus solely to inference tasks.

This setback could have broader market implications. Microsoft’s success in building competitive AI chips was expected to be a disruptive force for Nvidia, whose chips power nearly every major generative AI service today. Had Microsoft succeeded in scaling its hardware in-house, it would have sharply reduced its reliance on Nvidia GPUs, potentially denting Nvidia’s future revenue. The scenario mirrors what happened with Intel, which suffered major revenue losses when Apple began producing its own M-series chips, eliminating its need for Intel processors in Macs.

But with Microsoft now facing delays and a performance gap, Nvidia appears to have a longer runway. CEO Jensen Huang has downplayed concerns about competition, stating: “What’s the point of building an ASIC if it’s not going to be better than the one you can buy?”

Microsoft’s challenges now seem to underscore his point.

The company’s earlier chip, the Maia 100, unveiled in 2023, has mostly remained in internal testing. Designed before the explosion of large language models, the Maia chip is geared more toward image processing than generative AI. Sources inside the company say Maia has not powered any of Microsoft’s flagship AI services, such as Copilot or OpenAI integrations.

If current trends continue, Clea, the final chip in Microsoft’s roadmap, may become the company’s first truly competitive offering. It’s not expected until 2027, which could leave Microsoft trailing Nvidia—and others—for at least another two years. That’s a lifetime in AI.

Until then, Microsoft will likely remain a major customer of Nvidia, funneling billions into buying its GPUs for use in Azure and enterprise AI services. The delay not only affects Microsoft’s cost structure but also stalls its ability to compete aggressively with Google, Amazon, and Meta, all of which are also developing custom silicon to control performance and spending at scale.

In an era where AI infrastructure is seen as the next frontier of computing, Microsoft’s struggles underscore how difficult it is, even for the world’s most powerful tech firms, to move away from Nvidia’s grip.

Achieving “career peace,” is deeply rooted in the concept of self-leadership. It’s about proactively steering one’s professional journey rather than merely reacting to external forces. This involves gaining a profound understanding of one’s strengths, passions, and long-term aspirations, akin to a firm defining its mission. By actively managing one’s career trajectory – seeking continuous learning, adapting to market shifts, and cultivating relevant skills – individuals can foster a sense of control and purpose, mitigating the anxiety often associated with professional uncertainty and fostering inner calm.

Self-leadership in a career context also emphasizes strategic navigation and personal resilience.  This proactive approach to career management, including seeking mentorship and building robust networks, transforms potential challenges into opportunities for advancement. Ultimately, career peace emerges not from the absence of challenges, but from the empowered ability to lead oneself through them, creating a fulfilling and sustainable professional life.

That was my take on the topic of “Career Peace through self-leadership”. An expert in this domain will be at Tekedia Mini-MBA today to teach on this. Vera Ng’oma is a leadership, personal and career development enabler and a passionate advocate for tangible excellence. She is founder of Excellica verangoma.com, an institute that helps build stronger workplaces, enables more women advance into leadership and young people to transition well into the world of work. At Tekedia Institute, she will explain physics of Career Peace and how to attend professional equilibrium through Self-Leadership.

Sat, Jun 28 | 7pm-8pm WAT | Career Peace through self-leadership – Vera Ng’oma, Excellica Leadership | Zoom link in class board.

Tekedia Institute will webcast this lecture here today here

As AI continues to redefine how economies function, Silicon Valley has largely focused on the upside—new markets, supercharged productivity, and the rise of solo entrepreneurs running billion-dollar startups. But with that optimism has come mounting anxiety over a future where machines could replace swathes of the global workforce, especially in entry-level and white-collar sectors. Now, one of AI’s key players is trying to meet that tension head-on.

Anthropic, the San Francisco-based AI company behind Claude, on Friday launched its Economic Futures Program, a new initiative that aims to understand and prepare for the seismic labor and economic shifts that generative AI is expected to trigger. Unlike many industry efforts that focus on showcasing innovation, Anthropic’s program is explicitly focused on researching the real-world impact of AI on jobs, productivity, fiscal policy, and inequality.

“Everybody’s asking questions about what are the economic impacts [of AI], both positive and negative,” Sarah Heck, Anthropic’s Head of Policy Programs and Partnerships, told TechCrunch. “It’s really important to root these conversations in evidence and not have predetermined outcomes or views on what’s going to [happen].”

The program follows remarks by Anthropic CEO Dario Amodei, who warned in May that AI could eliminate up to 50% of entry-level white-collar jobs within the next five years, potentially pushing unemployment rates as high as 20%. That projection shared more candidly than most industry leaders are willing to admit, casts a shadow over the glowing productivity statistics that AI developers are touting to investors and governments.

Anthropic’s new initiative is part of a small but growing movement among AI firms to position themselves not just as builders of disruptive technologies but as stewards of social stability in the aftermath. The Economic Futures Program will:

• Fund empirical research on AI’s effect on labor markets and value creation.
• Convene policy symposiums in Washington, D.C. and Europe.
• Partner with academic institutions and nonprofits to build datasets tracking AI’s impact on economic systems.

Anthropic is kicking off the program with rapid grants of up to $50,000, available to individuals and teams who can deliver data-driven insights within six months. Unlike traditional research funding, peer review isn’t a requirement—what matters is speed and relevance. The company will also offer Claude API credits to assist researchers in analyzing and prototyping.

“We want to understand the transitions,” Heck said. “How are new jobs being created that nobody ever contemplated before? How are certain skills remaining valuable while others are not?”

While many policy efforts today focus narrowly on job displacement, Anthropic wants to broaden the aperture, including investigations into:

• Workflow redesign in AI-enhanced industries.
• Shifts in fiscal policy and tax structure as companies adopt AI.
• How governments should invest in education and training to build new value pipelines.

Heck emphasized that this is not a lobbying effort. Instead, the company is trying to inject real data into policy conversations that are currently filled with speculation and polarization.

Contrasts with OpenAI’s Blueprint

The move comes months after Anthropic’s chief competitor, OpenAI, released its Economic Blueprint in January. That document largely promotes adoption, infrastructure, and regional development—outlining frameworks for AI literacy, building AI economic zones, and scaling access to cloud computing. But it sidesteps the direct question of job loss, focusing instead on skilling up the workforce and creating new hubs for AI investment.

In contrast, Anthropic’s approach is more grounded in economic risk assessment. Its internal Economic Index, launched earlier this year, aggregates anonymized data to study labor shifts in real time—offering a level of transparency that few tech companies currently match.

While OpenAI’s Stargate project—a $100 billion plan to build cutting-edge data centers in partnership with Oracle and SoftBank—promises tens of thousands of construction jobs, it doesn’t directly address the fact that many of the same systems may soon automate entire departments across industries ranging from finance to marketing.

Anthropic’s Economic Futures Program also comes as part of a larger reputational recalibration happening across the tech sector. With governments from the U.S. to the EU scrutinizing AI’s social consequences, companies are increasingly trying to show that they’re not just disruptors, but also responsible stakeholders.

Earlier this week, Lyft launched a forum to gather input from drivers as it begins to roll out robotaxis—a rare case of a platform actually engaging workers likely to be displaced by automation.

Anthropic appears to be taking a similar stance, admitting that while AI might boost GDP, those gains won’t be equally distributed unless the consequences are managed intentionally and transparently.

“If there is job loss, then we should convene a collective group of thinkers to talk about mitigation,” Heck said. “If there will be huge GDP expansion, great. We should also convene policymakers to figure out what to do with that.”

With this initiative, Anthropic is betting that responsible AI will not just be an ethical imperative, but also a long-term strategic advantage—especially in an industry whose economic impact could rival the scale of the industrial revolution.