What Tech Leaders said in last 10 days

October 30, 2026 No Comments

AI is reshaping the technology industry, but the questions facing its leaders go well beyond model performance. How much infrastructure should companies build? Who will own enterprise intelligence? How will AI agents change software, security, and the way people work? This selection of interviews, published October 20–30, 2026, brings together perspectives from leaders at AWS, Microsoft, Amazon, NVIDIA, IBM, Palo Alto Networks, OpenAI, and Google. The conversations offer a useful view of where these companies are placing their bets—and the practical and economic challenges ahead.

Eight tech leaders featured in Stackpane's What Tech Leaders Said roundup, wide landscape format

Oct. 30, 2026 – Matt Garman — AWS – Bloomberg Tech / Ed Ludlow

Main topics: Cloud growth, AI demand, and funding infrastructure expansion.

YouTube link: Watch on YouTube

Oct. 26, 2026 – Satya Nadella — Microsoft – 20VC / Harry Stebbings

Main topics: Open versus closed models, trusting AI agents, owning enterprise intelligence, and AI economics.

YouTube link: Watch on YouTube

Oct. 23, 2026 – Andy Jassy — Amazon – a16z / Ben Horowitz and Erik Torenberg

Main topics: AI infrastructure investment, AWS’s growth potential, inference costs, and AI’s impact on Amazon’s operations.

YouTube link: Watch on YouTube

Oct. 21, 2026 – Jensen Huang — NVIDIA GTC Berlin Keynote

Main topics: AI infrastructure demand, neoclouds, investment spending, and Europe’s AI opportunity.

YouTube link: Watch on YouTube

Oct. 21, 2026 – Arvind Krishna — IBM & Nikesh Arora — Palo Alto Networks – All-In

Main topics: Quantum threats to encryption, open-weight versus frontier AI, cyber warfare, and software consolidation.

YouTube link: Watch on YouTube

Oct. 21, 2026 – Sam Altman — OpenAI – Matthew Berman

Main topics: Recursive self-improvement, AI consciousness, scientific discovery, personal assistants, and hardware ambitions.

YouTube link: Watch on YouTube

Oct. 21, 2026 – Sundar Pichai — Google / Alphabet – Sources / Alex Heath

Main topics: Gemini, DeepMind leadership, compute shortages, AI infrastructure spending, personal agents, Waymo, and quantum computing.

YouTube link: Watch on YouTube

Oct. 20, 2026 – Sundar Pichai — Google / Alphabet – Big Technology / Alex Kantrowitz

Main topics: Competition with OpenAI, compute allocation between AI research and Cloud customers, AI agents, monetization, and Apple/Anthropic partnerships.

YouTube link: Watch on YouTube

Microsoft Offers Voluntary Exits to 60,000 Eligible Employees

October 26, 2026 No Comments
Microsoft voluntary exit offer: approximately 60,000 employees eligible

This is huge: Microsoft is offering voluntary exits to approximately 60,000 eligible employees at level 70 and below on participating teams. The eligible range includes employees from junior roles through senior leadership; it is not limited to junior and mid-level staff.

Eligible employees must decide whether to accept the offer in early December. Those who participate will leave on December 31, 2026.

The eligible pool represents approximately 27% of Microsoft’s reported global workforce of 223,000 employees. A subset of that pool is expected to accept the offer; the number who will leave is not yet known.

In April, Microsoft announced a voluntary retirement program for which around 7% of its U.S. employees were eligible. Then, in July, Microsoft announced plans to lay off 4,800 employees.

A detailed report from Jordan Novet, who covers Microsoft for CNBC, is available here.

My Perspective: Retain, Retrain and Redeploy

I believe opportunities elsewhere, combined with inaction inside companies, could lead to significant talent movement over the next few years.

Poor judgment today could leave companies stranded without the people they need tomorrow. When experienced employees leave, they take more than their current skills with them: they take institutional knowledge, customer relationships and an understanding of how the business works.

After more than 30 years in this industry, I believe companies should make every reasonable effort to retain, retrain and redeploy their workforce before laying people off and hiring others with newer skills.

That does not mean companies should never hire from outside. They will need new expertise. But they should first ask: Can we develop the skills we need in the people who already know our business?

The companies that get this right will give employees more than training. They will give them time to learn, opportunities to apply new skills and a reason to build their future inside the company.

Microsoft experience levels

These are illustrative software engineering levels; titles and responsibilities can vary by role and organization:

  • Entry-level / new graduates: Learning the ropes.
  • SDE II (L61–L62): Mid-level engineers owning features and growing their scope.
  • Senior SDE (L63–L64): Leading projects and mentoring others.
  • Principal and above (L65+): Shaping technology strategy and broader company goals. The eligible range extends through level 70, including some senior leaders.

#Microsoft #TechCareers #SoftwareEngineering

Google Brings Gemini Agents to Government Workflows

October 20, 2026 No Comments

Google Cloud has introduced Gemini for government services, now in preview, to help federal, state, and local agencies apply AI to everyday operations. The focus is on work shaped by regulations, sensitive information, legacy systems, and human accountability.

The offering combines government-specific instructions, connections to existing agency systems, partner-built agents, and centralized governance. Google says agencies can connect their systems without migrating data while preserving existing access permissions.

My take: The workflow is where AI earns its value

Government use cases are a useful test of enterprise AI because producing an answer is only one part of the job. That answer must reflect the applicable rules, respect access controls, cite supporting evidence, and fit an established approval process.

This announcement reinforces a broader shift I see across the industry: AI value increasingly comes from connecting models to institutional knowledge and operational systems. The model matters, but so do integration, information architecture, governance, and the people responsible for decisions.

From a systems economics perspective, agencies should measure the full cost of completing a task. A faster document review has limited value if employees spend the time saved correcting errors or checking unsupported conclusions.

The measures of success should include:

  • Service delivery: Shorter processing times and smaller backlogs.
  • Quality: Fewer errors, clearer explanations, and less rework.
  • Accountability: Traceable evidence and meaningful human oversight.
  • Economics: Lower cost per successfully completed case, including integration and review.

Because there is no one AI workload, there is no one KPI for AI. Government adoption will need to demonstrate value workflow by workflow. This preview gives Google an opportunity to show how governed agents can translate technical capability into better public service.

Where Google sees immediate applications

  • Policy research: Find answers grounded in regulations and agency guidance.
  • Public records requests: Locate and organize documents for Freedom of Information Act reviews.
  • Document redaction: Identify sensitive information and propose redactions for human approval.
  • Procurement: Evaluate vendor bids against requirements and document scoring.
  • Grant audits: Flag questionable expenses and missing documentation.
  • Application intake: Validate submitted documents and identify incomplete information.
  • Public communication: Turn complex guidance into plain-language notices.

The platform supports connections across Oracle, Salesforce, ServiceNow, Google Workspace, and Microsoft 365. Its security foundation includes an option to deploy within Google Cloud’s Assured Workloads FedRAMP High Data Boundary. Google also says agency information will not be used to train its foundation models without permission.

More from Google on this announcement here.

Why Domain-Specific AI Can Be Faster, Safer, and More Economical

October 9, 2026 No Comments

I believe domain-specific AI will become better, safer, and more economical for the enterprise work it is designed to perform. That is why I have repeatedly praised Google’s focus on industry-specific AI.

At Gemini at Work on October 8, 2026, Thomas Kurian reinforced that direction. For me, the significance goes beyond another product announcement: it points to where enterprise AI can create lasting value. We had an  analysts Q&A session with Google execs after the announcement on October 8 which brought some additional clarity around the announcements. 

Google’s industry focus is taking shape

Google’s Gemini at Work announcement identifies Financial Services and Legal as in preview, with Government, Healthcare, and Retail coming soon. These are the availability labels announced at the event, rather than a claim that all five offerings are generally available.

The direction matters. Enterprise buyers need AI that understands the context of their work: its terminology, information sources, workflows, permissions, and standards of quality. A capable model is the starting point. Putting it to productive use requires the surrounding system.

The economic unit is a useful outcome

This is where systems economics becomes central. Token pricing alone does not tell an enterprise what AI costs. The full cost includes data preparation, integration, inference, review, rework, and ongoing operations. A cheaper response that requires repeated corrections may produce a more expensive workflow. A more capable system may justify its cost if it consistently reduces that work.

My preferred question: What does it cost to complete a useful, verifiable business task? Domain-specific AI earns its economic advantage when it improves that answer after the cost of specialization is included.

Why this matters for Google

I see Google’s combination of frontier-model development and hyperscale cloud delivery as a significant strategic advantage. An industry focus gives it a way to connect those capabilities to the environments where enterprise work happens.

The opportunity is to make more of the required infrastructure, integration, and workflow knowledge reusable across customers. Execution will determine how much of that opportunity becomes a practical advantage.

I have praised Google AI’s Industry/Vertical focus, repeatedly since they have announced it. In this week’s announcements, Thomas Kurian reiterated that focus with latest Gemini At Work announcements from Mountain View! 

This is Google’s huge differentiator as a Frontier Lab and a Hyper Scaler/Cloud Provider. I am sure that they will add more industries as time progresses, tech matures, and more industries come forward to collaborate.

Google AI’s Industry Focus (so far)

– Financial Services (in preview)

– Legal (in preview)

– Government (coming soon)

– Healthcare (coming soon)

– Retail (coming soon)

 My conviction remains: domain-specific AI can turn broad model capability into better enterprise outcomes. Its value will be proven through the quality, safety, and economics of the work it completes.

Adapted and expanded from my original LinkedIn post. Graphic reused from that post. Product availability references reflect Google’s October 8, 2026 announcement.

Sarbjeet Johal on Open Source AI and Cloud Economics | theCUBE + NYSE Wired

September 1, 2026 No Comments

Where is AI creating lasting value—and who is positioned to capture it? In this conversation from theCUBE + NYSE Wired: Mixture of Experts, StackPane founder Sarbjeet Johal joins John Furrier to explore the connections between open source, global AI adoption, and the economics of AI infrastructure.

Open source and the global AI economy

Johal discusses how open source and open-weight models can expand access to AI beyond the largest technology companies. The conversation considers NVIDIA’s role in the open source ecosystem, the opportunities for developing nations and local data centers, and the pressures automation places on systems integrators.

Coding agents are another focus. Johal connects their importance to the broader work of digitizing industries, including healthcare, transportation, and infrastructure.

The economics behind AI infrastructure

The discussion also examines neoclouds and the factors that shape their businesses: token output, operating costs, and power efficiency. Johal and Furrier explore how these considerations affect emerging infrastructure providers such as CoreWeave.

They round out the conversation with the competition for AI talent, differing paths to public markets, and the partnerships established software companies are using to respond to AI disruption.

Explore the conversation

Video published by SiliconANGLE theCUBE on September 1, 2026. Watch the original interview on YouTube.

Open-Weight AI Models: Costs, Continuity, and Enterprise Risk, with Efi Pylarinou

August 25, 2026 No Comments

What does access to model weights mean for an enterprise that needs dependable AI? Sarbjeet Johal joins Dr. Efi Pylarinou on Fintech Conversations & Insights to discuss the practical economics and operational risks of open-weight models.

Ownership and operational continuity

The conversation distinguishes open-weight models from open source and examines the infrastructure needed to run them. Johal explores why possessing model weights does not, by itself, resolve questions of continuity, backward compatibility, or long-term business dependence.

Matching models to enterprise requirements

Johal and Pylarinou discuss model routing, experimentation, and the different requirements of startups and established enterprises. They also consider why systems of record and systems of engagement place different demands on AI accuracy and consistency.

The next enterprise AI architecture

The discussion looks ahead to specialized models, independent software vendors, and composite systems. It connects those possibilities with the continuing role of developers and the opportunity to make AI more broadly accessible.

Explore the conversation

Video published on YouTube on September 2, 2026, was recorded on August 25, 2026. Watch the original conversation on YouTube.

Agentic Runtime Security | Suja Viswesan and Sarbjeet Johal at IBM Think 2026

June 25, 2026 No Comments

How should security change when software agents can act on behalf of people? In this conversation from IBM Think 2026, Sarbjeet Johal speaks with IBM’s Suja Viswesan about securing agentic AI at runtime.

Identity, access, and accountability for agents

The discussion examines why controls designed for human users may need to evolve for agents. Topics include traceability, accountability, agentic identity and access management, and the risks of accumulating access or escalating privileges.

Security throughout an agent’s operation

Johal and Viswesan explore the distinction between runtime security and design-time security, as well as the relationship between human and agent privileges. They also discuss retaining context while managing permissions, the implications of composite applications, and security across products and vendors.

Explore the conversation

Video published on YouTube on June 25, 2026. Watch the original conversation on YouTube.

Cisco Live 26: Customer Experience, AI & Resilient Operations | Eric Knipp & Sarbjeet Johal

June 13, 2026 No Comments

At Cisco Live 2026, Sarbjeet Johal speaks with Eric Knipp, Cisco’s SVP and General Manager of Customer Experience, about the announcements shaping Cisco’s customer experience business and the work of bringing new capabilities into production.

Customer experience, security, and partner services

The discussion explores Cisco’s customer experience philosophy, Cisco IQ, and Cisco Cloud Control. Johal and Knipp also examine the connections between security and customer experience, the role of professional services, and how playbooks and APIs can support partners.

From automation to resilient operations

The conversation turns to practitioner training, network as code, and the relationship between agentic AI, APIs, and automation. Digital twins and systems economics round out the discussion, connecting technical capabilities with operational resilience and the journey from planning to production.

Explore the conversation

Video published on YouTube on June 13, 2026. Watch the original conversation on YouTube.

Microsoft Build Announcements and the Future of AI | Rob Tiffany and Sarbjeet Johal

June 8, 2026 No Comments

Rob Tiffany and Sarbjeet Johal discuss Microsoft Build announcements and the future of AI in this dialogue recorded on June 5, 2026.

Watch the conversation

The video brings Tiffany and Johal together for a discussion of Microsoft Build and AI’s direction. Watch the full conversation below for their perspectives.

Video published on YouTube on June 8, 2026. Watch the original conversation on YouTube.